Hicl Vol 23 PDF

econstor
A Service of
zbw
Leibniz-Informationszentrum
Wirtschaft
Make Your Publications Visible.

Leibniz Information Centre
for Economics
Kersten, Wolfgang (Ed.); Blecker, Thorsten (Ed.); Ringle, Christian M. (Ed.)
Proceedings
Digitalization in Supply Chain Management and
Logistics: Smart and Digital Solutions for an Industry
4.0 Environment
Proceedings of the Hamburg International Conference of Logistics (HICL), No. 23
Provided in Cooperation with:

Hamburg University of Technology (TUHH), Institute of Business
Logistics and General Management
Suggested Citation: Kersten, Wolfgang (Ed.); Blecker, Thorsten (Ed.); Ringle, Christian M. (Ed.)
(2017) : Digitalization in Supply Chain Management and Logistics: Smart and Digital Solutions
for an Industry 4.0 Environment, Proceedings of the Hamburg International Conference of
Logistics (HICL), No. 23, ISBN 978-3-7450-4328-0, epubli GmbH, Berlin,
http://dx.doi.org/10.15480/882.1442
This Version is available at:

http://hdl.handle.net/10419/209192
Standard-Nutzungsbedingungen: Terms of use:
Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Documents in EconStor may be saved and copied for your
Zwecken und zum Privatgebrauch gespeichert und kopiert werden. personal and scholarly purposes.
Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle You are not to copy documents for public or commercial
Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich purposes, to exhibit the documents publicly, to make them
machen, vertreiben oder anderweitig nutzen. publicly available on the internet, or to distribute or otherwise
use the documents in public.
Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen
(insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, If the documents have been made available under an Open
gelten abweichend von diesen Nutzungsbedingungen die in der dort Content Licence (especially Creative Commons Licences), you
genannten Lizenz gewährten Nutzungsrechte. may exercise further usage rights as specified in the indicated
licence.
https://creativecommons.org/licenses/by-sa/4.0/
www.econstor.eu
Proceedings of the Hamburg International Conference of Logistics (HICL) – 23
Wolfgang Kersten, Thorsten Blecker and

Christian M. Ringle (Eds.)
Digitalization in Supply Chain

Management and Logistics
HICL Proceedings
Editors: Kersten, W., Blecker, T., Ringle, C.M. and Jahn, C. 2017
Digitalization in Supply Chain Management and Logistics
Digitalization in Maritime Logistics and Sustainable Supply Chain Management.
Editors: Kersten, W., Blecker, T. and Ringle, C.M. 2015

Innovations and Strategies for Logistics and Supply Chains ISBN 978-3-7375-6206-5
Sustainability in Logistics and Supply Chain Management ISBN 978-3-7375-4057-5
Operational Excellence in Logistics and Supply Chains ISBN 978-3-7375-4056-8

Innovative Methods in Logistics and Supply Chain Management IS BN 978-3-7375-0341-9
Next Generation Supply Chains ISBN 978-3-7375-0339-6

Pioneering Solutions in Supply Chain Performance Management ISBN 978-3-8441-0267-3
Sustainability and Collaboration in Supply Chain Management ISBN 978-3-8441-0266-6

Managing the Future Supply Chain ISBN 978-3-8441-0180-5
Pioneering Supply Chain Design ISBN 978-3-8441-0181-2
…find more proceedings on https://hicl.org/publications

Proceedings of the Hamburg International Conference of Logistics (HICL)
23
Digitalization in Supply Chain

Management and Logistics
Smart and Digital Solutions for an Industry 4.0 Environment
Prof. Dr. Dr. h. c. Wolfgang Kersten

Prof. Dr. Thorsten Blecker
Prof. Dr. Christian M. Ringle
(Editors)
This work is licensed under the Creative Commons Attribution-ShareAlike 4.0 In-
ternational License. To view a copy of this license, visit https://creativecommons.
org/licenses/by-sa/4.0/ or send a letter to Creative Commons, PO Box 1866, Moun-
tain View, CA 94042, USA.
Edition 1st edition, September 2017

Publisher epubli GmbH, Berlin, www.epubli.de
Editors Wolfgang Kersten, Thorsten Blecker, Christian M. Ringle
Cover (Photo) Nick Scheerbart – unsplash.com/@nck

Layout Johannes Beyrich, Niels Hackius, Henriette Porten, Sven
Reimers
ISBN 978-3-7450-4328-0
ISSN (print) 2635-4430
ISSN (online) 2365-5070
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
I Innovation and Technology Management 1

Blockchain in Logistics and Supply Chain: Trick or Treat? . . . . . . . . . . . 3
Niels Hackius, Moritz Petersen
A Peer-To-Peer Platform for Decentralized Logistics . . . . . . . . . . . . . 19

Olivier Gallay, Kari Korpela, Niemi Tapio, Jukka K. Nurminen
IoT-coordinated Logistics in Product-Service Systems . . . . . . . . . . . . 35

Jürgen Anke
Towards a Web Based Transportation Infrastructure . . . . . . . . . . . . . 55

Florian Kleedorfer, Christian Huemer
Lead-time Optimization Potential of Digitization in Air Cargo . . . . . . . . . 75

Benjamin Bierwirth, Kai-Oliver Schocke
Investigating the Factors Influencing the Acceptance of Fully Autonomous Cars 99

Ahmed Ziad Benleulmi, Thorsten Blecker
Data Source Taxonomy for Supply Network Structure Visibility . . . . . . . . 117

Johannes Zrenner, Ahmad Pajam Hassan, Boris Otto, Jorge Marx Gómez
Transforming Blood Supply Chain Management with Internet of Things Paradigm 139
Wijai Boonyanusith, Phongchai Jittamai
E-Cash & E-Vouchers: the Digitalization of the Humanitarian Aid and Logistics 157
Loïc Cohen, Vincent Salaun
v
Contents
Business Model Innovation: A Case of the Offshore Lifting Equipment Supplier 175
Mikhail Shlopak, Richard Myhre, Gabriele Hofinger Jünge
Supply Chain Flexibility and SMEs Internationalization. . . . . . . . . . . . . 195

Sylvia Mercedes Novillo Villegas, Hans–Dietrich Haasis
II Advanced Manufacturing and Industry 4.0 213
Data Mining and Fault Tolerance in Warehousing . . . . . . . . . . . . . . . 215

Christopher Reining, Omar Bousbiba, Svenja Jungen, Michael ten Hompel
Influence of Patterns and Data-Analytics on Production logistics . . . . . . . 233

Benjamin Kormann, Susanne Altendorfer-Kaiser
Use Case of Self-Organizing Adaptive Supply Chain . . . . . . . . . . . . . 255

Julia Wagner, Henning Kontny
Supply Chain Process Oriented Technology-Framework for Industry 4.0 . . . 275

Florian Schlüter, Endric Hetterscheid
Cooperation Strategies among SMEs for Implementing Industry 4.0 . . . . . 301

Julian Marius Müller, Lukas Maier, Johannes Veile, Kai-Ingo Voigt
Design of a Cyber-Physical Production System for Semiconductor Manufacturing 319

Sophia Keil
III Risk and Security Management 341
Cybersecurity in Ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343

Jenna Ahokas, Tuomas Kiiski, Jarmo Malmsten, Lauri Ojala
Smart Supply Chain Risk Management - A Conceptual Framework . . . . . . 361

Florian Schlüter, Michael Henke
Framework for Digitalized Proactive Supply Chain Risk Management . . . . . 381

Katharina Diedrich
vi
Contents
An Adaptive Supply Chain Cyber Risk Management Methodology . . . . . . 405

Stefan Schauer, Martin Stamer, Claudia Bosse, Michalis Pavlidis, Haralambos Mouratidis, Sandra
König, Spyros Papastergiou
Simulation-Framework for Illicit-Goods Detection in Large Volume Freight . . 427

Lutz Kretschmann, Torsten Münsterberg
Relational View on Collaborative Supply Chain Disruption Recoveries . . . . 449

Marie Brüning, Julia Bendul
Automotive Supply-Chain Requirements for Knowledge Management . . . . 467

Ann-Carina Tietze, Jan Cirullies, Boris Otto
Bibliometric Analysis of Risk Management in Seaports . . . . . . . . . . . . 491

Ayman Nagi, Marius Indorf, Wolfgang Kersten
vii
Preface
This year‘s edition of the HICL proceedings are released at a time of a profound
shift: Digitalization puts existing business models to the test and questions exist-
ing business processes and logic. Companies find themselves in an increasingly
volatile environment and face an ever-increasing demand for flexibility. Logistics
and supply chain management shape these processes and long for innovative
concepts to achieve the requirements of the digital age.
The dense interweaving of the physical and digital space, of course, poses new
risks with regards to data security and creates new requirements with respect to
data exchange and acquisition. In contrast, it creates millions of opportunities to
optimize, to become more flexible and to reduce risks in supply chains.
This book focuses on the hot topics of digitalization in the supply chain. It con-
tains manuscripts by international authors providing comprehensive insights into
topics like risk management, cybersecurity, the internet of things or Blockchain
and provides future prospects for the field of supply chain management. All
manuscripts contribute to theory development and verification in their respec-
tive area of research.
We would like to thank the authors for their excellent contributions, which ad-
vance the logistics research progress. Without their support and hard work, the
creation of this volume would not have been possible.
Prof. Dr. Dr. h. c. Wolfgang Kersten

Hamburg, October 2017 Prof. Dr. Thorsten Blecker
Prof. Dr. Christian M. Ringle
ix
Part I
Innovation and Technology

Management
Blockchain in Logistics and Supply
Chain: Trick or Treat?
Niels Hackius1 , Moritz Petersen2
1 – Hamburg University of Technology
2 – Kühne Logistics University
Blockchain is an emergent technology concept that enables the decentralized and
immutable storage of verified data. Over the last few years, it has increasingly
attracted the attention of different industries. Especially in Fintech, Blockchain is
hyped as the silver bullet that might overthrow today’s payment handling. Slowly,
the logistics and supply chain management community realizes how profoundly
Blockchain could affect their industry. To shed light on this emerging field, we
conducted an online survey and asked logistics professionals for their opinion on
use case exemplars, barriers, facilitators, and the general prospects of Blockchain
in logistics and supply chain management. We found most of our participants
are fairly positive about this new technology and the benefits it offers. However,
factors like the hierarchical level, Blockchain experiences, and the industry sector
have a significant impact on the participants’ evaluation. We reason that the
benefits over existing IT solutions must be carved out more carefully and use cases
must be further explored to get a rather conservative industry, like logistics, more
excited about Blockchain.
Keywords: Blockchain; Logistics; Supply Chain Management; Use Cases
3
Blockchain in Logistics and Supply Chain: Trick or Treat?
1 Introduction
Blockchain is everywhere. Invented by Satoshi Nakamoto, one or more mysterious

individuals unmasked until today, it has been more of an insider’s tip for the
longest part of its existence. It became known to a larger audience in September
2015: nine financial companies – Goldman Sachs, Barclays, J.P. Morgan, and
others – joined forces to build a new Blockchain-based infrastructure for financial
services (Underwood 2016). By then, Blockchain had become the latest hype
in Fintech, with almost daily announcements of new startups and corporate
projects. It took longer until the logistics and supply chain management (SCM)
community caught on and slowly realized the impact Blockchain might have
on their industry. One major promise of Blockchain is to create transparency –
every member of the network has access to the same data, providing a single
point of truth (Tapscott & Tapscott 2016). Supply chain transparency is one of
the most important and hardest to achieve improvement areas for logistics and
SCM (Abeyratne & Monfared 2016). It comes as no surprise that some logistics
experts consider Blockchain to offer “enormous potential” (O’Marah 2017), to be
a “much-needed platform for economic renewal” (Casey & Wong 2017), and to
“transform the supply chain and disrupt the way we produce, market, purchase
and consume our goods” (Dickson 2016). Taken together, Blockchain might be
nothing less than the “holy grail” (Popper & Lohr 2017). However, as it often is
the case with emerging technology, the hype around Blockchain seems primarily
driven by technology providers, consultants, and journalists. Logistics operators –
especially small and medium-sized companies – declare to have little knowledge
about Blockchain (Kersten et al. 2017). This can be explained through the novelty
of the technology but also through the lack of convincing use cases that clearly
show Blockchain’s benefit over existing IT solutions. Logistics and SCM research
on Blockchain is still in its infancy (Zhao et al. 2016) and ought to look into possible
applications (Yli-Huumo et al. 2016).
The research questions for this paper emerge: “What might be suitable applica-
tions for Blockchain technology in logistics and SCM?” and “Should Blockchain in
logistics and SCM be considered a treat or rather a trick?” The remainder of the pa-
per is structured as follows: First, we summarize the basic features of Blockchain.
Then, we introduce four use case exemplars explored in theory and practice.
Subsequently, we present the findings of an international survey we conducted
within the logistics industry to look into the prospects of the four use cases and
expectations and apprehensions towards Blockchain. We conclude by discussing
4
2 Basics of Blockchain
the findings and daring a prognosis on the future of Blockchain in the logistics
industry.
2 Basics of Blockchain
The Blockchain is a distributed digital ledger of transactions that cannot be tam-

pered with due to the use of cryptographic methods (Pilkington 2016). This short
explanation includes the three most important properties of a Blockchain: de-
centralized, verified, and immutable (see Figure 1). (1) It is decentralized because
the network is entirely run by its members, without relying on a central author-
ity or centralized infrastructure that established trust. To add a transaction to
the ledger, the transaction must be shared within the Blockchain’s peer-2-peer
network. All members keep their own local copy of the ledger. (2) It is verified be-
cause the members sign the transactions using public-private-key cryptography
before sharing them with the network. Therefore, only the owner of the private
key can initiate them. However, the members can stay anonymous because the
keys are not linked to real-world identities. (3) It is immutable through its con-
sensus algorithm: One or more transactions are grouped together to form a new
block. All members of the network can verify the transactions in the block. If no
consensus on the validity of the new block is reached, the block is rejected. Like-
wise, if consensus exists that the transactions in the block are valid, the block is
added to the chain. A cryptographic hash is generated for each block. Each block
not only holds transaction records but also the hash of the previous block. This
creates a block interdependency linking up to a chain – the Blockchain. Altering a
Decentralized Verified Immutable
through peer-2-peer through signatures through consensus

network algorithm
+ + X X
Figure 1: Basic Properties of Blockchain
5
transaction on the Blockchain retroactively would require not only to alter the
local records on most of the networks members’ devices but also altering the
cryptographic hash of every block down the chain.
A distributed system, like a Blockchain, holds benefits over centralized architec-
tures as it provides the same, verified information to all network members. It
creates trust between the parties by eliminating the need for trust. Blockchain can
record the transfer of assets between two parties, without the need of a trusted
intermediary. Such assets could be digital money, but also carbon credits or other
deeds of ownership (Tapscott & Tapscott 2016).
The Bitcoin Blockchain is the first implementation of the Blockchain principles
and only supports simple transactions. It also shows how reliable Blockchain is,
as it has run error-free since January 2009 and has a current market capitalization
of over 35 bn. Euro. Centralized infrastructures, on the contrary, are increas-
ingly hacked – classic middlemen, like banks or dating websites, provide ample
examples (Tapscott & Tapscott 2016). Today, many more advanced Blockchain
implementations exist. While most are open to the public (permissionless), there
are also private (permissioned) Blockchain implementations, where the rights
to read and write are controlled by a central authority (Pilkington 2016). Some
Blockchain implementations support so-called “smart contracts” or applications
living on the Blockchain (Christidis & Devetsikiotis 2016). Smart contracts are con-
ditions written in code. The delivery of a parcel can serve as a simple example: To
counter the risk of a loss, a smart contract can be designed such that the payment
by the sender is only released once the shipping company confirms the delivery.
This allows for a transaction to be automated, yet documented and controlled.
Of note, Blockchain provides not only benefits but also has challenges attached
to it (Petersen et al. 2016; Yli-Huumo et al. 2016; Xu 2016). Most result from
the early maturity phase of the technology. While these challenges (e.g., lim-
ited throughput) have to be addressed from a technological perspective, they
should not distract possible users from evaluating the benefits of the underlying
principles.
3 Blockchain in Logistics and SCM
As introduced, Blockchain is considered to offer large potential for improving

processes and enhancing business models in logistics and SCM. However, accord-
6
1 Global container shipping still involves a lot of

paperwork – costing time and money. Also,
paper-based freight documents like the bill of
lading are prone to loss, tampering, and fraud.
Ease Paperwork Processing
2 Counterfeit medicine is a growing problem for

pharmacy supply chains. This especially pertains
to expensive, innovative medicine like cancer
drugs. Pharmacies have to make sure to sell “the
Identify Counterfeit Products right thing” to the consumers.
3 In the food supply chain, foodborne out-breaks

are a challenge for retailers. They have to get a
quick overview of where the food came from
and which other products are also affected and
Facilitate Origin Tracking have to be removed from the stores.
4 More and more logistics objects are equipped

with sensors that generate data along the
supply chain – e.g. about the status of a
shipment. This data has to be stored in an
Operate the Internet of Things immutable, accessible way.
Figure 2: Overview of Use Case Exemplars
ing to a recent study on trends in logistics and SCM, Blockchain is only known to
some logistics experts and even fewer pursue implementation plans (Kersten et
al. 2017). In this section, we shed light on Blockchain’s potential through intro-
ducing use case exemplars. These exemplars represent four major ideas currently
explored in both theory and practice. They are also four single ideas out of a yet
unmapped sea of opportunities. In choosing them, we tried to collate a broad
and multifaceted picture later used for investigating the prospects of Blockchain
for logistics and SCM. Figure 2 summarizes the use case exemplars. They are
introduced in more detail in the following sections.
3.1 Ease Paperwork Processing in Ocean Freight
International container transports have a long trail of paperwork associated with

them. For example, shipping refrigerated goods from East Africa to Europe re-
quires stamps and approvals from around 30 people and organizations that must
7
interact with each other on over 200 occasions. Also, documents like the bill of
lading might be subjected to fraud (Popper & Lohr 2017). Taken together, the
cost of the trade-related paperwork processing is estimated to be between 15
and 50 percent of the costs of the physical transport (Groenfeldt 2017; Popper
& Lohr 2017). To tackle such process inefficiencies and digitize paper records,
IBM and Maersk joined forces in 2015. They eventually settled for a permissioned
Blockchain solution as means to connect the vast global network of shippers,
carriers, ports, and customs. The implementation details still must be worked out.
However, a round of pilots in 2017 has succeeded. In these pilots, every relevant
document or approval was shadowed on the Blockchain, meaning the legacy IT
systems were not replaced but augmented. Using a standardized interface, every
partner is empowered to have full visibility of the container status (Allison 2017).
Until the end of 2017, Maersk hopes to shadow one in seven of their container
shipments on the Blockchain – around 10 million boxes per year (Groenfeldt 2017).
The problems associated with extensive paperwork are not limited to this specific
use case but hamper all kinds of trade flows (Chu et al. 2016; Morabito 2017).
3.2 Identify Counterfeit Products
The provenance of high-value items often relies on paper certificates that can
get lost or tampered with: whether a diamond’s certificate is genuine or fake
– and if the diamond was stolen – is not always easy to determine. The same
holds true for expensive wine, watches, or handbags (Lomas 2015). Since, for
example, a diamond’s serial number can easily be cut, the startup Everledger
takes an alternative approach and records 40 data points that uniquely identify a
diamond. Using these publicly available records on the Blockchain, a potential
buyer can clearly determine if the seller is the actual owner of the diamond and
can also make sure he is not buying a “blood diamond” mined in a war zone
(Underwood 2016). Everledger plans to extend this fraud detection system into
a provenance platform for many high-value items (Lomas 2015). In the medical
sector, counterfeit drugs are a known problem that – for example with anti-cancer
drugs – can even have lethal consequences if patients do not receive the treatment
as prescribed (Mackey & Nayyar 2017). Blockchain could improve patient safety
through establishing supply chain transparency from manufacturers through
wholesale and pharmacies to the individual patients. Through barcodes or auto
ID technology, patients could be empowered to check whether they received the
actual drugs (DeCovny 2017; Mackey & Nayyar 2017). Blockchain is considered to
8
make it much more difficult to tamper with products or to channel in products of

illegal origin (Sutherland et al. 2017; Apte & Petrovsky 2016; Morabito 2017).
3.3 Facilitate Origin Tracking
If faced with a foodborne disease outbreak, retailers have a hard time figuring out
where the bad ingredients came from and to which stores they were delivered
(Tian 2016). Today, it can take weeks to track down the source of the contamina-
tion and restore consumers’ confidence in food safety (Popper & Lohr 2017). To
facilitate origin tracking for food items, Walmart partnered with IBM in 2016. Like
with Maersk, Blockchain is used to augment the supply chain partners’ existing IT
systems through a transparent, superordinate ledger, tracking the movements
of food items. This shared forum is considered a substantial improvement over
Walmart’s earlier trials involving barcodes or auto ID technology – solutions that
required central databases and trust between the participants (Hackett 2016). In
some first pilots, Walmart and IBM digitally tracked both domestic movements –
pork from small Chinese farms to Chinese stores – and international movements
– produce from Latin America to stores in the United States (Popper & Lohr 2017).
In these pilots, data like the farm origin, batch numbers, factory and processing
data, expiration dates, and shipping details were written on the Blockchain and
instantly became available to all network members. With a foodborne disease
outbreak, this data enables Walmart to track down the origin in a matter of sec-
onds. During the year, further pilots with more data attributes are scheduled.
Ultimately, Walmart believes Blockchain could also reduce food waste if the newly
available data on shelf life is used as a parameter for supply chain optimization
(Shaffer 2017).
3.4 Operate the Internet of Things
The Internet of Things (IoT) means everyday objects – essentially everything

with a plug – get equipped with electronics and can exchange data over the
internet. A Gartner report estimates there will be over 20 billion connected things
by 2020 (Gartner 2015). However, the current internet architecture with its server
infrastructure might not handle such an amount of devices and data (Eastwood
2017). Single servers represent a single point of failure and raise data security
concerns. The public Blockchain ledger is considered a solution to connect and
9
manage IoT devices reliably (Pilkington 2016; Christidis & Devetsikiotis 2016).
Given the large amount of possible IoT objects (vehicles, shipments, etc.), logistics
might be one of the most promising applications for IoT and Blockchain (Zheng et
al. 2017). First large companies start to work in this area. For example, Walmart
was recently granted a patent that aims at improving last mile logistics through
connecting delivery drones to the Blockchain (Hackett 2017). Such IoT devices
connected to the Blockchain could also be provided with a digital currency. This
would enable them to interact autonomously with other parties and – through
smart contracts – to pay fees and duties by themselves, e.g., for priority access to
restricted air corridors (Christidis & Devetsikiotis 2016; Petersen et al. 2016).
4 Survey about the Prospects of Blockchain in Logistics

and SCM
To investigate how experts from the logistics industry evaluate Blockchain, we

designed and conducted an online survey. This section describes its setup and
the most important findings.
4.1 Setup and Data Collection
The survey was implemented using Typeform and had four major parts: First, we
inquired the participants’ general knowledge of logistics and SCM and Blockchain.
Second, we introduced the four use case exemplars and inquired the participants’
evaluation of Blockchain benefits and adoption likeliness. Third, we asked the
participants about their general opinion on the main beneficiaries of Blockchain in
logistics, likely adoption barriers, and the expected effect on established logistics
processes. Finally, we inquired job and company details.
We collected the data between April 28th and June 13th, 2017. Participants were
mainly recruited through social media, e.g., posts in logistics and Blockchain
interest groups on LinkedIn, Xing, or Meetup. The BVL (Bundesvereinigung Lo-
gistik e.V.) shared our call for participants through their social media channels.
We incentivized participation through small donations. After data collection was
concluded, we donated 50 Euro to “Zeit für Zukunft” (Hamburg-based mentoring
program) and 50 Euro to “Ingenieure ohne Grenzen” (Berlin-based aid organiza-
tion) on behalf of our participants.
10
4 Survey about the Prospects of Blockchain in Logistics and SCM
Especially for anonymous internet surveys, thoroughly examining the data is

advised to identify careless responses (Meade & Craig 2012). We excluded three
out of 155 initially collected data sets from the analyses due to clear answer
patterns or answers that made no sense. Then, we prepared the data for statistical
analysis following the guidelines of Hair et al. (2009). We used IBM SPSS Statistics
22 for the analyses.
4.2 Findings
We present the results of the study in the following, mainly through providing the
mean values of the participants’ evaluations. If our analyses proved them to be
statistically significant at the 5%-level, we present more detailed findings and
explore the differences between groups of participants.
Figure 3 shows an overview of the 152 participants: The clear majority works
in consulting, followed by logistics services and sciences. More than half of the
participants come from Germany, followed by the US, Switzerland, and France.
Most participants work for small and medium-sized companies with a headcount
of under 250 people and an annual turnover of less than USD 50m. This distri-
bution is mainly caused by a high number of participants from small consulting
companies. If considered on their own, around 60% of the participants from the
logistics services industry work for large companies with over 3.000 employees
and more than USD 500m turnover.
We then asked the participants about their company’s stance towards Blockchain.
Figure 4 summarizes the results. 43% declare they do not look into Blockchain just
yet or observe the development from a distance. 37% of the participants investi-
gate use cases, and 20% have implemented first Blockchain solutions. Again, this
distribution is caused by consulting companies as almost three-quarters of them
either investigate or implement Blockchain solutions. Looking at the logistics
services companies alone, around 65% declare to be – if at all – watching from a
distance. Only two experiment with Blockchain technology – one start-up and
one logistics services behemoth.
Next, we introduced the four use case exemplars. We provided the information in
Figure 2 and asked the participants to evaluate each use case regarding (1) the
benefit of Blockchain and (2) the likeliness of adopting Blockchain. Figure 5 sum-
marizes the findings through the mean values and standard deviations. Overall,
Blockchain is evaluated to offer considerable benefits for all use cases (between
11
12% 21% 35% 53%

Manu- Logistics Other Germany
facturing Services
16% 6%
Science Industry Retail Country
45% 13%
Consulting USA
34% 50% 31% 51%

>3,000 <250 >$500m <$50m
Company Company
Headcount Turnover
16% 18%
250 - $50m -
3,000 $500m
Figure 3: Overview of Participants and their Companies
20% 16%
We implement We don‘t look
Blockchain into that
solutions
37% 27%
We investigate We observe the
What is your company‘s possible use development
stance towards Blockchain? cases from a distance
Figure 4: Company’s Stance towards Blockchain
12
4 Survey about the Prospects of Blockchain in Logistics and SCM
1 Benefit of Blockchain 5.70

very small – 1 7 – very large
Likeliness of adoption 4.88
Ease Paperwork very unlikely – 1 7 – very likely
Processing

Identify Counterfeit very unlikely – 1 7 – very likely
Products

Facilitate Origin very unlikely – 1 7 – very likely
Tracking

Operate the very unlikely – 1 7 – very likely
Internet of Things
Figure 5: Evaluation of Use Cases
5.09 for the second and 5.70 for the first and fourth use case). In all instances, the
likeliness of adoption receives lower ratings than the benefit (between 4.48 for
the third and 5.22 for the fourth use case). However, the opinions are not uniform
across different groups of participants. Middle managers rate both benefit and
adoption likeliness lower than c-level executives do across all four use cases. Also,
participants more experienced in Blockchain applications (for example, through
own implementations) give better ratings than participants that only think about
Blockchain. Finally, the participants’ sector is also a differentiator. Especially
regarding paperwork processing though Blockchain applications, employees
from logistics service providers are more skeptical about the actual benefit than
consultants or scientists.
After evaluating the use cases, we inquired the participant’s opinion about the
main beneficiaries of Blockchain applications in logistics. The findings are shown
13
80% Logistics service providers

76% Senders
74% Receivers
74% Technology providers
41% Consultants
Who will benefit from Block-
31% Scientists
chain in the logistics context?
Figure 6: Beneficiaries of Blockchain
56% Regulatory uncertainty

50% Different parties have to join forces
49% Lack of technological maturity
49% Lack of acceptance in industry
41% Data security concerns
What are likely barriers for 40% Benefits are not clear
Blockchain adoption in the
28% Dependence on Blockchain operators
logistics industry?
Figure 7: Barriers to Blockchain Adoption
in Figure 6. Around three-quarters of the participants expect logistics service

providers, senders, receivers, and technology providers to benefit. Only 41%
see consultants and only 31% see scientists as being beneficiaries. Again, these
results are influenced by factors like hierarchical level or sector. Only one in five
middle managers believes consultants will benefit from introducing Blockchain,
while more than half of the c-level executives from the sample share this opin-
ion. Further, a disproportionately high share of employees from logistics service
providers believes technology providers will benefit, while consultants focus
more on supply chain actors like senders and receivers.
We asked the participants about likely barriers to Blockchain adoption in the

logistics industry. Figure 7 summarizes the findings. For 56% of the participants,
regulatory uncertainty might be a barrier. Around half of the participants also
refer to the fact that different parties must join forces, a lack of technological
14
5 Discussion and Implications
Blockchain will Blockchain will radi-

barely effect them 7.08 cally transform them
1 10
If Blockchain would be implemented industry-wide in logistics, what

would be the effect on established processes and business models?
Figure 8: Effect on Processes and Business Models
maturity, and a lack of acceptance in the industry as major barriers. Data security
concerns (41%), unclear benefits (40%), and too much dependence on Blockchain
operators (28%) are also named to be likely showstoppers. Again, the individual
opinions significantly depend on factors like the hierarchical level, the sector, and
the participants’ experience with Blockchain.
Finally, we inquired the participants’ overall evaluation of Blockchain’s effect on
established processes and business models in logistics. The result is shown in
Figure 8. With an average evaluation of 7.08, the participants believe Blockchain
would have a strong effect on the industry, even though the transformation might
not be as radical as some trade press articles herald. No big differences can
be found between any of the groups. The size of the standard deviation (1.88)
proves there are skeptics and enthusiasts in the sample. However, they cannot be
assigned to a specific group, like consultants or early adopters of Blockchain.
5 Discussion and Implications
Our study revealed valuable insights about Blockchain adoption in the logistics
and SCM context. Despite realizing the impact Blockchain might have on their
industry, companies seem hesitant to dedicate resources to look into possible
Blockchain applications. Apart from that, the findings also provide insights into
the perspectives of different participant groups. Taking the hierarchical level of
the participants into account, our data suggests middle managers are much less
enthusiastic about Blockchain than c-level executives or operational employees.
They give significantly lower ratings on Blockchain benefit and adoption likeliness
for the use cases, see fewer beneficiaries, and expect more showstoppers. For
example, 60% of the middle managers raise concerns about data security, while
15
only 28% of c-level executives share this reservation. An explanation for this lack
of enthusiasm might lie in the better overview middle managers have of their
processes. Since they are likely responsible for implementing new IT solutions
(at least from the business perspective), they might feel Blockchain is overhyped
and is just another IT development praised as the silver bullet.
Our data suggests significant differences between logisticians (working for logis-
tics service providers, manufacturers, and retailers), on one hand, and consultants
and scientists on the other. These differences are striking in terms of beneficiaries
and adoption barriers. Logisticians have difficulties getting a clear idea of the
benefits and use cases, while consultants and scientists worry about the techno-
logical maturity of Blockchain. Our results underline the importance of carving
out possible use cases for logistics and supply chain management. If people from
a rather conservative industry, like logistics, are expected to buy into new tech-
nology, the benefit must be very clear. Just because something is new doesn’t
mean logisticians get too excited.
A third differentiator is the level of Blockchain experiences. Our data suggests the
more experienced participants are (e.g., exploring use cases instead of just observ-
ing the development in the industry), the more positive they evaluate Blockchain.
We find steadily growing ratings of benefit and adoption likeliness across the four
experience levels shown in Figure 4. Also, more experienced participants identify
more beneficiaries. Their perspective on possible barriers shifts: While only one
fourth of the little-experienced participants expect collaborating with different
partners to be a showstopper, around 60% of participants having implementation
experiences indicate a high level of collaboration and commitment might be a bar-
rier. Our findings show small-scale experiments with Blockchain applications are
vital to understand the barriers and benefits of Blockchain. Logisticians should
engage in experiments to find out if and how Blockchain could be of use for their
own company.
6 Conclusion
In this paper, we presented a study on the current state of Blockchain in logistics

and SCM. To shed light on the first research question pertaining to possible appli-
cations of Blockchain in logistics and SCM, we introduced four use case exemplars
that are under exploration in theory and practice. To provide insights into the
second research question about Blockchain being a trick or a treat for logistics
16
References
and SCM, we presented the findings of a survey. The survey was conducted in the
logistics and SCM industry and investigated the participants’ opinion on use cases,
showstoppers, and benefits of Blockchain. As our findings show, Blockchain is
expected to have a considerable impact on the logistics industry and should be
considered a treat. However, our findings also show logisticians should start
“chewing” on Blockchain soon. They are well-advised to find out how much of
the overall Blockchain hype could be developed into a value-add for their service
portfolio before someone else does.
References
Abeyratne, S. A. and R. P. Monfared (2016). “Blockchain Ready Manufacturing Supply Chain Using
Distributed Ledger”. In: International Journal of Research in Engineering and Technology 5.9,
pp. 1–10.
Allison, I. (2017). Maersk and IBM Want 10 Million Shipping Containers on the Global Supply Blockchain
by Year-End.
Apte, S. and N. Petrovsky (2016). “Will Blockchain Technology Revolutionize Excipient Supply
Chain Management?” In: Journal of Excipients and Food Chemicals 7.3, pp. 76–78.
Casey, M. J. and P. Wong (2017). Global Supply Chains Are About to Get Better, Thanks to Blockchain.
Christidis, K. and M. Devetsikiotis (2016). “Blockchains and Smart Contracts for the Internet of
Things”. In: IEEE Access 4, pp. 2292–2303.
Chu, Y., J. Ream, and D. Schatsky (2016). CFO Insights: Getting Smart About Smart Contracts. De-
loitte.
DeCovny, S. (2017). Experts Discuss Tackling Pharma Supply Chain Issues With Blockchain.
Dickson, B. (2016). Blockchain Has the Potential to Revolutionize the Supply Chain.
Eastwood, G. (2017). Why Blockchain is the Future of IoT.
Gartner (2015). Gartner Says 6.4 Billion Connected ”Things” Will Be in Use in 2016, Up 30 Percent
From 2015. Tech. rep.
Groenfeldt, T. (2017). IBM And Maersk Apply Blockchain To Container Shipping.
Hackett, R. (2016). Walmart and IBM Are Partnering to Put Chinese Pork on a Blockchain.
Hackett, R. (2017). Wal-Mart Explores Blockchain for Delivery Drones.
Hair, J. F., W. C. Black, B. J. Babin, and R. E. Anderson (2009). Multivariate Data Analysis. Upper
Saddle River: Prentice Hall.
Kersten, W., M. Seiter, B. von See, N. Hackius, and T. Maurer (2017). Trends and Strategies in Logistics
and Supply Chain Management – Digital Transformation Opportunities. Hamburg: DVV Media
Group.
Lomas, N. (2015). Everledger Is Using Blockchain To Combat Fraud, Starting With Diamonds.
Mackey, T. K. and G. Nayyar (2017). “A Review of Existing and Emerging Digital Technologies to
Combat the Global Trade in Fake Medicines”. In: Expert Opinion on Drug Safety 16.5, pp. 587–602.
Meade, A. W. and S. B. Craig (2012). “Identifying Careless Responses in Survey Data”. In: Psycholog-
ical Methods 17.3, pp. 437–455.
17
Morabito, V. (2017). Business Innovation Through Blockchain. Cham: Springer International Pub-
lishing.
O’Marah, K. (2017). Blockchain For Supply Chain: Enormous Potential Down The Road.
Petersen, M., N. Hackius, and W. Kersten (2016). “Blockchains für Produktion und Logistik: Grund-
lagen, Potenziale und Anwendungsfälle”. In: Zeitschrift für wirtschaftlichen Fabrikbetrieb 111.10,
pp. 626–629.
Pilkington, M. (2016). “Blockchain Technology: Principles and Applications”. In: Research Handbook
on Digital Transformations. Ed. by F. X. Olleros and M. Zhegu. Edward Elgar Publishing, pp. 1–39.
Popper, N. and S. Lohr (2017). Blockchain: A Better Way to Track Pork Chops, Bonds, Bad Peanut
Butter?
Shaffer, E. (2017). Walmart, IBM Provide Blockchain Update.
Sutherland, W. J., P. Barnard, S. Broad, M. Clout, B. Connor, I. M. Côté, L. V. Dicks, H. Doran, A. C.
Entwistle, E. Fleishman, M. Fox, K. J. Gaston, D. W. Gibbons, Z. Jiang, B. Keim, F. A. Lickorish,
P. Markillie, K. A. Monk, J. W. Pearce-Higgins, L. S. Peck, J. Pretty, M. D. Spalding, F. H. Tonneijck,
B. C. Wintle, and N. Ockendon (2017). “A 2017 Horizon Scan of Emerging Issues for Global
Conservation and Biological Diversity”. In: Trends in Ecology and Evolution 32.1, pp. 31–40.
Tapscott, D. and A. Tapscott (2016). Blockchain Revolution. 1st ed. New York: Penguin Random
House.
Tian, F. (2016). “An Agri-food Supply Chain Traceability System for China Based on RFID & Blockchain
Technology”. In: 2016 13th International Conference on Service Systems and Service Management
(ICSSSM), pp. 1–6.
Underwood, S. (2016). “Blockchain Beyond Bitcoin”. In: Communications of the ACM 59.11, pp. 15–
17.
Xu, J. J. (2016). “Are Blockchains Immune to All Malicious Attacks?” In: Financial Innovation 2.25,
pp. 1–9.
Yli-Huumo, J., D. Ko, S. Choi, S. Park, and K. Smolander (2016). “Where Is Current Research on
Blockchain Technology? - A Systematic Review”. In: PLoS ONE 11.10, pp. 1–27.
Zhao, J. L., S. Fan, and J. Yan (2016). “Overview of Business Innovations and Research Opportunities
in Blockchain and Introduction to the Special Issue”. In: Financial Innovation 2.28, pp. 1–7.
Zheng, Z., S. Xie, H.-N. Dai, and H. Wang (2017). “Blockchain Challenges and Opportunities: A
Survey”. In: International Journal of Electric and Hybrid Vehicles, pp. 1–23.
18
A Peer-To-Peer Platform for
Decentralized Logistics
Olivier Gallay1 , Kari Korpela2 , Niemi Tapio1 , Jukka K. Nurminen3
1 – University of Lausanne
2 – Lappeenranta University of Technology
3 – VTT Technical Research Centre of Finland
We introduce a novel platform for decentralized logistics, the aim of which is to

magnify and accelerate the impact offered by the integration of the most recent
advances in Information and Communication Technologies (ICTs) to multi-modal
freight operations. The essence of our peer-to-peer (P2P) framework distributes
the management of the logistics operations to the multiple actors according to
their available computational resources. As a result, this new approach prevents
the dominant players from capturing the market, ensures equal opportunities
for different size actors, and avoids vendor lock-in. The latest ICTs such as In-
dustrial Data Space (IDS), Blockchain, and Internet-of-Things (IoT) are used as
basic building blocks which, together, enable the creation of a trusted and inte-
grated platform to manage logistics operations in a fully decentralized way. While
IDS technology allows for secured data exchange between the different parties
in the logistics chain, Blockchain technology handles transaction history and
agreements between parties in a decentralized way. IoT enables the gathering
of real-time data over the logistics network, which can be securely exchanged
between the different parties and used for managing the decision-making related
to the control of the freight transportation activities. The practicability and the
potential of the proposed platform is demonstrated with two use cases, involving
various actors in the logistics chains.
Keywords: Peer-to-Peer Logistics; Blockchain; Industrial Data Space; Internet-

of-Things
19
A Peer-To-Peer Platform for Decentralized Logistics
1 Introduction
Digitized logistics integration, beside drastically decreasing manual work, opens

the door for novel ways of optimizing the logistics processes and transport routes.
It also enables real-time monitoring of transportation flows in order to react
dynamically to unexpected circumstances. As a result, the integration of the
newest Information and Communication Technologies (ICT) offers great potential
to improve cost-effectiveness in logistics and enables new business models based
on real-time economics. Today, 70% of companies lack operational performance
data along the entire value chain, (World Economic Forum, 2016). To face this issue
within the European Union (EU), the long-term goals and the related roadmaps of
the European Technology Platform ALICE (c.f. etp-logistics.eu) aim at a 10%–30%
increase of efficiency in the EU logistics sector, which means EUR 100–300 billion
cost relief for the European industry. A true “people, planet and profit” oriented
logistics and supply chain sector contributes not only in industry competitiveness
but also in meeting environmental policy targets.
One of the main challenges in logistics is that various independent parties are
involved (freight forwarders, 3rd party logistic service providers, multimodal
transport operators, carriers, etc.). Communication between these parties needs
to be accurate, and the associated necessity is to create a common platform where
all parties share the necessary relevant information. Several attempts towards
the implementation of such a platform were carried out in several past EU-funded
research projects (such as CONTAIN, FREIGHTWISE, e-FREIGHT, or iCargo, c.f.
containproject.com, freightwise.tec-hh.net, eutravelproject.eu), and recently by
commercial products such as Uber Freight. The problem when adopting such
approaches is that some organization needs to run the common platform. While
having platform dominance often represents a good business opportunity for
the company running the platform, other actors have to pay fees and follow the
whims of the platform owner. Furthermore, a single gatekeeper poses a risk for
diversity and fairness in the market. Multiple competing platform companies, on
the other hand, create the risk of a splintered market resulting into non-optimal
logistics decisions in separate silos.
An alternative solution – and the focus of this contribution – is to handle logistics

management in a peer-to-peer (P2P) fashion. While P2P logistics is sometimes
understood as crowd-sourced transportation, we concentrate here on the con-
ventional transportation industry. The essence of our P2P approach is that the
logistics management system is fully distributed to the computational resources
20
2 Related Work
of the actors. The outcome is a distributed system with a network of interactions

between parties. The proposed framework aims at using and integrating two
recent P2P technologies: Industrial Data Space (IDS), developed at Fraunhofer
Institute in Germany, and Blockchain, which was initially a key part of cryptocur-
rencies (Mukhopadhyay et al., 2016), but is now a separate technology on its
own.
The proposed P2P approach blocks dominant players from capturing the market,
ensures equal opportunities for different size actors, and avoids vendor lock-in.
The essence of the introduced P2P platform is that the cloud-enabled transport
management system is distributed over multiple actors according to their com-
putational capabilities. The result is a distributed system with a network of inter-
actions between actors and entities, which enables smart and innovative data
based services.
Ultimately, we present a framework that goes beyond the paradigm of a central-
ized platform, and we propose a fully P2P architecture for logistics management.
The platform combines three novel technologies: IDS, Blockchain, and IoT. All
these technologies have been tested and used in their specific fields but never
been combined in the present way. Therefore, the proposed system provides the
logistics field with a totally new management approach, and open possibilities
for new innovations and businesses utilizing this platform.
2 Related Work
The key question raised in this paper is how can a logistics management system
based on P2P ideas help and provide new value in the digitized supply chain, and
solve the contradiction between interoperability and data sovereignty. While the
technical components we are introducing in the present contribution have already
been investigated in various contexts including logistics, they have never been
applied in an integrated fashion to logistics as an enabler for smart and innovative
data based services. Some companies are already marketing P2P logistics services
(e.g. p2plogistics.co.uk), however it is important to note that they are simply using
a centralized client-server architecture platform to match the transport needs
and crowd-sourced providers. Ideas towards a P2P approach have first been
envisioned on a general level in a Hewlett-Packard patent (Chen and Hsu, 2007),
but application of such ideas to logistics has not been systematically studied nor
experimented so far.
21
A recent survey by Sternberg and Andersson (2014) gives an exhaustive overview

of the existing scientific research contributions on distributed freight manage-
ment. In particular, the authors observe that “decentralized intelligence in lo-
gistics can be viewed as a disruptive architectural innovation operating on an
IT-infrastructure level, displaying massive network effects that has not yet ma-
terialized”. Sternberg and Andersson see the problem of reaching the critical
mass of adopters as a key challenge, and highlight the importance of industrial
experiments.
Industrial Data Space (IDS) reference architecture and technology have been
created by the Fraunhofer Institute and the Industrial Data Space Association (c.f.
industrialdataspace.org). The association has currently 50+ members ranging
from manufacturing, production, logistics, ICT to services, such as DB Schenker,
ThyssenKrupp, Schaeffler, FESTO, Bosch, SICK, Salzgitter and SMEs like Setlog
and Quinscape. Experiments on IDS technology are ongoing in 10+ European
IDS hubs. The IDS Connector software component (Otto, 2017), which allows
for secured data sharing and processing, is currently standardized in a fast track
procedure by DIN, the leading German standardization institution, and by ISO in
parallel. IDS is recognized as a key enabler for future industry platforms by the
EU’s Digitising European Industry initiative (European Commission, 2016; Jarke,
2017; Jesse, 2017).
Yuan et al. (2016) claim in a recent contribution that Blockchain can revolutionize
intelligent transport systems, and as a result the use of Blockchain in various in-
dustrial and logistics applications is currently widely investigated. Blockchain has
already been applied to logistics in several R&D projects and academic research.
For instance, Apte and Petrovsky (2016) discuss the implementation of Blockchain
within pharmaceutical supply chains, and Zhang et al. (2017) combine Internet-
of-Things (IoT) and Blockchain in order to propose a novel business model for
IoT services. Tian (2016) combines RFID and Blockchain for building a traceable
agri-food supply chain. A recent contribution by Korpela, Hallikas, and Dahlberg
(2017) studies the challenges raised by the integration of Blockchain in supply
chain frameworks. Blockchain design functionalities are shown to support good
integration for ledger and smart contracting, however for a global integration
within a supply chain context, a common standardized data model is additionally
required to secure the interoperability, which is performed by the IDS technology
in the platform proposed in this contribution. Blockchain technology and various
associated industrial business cases are currently analyzed in a national project
in Finland (BOND - Blockchains bOosting FiNnish InDustry, c.f. vtt.fi/sites/BOND).
A Scandinavian consortium (DBE Core, c.f. dbecore.com/portfolio/what-next),
22
3 A smart platform for decentralized value chain operations
including IBM, is working on the integration of Blockchain open logistics APIs, as

well as on associated new business models for open transportation ecosystems.
SmartLog (c.f. kinno.fi/en/smartlog), a EU-funded research project, is currently
investigating the first Proof-of-Concept implementation for a solution involving
Blockchain and IoT in the logistics industry.
The business opportunities raised by the appearance of IoT have been acknowl-
edged by Gartner to represent a major market prospect in the next decade (Rivera,
2014). Papert and Pflaum (2017) give recommendations on how IoT could be
implemented by logistics companies. One of their key findings is that the IoT plat-
form should be based on open interfaces and open source software to support
communication among different applications. There exist several recent studies
on various ways to apply IoT to logistics, transportation, and vehicles (Da Xu, He,
and Li, 2014; Guerrero-Ibanez, Zeadally, and Contreras-Castillo, 2015; Guo et al.,
2017; Contreras, Zeadally, and Guerrero-Ibanez, 2017).
In terms of Technology Readiness Level (TRL), the Blockchain technology is al-
ready in massive use in digital currencies, but for other use cases it is still under
development (TRL5). IDS Connector technology is used in industrial pilots (TRL5).
Positioning and other IoT sensing technologies have been in production use for a
long time (TRL9). The integration of IDS, Blockchain, and IoT has not been fully
demonstrated in industrial use (TRL5) for logistics system management.
3 A smart platform for decentralized value chain

operations
3.1 Concept
The ultimate objective of this contribution is to take advantage of novel P2P

technologies, in order to enable the creation of a trusted, decentralized logistics
platform. Figure 1 gives a general sketch of the proposed platform. IDS technol-
ogy allows for data exchange between the different actors in the logistics chain.
Its key feature is data governance, in the sense that it enables parties to work
together without releasing confidential information to each other. IDS Connec-
tors can implement in a decentralized fashion any type of taylored algorithms,
e.g. for calculating transportation options matching the customers’ needs, or
for merging procurement systems. Since these algorithms are embedded inside
23
Figure 1: High-level system architecture with example data flows.
IDS Connectors, proprietary business logic does not need to be disclosed. In a

similar way, the customer company can work privately with the IDS Connector
using sensitive data without disclosing it to anyone. Trust is further enhanced by
the option to have IDS Connectors certified by 3rd parties.
Like IDS, Blockchain technology works in a P2P fashion, managing transaction
history and verifying agreements between parties in a decentralized way. The
distributed trust, which Blockchain enables, is a perfect match to the distributed
IDS technology. Indeed, Blockchain technology allows for the establishment
of Smart Contracts, which, once agreed, can be run by autonomous programs
without any human intervention.
IoT is the third key building block needed in order to allow timely information in
a dynamic environment (Al-Fuqaha et al., 2015). This consists in a large collection
of networked sensors or other devices communicating with each other and de-
livering data to higher level applications using a suitable middleware (Razzaque
et al., 2016). The data collected by IoT sensors can be stored, processed, and
shared by applying cloud-computing (Botta et al., 2016).
In Figure 1, IDS Connectors represent containers (or other transport units), vehi-
cles, and transport companies. The meta-connector is a virtual entity searching,
optimizing, and managing end-to-end transport routes and processes, which
24
can span multiple means of transportation and multiple transport companies. A

brokering service maintains a directory of available IDS Connectors. IDS allows for
standardization of the data transfers and the data formats used. As a result, IDS
provides all actors of the logistics chain with an easy access to the platform, and
to the services it offers. IDS Connectors communicate with each other, and with
the IDS broker. For instance, when a company needs a transport service, it uses a
dedicated meta-connector, which performs route optimization by combining the
best transport means and service providers to form a single transport chain.
The transport unit IDS Connector also collects and manages information from IoT
sensors installed in the containers or in other transport units. The essential events
in IoT data, such as crossing national borders or sensory readings of changed
freight conditions (e.g. location, temperature, humidity level, etc.), can also be
stored into the Blockchain. In this way, e.g. automatic payment can be autho-
rized after the transport unit has been delivered to the final customer and the
merchandises in it have been stored in agreed conditions (e.g. temperature). The
transport unit IDS Connector can also offer real-time information on the location
and other sensor data to the company that ordered the transport service.
Let us now focus on one of the most natural process to be handled, where optimal
transport routes have to be negotiated. Thanks to the present P2P platform, a
multi-modal routing algorithm can be implemented within the IDS Connectors in
order to determine the best transportation means and schedule for each freight
request arising in the logistics system. In addition to monetary objectives, the
algorithms could use, e.g. environmental factors as an optimization goal. Since
any user can create new IDS Connectors, adding new specific routing algorithms is
possible, depending on the customers’ needs. The route planning of the transport
units might typically consist of:
— An offline algorithm, which first determines an ideal multi-modal routing
schedule, based on the data available at the time of the transportation
request.
— The initial schedule may thereafter be modified and adapted via the
execution of an online algorithm, which determines dynamic recourse
actions in case of perturbations to the original schedule (delays, disrup-
tions).
For the latter online algorithm above, there are many events which have an impact
on the time of arrival of a transport unit: traffic perturbations, rest periods of the
driver, as well as breakdowns. For these reasons, IoT is implemented in order to
25
constantly track freight units and vehicles along their journey. Real-time data
is collected from the numerous IoT sensors deployed over the logistics network
as well as from the freight units and vehicles measured via GPS-tracking. Open
data, including real-time traffic information, can also be considered by the online
routing module. The estimated time of arrival can be automatically calculated
with the help of GPS coordinates that are sent automatically to the IDS Connector.
Time consuming events can also be manually sent (e.g. in the GS1 EDI XML format)
to the customer via the IDS platform, which will re-optimize the route and re-
calculate the estimated time of arrival. The driver himself can negotiate new
time slots through a smartphone-based platform. Hence, the online algorithm
is able to help in bringing more reliability to the logistics network and allows for
a substantial reduction of transportation times. When it comes to multi-modal
transportation networks, it allows to react to random perturbations during the
shipment by dynamically and automatically managing mode changes using the
IDS Connectors, and to manage the related contracting with the involved carriers
by using the Blockchain feature. Through the IDS technology, the routing module
can provide users emitting a freight request not only with the best option (in terms
of transportation costs, route, and combination of modes), but also with possible
routing alternatives and associated costs. This allows the users to autonomously
plan on the online platform their freight request based on their particular needs.
Blockchain technology is used to confirm and store contracts negotiated by IDS

Connectors. Also, transactions associated with the routing decision-making pro-
cesses, which involve the selection of various transport carriers and operators,
are recorded by the IDS platform on the Blockchain to ensure future traceabil-
ity, visibility, and transparency. As Smart Contracts are shared programs inside
Blockchains, they do not need any 3rd party validators, and can be executed
automatically, when all definitions in the contract have been fulfilled, e.g. when
IoT readings satisfy agreed conditions.
Such a decentralized logistics platform will have to be first implemented in smaller

use cases, as the ones presented in Section 4 below. This will allow hence to in-
vestigate in parallel the role and potential of new service and product innovations
by identifying service gaps. This is done by analyzing how companies and the
business ecosystem increase revenues and cut costs by using real-time business
processes, and identifying opportunities for cost-cutting through the elimination
of manual work.
26
3.2 Impact
The essence of the present decentralized P2P platform is to show the potential
for high interoperability through distributed Blockchain and IDS technologies. In
the following, we discuss the benefits of the proposed platform, grouped by the
enabling technologies.
Peer-To-Peer
The proposed platform, based on P2P technologies, allows for a flexible integra-
tion of ICT and operational processes of the different involved parties. It offers
a complete real-time mapping between physical freight and digital information
flows. Applying P2P technologies and cloud-based service architecture ensures
easy access to the system with user devices ranging from sensor platforms, mo-
bile phones/tablets, and PC computers. The P2P concept provides hence the
different supply chain partners with equal opportunities, no matter their size
and importance. Another key benefit of the P2P approach is scalability, as the
workload is divided, and there is hence no bottleneck in the system.
IDS and Data Sharing
The data sovereignty of IDS technology has been designed in order to enable
data exchange while keeping control of the data. Horizontal collaboration is
implemented via the effective information and data exchange that is achieved
through the IDS Connector technology. As only the IDS Connector APIs are fixed,
new ICT players can start introducing new functionalities by creating their own IDS
Connectors. The system is therefore open to both logistics and ICT parties to easily
join in. The APIs enable the integration of IDS Connectors to company internal
data systems, and thus to their specific operational processes. IDS Connectors
ensure data privacy and confidentiality of sensitive business data.
Blockchain
Blockchain technology provides the different supply chain partners with the
technical basis for trust while the data governance mechanisms of IDS will handle
information sharing. As a result, Blockchain technology is used to build a reliable
27
general electronic bookkeeping solution, as the Blockchain approach is highly

successful in guaranteeing the truthfulness of the data.
Internet-Of-Things
IoT technologies are used to monitor the location and state of freight units and
vehicles. As a result, real-time data, such as location and other IoT sensor data,
can be used in managing the logistics decision-making, which is directly used
to control the freight transportation activities. In this way, taylored real-time
optimization algorithms can be implemented in order to allow for reacting dy-
namically to deviations. Ease of deployment is a key design goal for IoT data
collection mechanisms.
Multi-Modal Routing Algorithms
The proposed platform allows for the development of novel static and dynamic
algorithms for optimal multi-modal transport decisions. Indeed, the above ICT
technologies facilitate the data and information sharing between the different
parties, and hence using them opens the door for smarter decision-making. More
efficient logistics will focus on shortening transportation routes, increasing ve-
hicle load factor, reducing empty runs by improved route planning, as well as
shortening the lead time and consequently reducing inventories. This will result
in creating savings in transportation costs. The resulting financial savings arise as
a result of both better combinations of loads and higher load factors, and better
selections among multiple routes and the specific advantages offered by the dif-
ferent transportation modes (flexibility of trucks, smaller environmental footprint
of railways, lower costs of maritime transport, etc.), which helps to shorten the
freight delivery routes, as well as to reduce CO2 emissions. All these aspects
will hence also have a positive impact on the ecological footprint of the logistics
system.
As highlighted in (Korpela, Hallikas, and Dahlberg, 2017), a combined stack of
standardized technologies is a major enabler for the accelerated development
of a many to many integration model. In general, the architecture of Blockchain
technology (design elements and functionalities) is commonly shared, and such
standardization has facilitated its integrated use over the Internet. IoT integration
is implemented in our platform with the use of standardized IDS Connectors.
28
4 Towards implementation and acceptance
4 Towards implementation and acceptance
Two initial use cases, which emphasize the different aspects of the proposed
concept, have been identified with various industrial partners. While each of
these two use cases requires some specific implementation work, they both
share the key technology platform introduced in this contribution. The concrete
aim of such initial uses cases is to highlight at a smaller-scale, and hence at a
reduced investment risk, how the logistics platform based on IDS and Blockchain
technologies is able to improve the efficiency of logistics operations.
4.1 Secure tracking of deliveries
This use case originates from a company that delivers turnkey solutions for in-
dustrial machine and facility builders. The customer base runs 14’000 turbines
for power generation, 700 paper machines and about 1 million power transmis-
sion systems in 70 different countries. Hundred sites are handling the delivery of
these customized solutions to the customers, using dozens of different transport
companies. Besides the shipment of heavy machinery parts, which have to travel
thousands of kilometers cross-borders, a solution includes thousands of other vi-
tal components that are needed to complete the delivery. Collecting information
during the in-transit phase is vital to be able to better predict the time of arrival
of the components, and help planning to ensure on-time delivery with reduced
lead-time.
This use case possesses all the facets needed to study how to facilitate on-demand
based operations in a globally distributed project context, and how new tech-
nologies related to P2P, talking containers (IoT), and dynamic routing can lower
the costs while improving performance in logistics related to large projects. In
particular:
— Routing: on-time delivery, real-time estimation of time of arrival, and

negotiation of a suitable delivery time. Both the shipper and the customer
can better organize their delivery processes and avoid congestion on
delivery platforms. Thanks to improved real-time routing information
enabled by the gathering of real-time IoT data and IDS-based routing
algorithms (from the carrier to the logistics service provider and the
customer), the installation site can prepare its processes for on-time or
29
late delivery. In case of perturbations, real-time re-routing allows to

minimize the lateness.
— Tracking: all parcels and containers are able to communicate their lo-
cation and condition measured through different IoT sensors that are
linked with the package and the components themselves.
— Security: all information throughout the delivery process is stored in a

secured Blockchain for multiple purposes. During the delivery phase,
the location and transportation circumstances of the containers are to
be recorded and stored securely. Once assembled, the operational his-
tory can be stored in the same Blockchain to cover the whole lifecycle
including the recycling procedure at the end of lifetime.
— Integration: existing transport management systems and mobile apps

can be integrated in order to ensure the ability of the logistics platform
to work together with legacy solutions and with standardized formats.
4.2 Blockchain and data integration for forest industry logistics
The Digital Business Ecosystem Consortium (DBE Core) drives the implementation
and adoption of logistics API integration to Blockchain and Smart Contracts for
end-to-end digital information exchange. The members of the DBE Core consist
of industry companies in two large business domains: the bio-refinery process
and the maritime ship building industry. They form a customer base for logistics
services in land operations across Finland, as well as cross-border operations to
global markets. The DBE Core companies are operating over 100 countries.
As a case example in Finland, the Lappeenranta region has one of the largest forest
industry clusters in the world. Every day more than 1000 full truckloads of raw
material is received in three regional bio-refinery industry locations, and some 300
truckloads per day are delivered to European customers. A new truck is unloaded
every 3 minutes. The import from Russia (which crosses EU borders) is over 15
million cubic meters of wood. The distributed stocks in the forest are geo-located,
and hundreds of Small and Medium Enterprises (SME) logistics companies handle
their transportation on a daily basis.
Logistics operations in different industry locations, supplier locations, ware-

houses, ports and railroads require a well-conducted and interoperable logistics
30
5 Conclusions and future work
management system. Logistics companies are generally using intermediate op-

erators to transfer and map the data between organizations and their systems.
Currently, lots of manual work is required to integrate the information, which
generates high costs. Therefore SME suppliers are exposed to be replaced by
Tier-2 or Tier-3 suppliers.
All stakeholders of DBE Core have decided jointly to move towards a common
logistic integration based on open Application Programming Interfaces (APIs).
The development is based on global business process standards (OASIS/UBL).
The development and piloting of Blockchain integration and Smart Contracting
can be established by the following steps: (a) defining DBE Core consortium
Blockchain (e.g. with IBM Bluemix and HyperLedger), (b) Logistic API integration to
Blockchain, (c) Smart Contract integration to control and automate transactions,
and (d) secure interoperability with IDS Connectors.
The P2P platform described in this paper will support the knowledge creation
and sharing of concepts and tools to reach the DBE Core visions and goals in the
following ways:
— Real-time visibility to logistics transportation: monitoring and control
of logistics processes enables the development of demand supply fore-
casting algorithms based on pull model for delivery control, in order to
reduce lead-time, waiting times, and manual work.
— Securing the data storing and sharing by Blockchain on B2B and IoT
transactions.
— Automating the supply and logistics integrations in gross organizational
environment.
5 Conclusions and future work
We present a fully decentralized platform for logistics management based on IDS

and Blockchain technologies which, combined with the gathering of real-time IoT
data, enables to address efficiently the challenges raised in the logistics networks
nowadays. The proposed P2P framework allows for supporting interoperability
between different actors in the logistic chains, sharing securely the data between
the different parties, and using it for optimal planning based on IoT data.
31
The platform allows for an implementation that is not affected by scalability is-
sues, and that is not limited by geographical borders. In particular, facilitation of
cross-border transportation is one of the main issues that the platform facilitates.
By lowering the threshold for new actors to join in, it is expected that new oppor-
tunities will arise both in dynamic transportation optimization, and in 3rd party
development of new components and services to the platform.
Interestingly, the overall concept discussed in this paper also has applicability
beyond logistics systems. There are a number of areas which share similar key
characteristics: (i) cooperation and data transfer between multiple independent
parties is essential, (ii) splitting the market into separate silos run by different
companies reduces the benefits, and (iii) a dominant platform company can have
an unhealthy role in controlling the market. Examples of such services include
auction sites, travel rental booking sites, and most importantly social media sites.
In these contexts, we typically observe strong network effects where the value
of the network rises with the number of participants it is able to attract. As a
consequence, the dominant player has a major advantage and is able to push the
smaller players away from the market, resulting into monopolistic positions (for
an overview see e.g. Haucap and Heimeshoff, 2014). While the shared economy
helps smaller players to enter the market to perform the actual work (Einav, Far-
ronato, and Levin, 2016), the competition at the more lucrative management layer
tends to fade. Being able to provide the similar service in a P2P fashion repre-
sents a way to gain the benefits of network effects without facing the problems of
dominant, monopolistic parties. For instance, Munsing, Mather, and Moura (2017)
have explored Blockchain technology in connection with electricity networks to
eliminate the need of a microgrid operator.
It is important to note that towards the implementation of the proposed platform,
business model implications will have to be investigated in parallel, in order to
unveil the associated new innovation opportunities and ecosystem effects, as well
as the additional needs for development, standardization, and other activities,
which are necessary to bring the enhanced logistics solutions into operational
use.
References
Apte, S. and N. Petrovsky (2016). “Will blockchain technology revolutionize excipient supply chain
management?” In: Journal of Excipients and Food Chemicals 7.3, pp. 76–78.
32
References
Botta, A., W. De Donato, V. Persico, and A. Pescapé (2016). “Integration of cloud computing and
internet of things: A survey”. In: Future Generation Computer Systems 56, pp. 684–700.
Chen, Q. and M. Hsu (2007). Peer-to-peer inter-enterprise collaborative process management
method and system. US Patent US 09/823,581.
Contreras, J., S. Zeadally, and J. A. Guerrero-Ibanez (2017). “Internet of vehicles: Architecture,
protocols, and security”. In: IEEE Internet of Things Journal.
Da Xu, L., W. He, and S. Li (2014). “Internet of things in industries: A survey”. In: IEEE Transactions
on Industrial Informatics 10.4, pp. 2233–2243.
Einav, L., C. Farronato, and J. Levin (2016). “Peer-to-peer markets”. In: Annual Review of Economics
8, pp. 615–635.
European Commission (2016). “Digitising european industry: Reaping the full benefits of a digital
single market”. In:
Al-Fuqaha, A., M. Guizani, M. Mohammadi, M. Aledhari, and M. Ayyash (2015). “Internet of things: A
survey on enabling technologies, protocols, and applications”. In: IEEE Communications Surveys
& Tutorials 17.4, pp. 2347–2376.
Guerrero-Ibanez, J. A., S. Zeadally, and J. Contreras-Castillo (2015). “Integration challenges of
intelligent transportation systems with connected vehicle, cloud computing, and internet of
things technologies”. In: IEEE Wireless Communications 22.6, pp. 122–128.
Guo, L., M. Dong, K. Ota, Q. Li, T. Ye, J. Wu, and J. Li (2017). “A secure mechanism for big data
collection in large scale internet of vehicle”. In: IEEE Internet of Things Journal 4.2, pp. 601–610.
Haucap, J. and U. Heimeshoff (2014). “Google, Facebook, Amazon, eBay: Is the Internet driving
competition or market monopolization?” In: International Economics and Economic Policy 11.1,
pp. 49–61.
Jarke, M. (2017). “Data spaces: Combining goal-driven and data-driven approaches in commu-
nity decision and negotiation support”. In: International Conference on Group Decision and
Negotiation. Springer, pp. 3–14.
Jesse, N. (2017). “The future of innovation with third party data”. In: https://www.talend.com/blog/
2017/07/18/building-data-sharehouse-agile-data-management-industrial-data-space-ids.
Korpela, K., J. Hallikas, and T. Dahlberg (2017). “Digital supply chain transformation toward
blockchain integration”. In: Proceedings of the 50th Hawaii International Conference on System
Sciences.
Mukhopadhyay, U., A. Skjellum, O. Hambolu, J. Oakley, L. Yu, and R. Brooks (2016). “A brief survey
of cryptocurrency systems”. In: 14th Annual Conference on Privacy, Security and Trust (PST), 2016.
IEEE, pp. 745–752.
Munsing, E., J. Mather, and S. Moura (2017). “Blockchains for decentralized optimization of energy
resources in microgrid networks”. In:
Otto, B. (2017). “Reference architecture model for the industrial data space”. In: http://industrialdataspace.
org/en/publications/industrial-data-space-reference-architecture-model-2017.
Papert, M. and A. Pflaum (2017). “Development of an ecosystem model for the realization of
Internet-of-Things (IoT) services in supply chain management”. In: Electronic Markets 27.2,
pp. 175–189.
Razzaque, M. A., M. Milojevic-Jevric, A. Palade, and S. Clarke (2016). “Middleware for internet of
things: A survey”. In: IEEE Internet of Things Journal 3.1, pp. 70–95.
Rivera, J. (2014). “Gartner says the internet of things will transform the data center”. In: http :
//www.gartner.com/newsroom/id/2684616.
33
Sternberg, H. and M. Andersson (2014). “Decentralized intelligence in freight transport: A critical

review”. In: Computers in Industry 65.2, pp. 306–313.
Tian, F. (2016). “An agri-food supply chain traceability system for China based on RFID & blockchain
technology”. In: 13th International Conference on Service Systems and Service Management
(ICSSSM), 2016. IEEE, pp. 1–6.
World Economic Forum (2016). Digital Transformation of Industries: Logistics Industry. Tech. rep.
Yuan, Y. and F.-Y. Wang (2016). “Towards blockchain-based intelligent transportation systems”.
In: IEEE 19th International Conference onIntelligent Transportation Systems (ITSC), 2016. IEEE,
pp. 2663–2668.
Zhang, Y. and J. Wen (2017). “The IoT electric business model: Using blockchain technology for
the internet of things”. In: Peer-to-Peer Networking and Applications 10.4, pp. 983–994.
34
IoT-coordinated Logistics in
Product-Service Systems
Jürgen Anke1
1 – Hochschule für Telekommunikation Leipzig
Product-Service Systems (PSS) enable new value to customers compared to tradi-

tional selling of goods, e.g. through new business models such as Performance-
based contracting (PBC) and Pay-per-use (PPU). This turns the manufacturer
into a service provider, which has to perform maintenance, repair and operation
(MRO) tasks for the products provided in a PSS. Therefore, the service provider
must organize the logistics of the equipment, spare parts, and consumables in a
timely manner. In this paper, we evaluate the impact of Internet of Things (IoT)
technology for the support of logistics-related tasks in PSS. More specifically, the
required tasks with relevance of logistics are identified using case studies and
existing literature. For each task, we derive trigger events and how they can be
discovered automatically. Based on this, we propose a generalized architecture
for IoT-enabled logistics processes in PSS, which can be used to better understand
the operation of such systems and support their design.
Keywords: product-service systems; cyber-physical systems; Internet of Things;

Performance-based contracting
35
IoT-coordinated Logistics in Product-Service Systems
1 Introduction
The proliferation of Internet of Things (IoT) technologies and ongoing servitization

efforts in manufacturing companies break ground for new business models, such
as pay-per-use (Weinberger, Bilgeri and Fleisch, 2016). One of the most frequently
cited example for this principle is the “power-by-the-hour” concept of Rolls-Royce
for aircraft engines (Neely, 2008). The concept of combining goods and services
into an integrated offer is called Product-Service System (PSS). In more elaborate
usage-based or results-based PSS, the task of managing and operating equipment
is shifted from the user to provider (Elmazoski, et al., 2016). While this reduces the
risk, required knowledge and maintenance resources at the customer, it causes
uncertainty about the overall cost as it is usage-dependent. For the provider,
which takes over the responsibility for maintaining the equipment, this model
offers continuous revenue through long-time contracts, customer lock in and
other benefits (Lockett et al., 2011).
The main aspect of PSS in regard to logistics is that products are not being sold
but instead provided as a service or as part of service (Cavalieri and Pezzotta,
2012). Therefore, the provider strives to keep the contracted equipment running
for two reasons: First, the revenue depends on the usage, and second, any penalty
payments due to service level agreement (SLA) violations have to be avoided.
To increase usage fees and prevent penalty payments for SLA violations with
PBC, the service provider aims to perform all required activities to maintain high
availability at the lowest possible cost. The employment of IoT-technology for
products in PSS offers large efficiency potentials in their operation. For example,
spare parts, consumables, and technicians need to be provided in timely manner
to fulfill the agreed upon SLAs. Consequently, the physical processes should
be coordinated efficiently through the IoT-enabled products in PSS. The major
challenge is to employ IT in a way that provides high availability for the user and
efficient servicing for the provider at the same time. However, the link between
the concept of PSS and underlying technology for it has just recently emerged in
literature (Ardolino et al., 2017; Sala et al., 2017).
In this paper, the role of information technology for the coordination of logistics
processes for PSS is investigated. Using several case studies as well as existing
literature, we identify information demand and data sources required for the
logistics tasks as part of the overall equipment servicing. Based on that, a general-
ized architecture for a cyber-physical system (CPS) is proposed, which integrates
information flows required to coordinate material flows for products equipped
36
2 Conceptual Foundation
with IoT-technology. From a technology perspective, IoT-data and events from

the product are processed with analytics in the cloud to trigger the execution of
business processes in the transaction systems at the service provider.
Our results are to help practitioners to design and manage PSS based on CPS.
They could also serve as a basis for services fourth-party logistics providers (4PLs),
which can then be contracted as part of the overall PBC-PSS service system by
the service provider. As a theoretical contribution, we aim to improve the con-
ceptual basis for smart service systems at the intersection of service science,
cyber-physical systems, logistics and big data (Demirkan et al., 2015; Medina-
Borja, 2015).
2 Conceptual Foundation
The key concepts of this research are product-service systems, business models,
IoT and cyber-physical systems, which are briefly introduced in this section.
The combination of tangible physical goods and intangible services into an inte-
grated offer is called a product-service system (PSS) (Mont, 2002; Baines, et al.,
2007). PSS take a marketing oriented perspective as they describe what com-
panies offer and what value customers can gain from these offers. According
to a popular classification proposed by Tukker (2004), PSS can be subdivided
into product-oriented, use-oriented and result-oriented PSS. The first category in-
volves the selling of product and the offer of accompanying services. Use-oriented
services still focus on the product but the product remains in the ownership of
the provider and it is billed on a per use basis. In result-oriented PSS, the con-
sumer and the provider agree on a defined result without the involvement of
any pre-determined product (Tukker, 2004). A trend contributing to the emer-
gence of PSS is the so-called “servitization”, which refers to the transformation of
product-oriented offers into services (Neely, 2007).
In contrast to the selling of goods, PSS allow for business models, in which the
product remains property of the service provider, while the user pays for the
actual service. Typical variants are pay-per-use and performance-based contract-
ing. In pay-per-use (PPU) models, the user is charged for the metered usage
of a good, e.g. by the hour as in car sharing services like Car2Go (Gassmann,
Frankenberger and Csik, 2014). Performance-based contracting (PBC) refers to
the “the contractual approach of tying at least a portion of supplier payment to
37
performance” (Selviaridis and Wynstra, 2014). In their survey of 241 papers on

PBC, Selviaridis and Wynstra (2014) recognize “performance specification and
evaluation, payment scheme design and its impact on supplier behavior and, to
a lesser extent, risk allocation and attitudes” as the main aspects in PBC research.
A comprehensive review of PBC from a business perspective is provided by Essig
et al. (2016).
The integration of Internet of Things technology into physical products has been
identified as basis for innovative business models (Allmendinger and Lombreglia,
2005; Herterich, Uebernickel and Brenner, 2015). This particularly holds true for
PBC and PPU a, which are both identified as business models benefitting from IoT,
especially from the “Remote Usage and Condition Monitoring” component in the
“digitally charged products” pattern (Fleisch, Weinberger and Wortmann, 2015).
The “IoT-enabling” of products constitutes of them being equipped with sensors,
embedded computation and communication capabilities to turn them into smart
products (Porter and Heppelmann, 2014; Valencia et al., 2015; Barbosa et al., 2016).
The remote access to the product status, operation conditions etc. for monitoring
and control provide the basis for data-driven product-related services (Herterich,
Brenner and Uebernickel, 2016; Kagermann, Riemensperger and Weckesser, 2016).
To support the servicing of PSS, it is therefore key to integrate IoT devices and
the events they generate with business processes, which has emerged as topic in
the recent years, e.g. by Schief et al. (2011) and Meyer, Ruppen and Magerkurth
(2013).
The theoretical foundation for these systems are Cyber-physical systems (CPS),
denoting the integration of computing with physical processes, which can affect
each other (Lee, 2008). In a more comprehensive understanding, CPS are socio-
technical systems, consisting of sensors, actuators, embedded systems, digital
networks, Internet services as well as coordination and management processes
(Broy, Cengarle and Geisberger, 2012). CPS can be considered as the technical
implementation of a PSS with smart, connected products (Sala et al., 2017).
In summary, within this research, we assume that all PBC and PPU business
models for technical products create a PSS. Furthermore, we consider CPS as the
conceptual basis for the implementation of such PSS with IoT-enabled technical
products, as CPS integrate physical goods, processes, people and information
technology into a holistic system.
38
3 Research Goal and Methodology
3 Research Goal and Methodology
The research goal of this paper is to identify how logistics processes in PSS can be
coordinated through IoT-enabled products. For that, the following methodology
is applied: Based on existing literature and several real-world cases, main logistics
tasks are identified. These tasks are consolidated into a framework, which relates
information demands, data sources and analytics to these logistics tasks. From
this, a generalized CPS architecture is derived to describe the coordination of
logistics processes in IoT-enabled PSS.
It is important to note that there is a variety of options to configure PSS offerings
(Freund and Stölzle, 2016). As Essig et al. (2016) point out, PBC is related to both
industrial marketing and operations and service management (OSM). For our
research, we focus on the OSM part and PSS with the following characteristics:
— The product part of the PSS is a technical product
— The PSS is offered in a pay-per-use or performance-based contracting
model
— Maintenance, Repair and Operations (MRO) are at least partly contained
in the offer
4 Related work
While the conceptual foundation addresses the basic concepts of this research,
the related work section presents recent work at the intersection between PSS,
CPS and PBC/PPU business models.
The potentials and challenges of the application of performance-based contract-
ing in manufacturing are surveyed by Holmbom, Bergquist and Vanhatalo (2014).
They find that there is a lack of empirical evidence on the improved profitability
for both the provider and the user, although this is one of the key motivators the
implementation of PBC. The procurement of performance in a PSS from a buyer’s
perspective is discussed by Elmazoski et al. (2016). They establish a framework
which links payment models and service orientation to identify different PSS
purchasing options for the buyer. Specifically, it discusses the incentives for buyer
and provider in the various models (Elmazoski, et al., 2016).
39
A number existing contributions deal with the planning and configuration of sup-
ply networks for services or PSS, e.g. by Lockett et al. (2011) and Xu et al. (2016).
However, they do not address the operational benefits of IoT for improved infor-
mation management for servitized products. Johnson and Mena (2008) propose
and integrated model for product supply chains and service supply chains in
order to cater of servitized products. Furthermore, they highlight the importance
of real-time information and the information flow management. Grubic (2014)
discusses the importance of real-time information as a measure to mitigate risk in
servitization, without relating to CPS. The value-based organization of business
processes for a product-service supply chain is described by He et al. (2016).
Several existing contributions deal with various aspects of IT-supported logis-
tics process coordination of smart products. Examples are the dispatching of
technicians (Bader et al., 2017), the replenishment process (Alfathi, Lyhyaoui and
Sedqui, 2015), spare parts availability (Chaudhuri and Ivcekno, 2017), replenish-
ment policy (Hosoda and Disney, 2012). The potentials of CPS for maintenance,
repair and operations (MRO) have been analyzed by Trentesaux et al. (2015) using
two case studies of aircrafts and trains. They, however, do not provide a link to
business models and focus on MRO tasks in general rather than PSS logistics in
general. Still, the contribution of Trentesaux et al. is a helpful work to illustrate
the principle of applying CPS to support maintenance in product fleets. Ardolino
et al. (2017) present a very recent and comprehensive study on the application
of digital technologies such as IoT, Cloud Computing and Predictive Analytics in
servitization. Furthermore they introduce helpful terminology for service trans-
formation paths such as “availability provider” (offering products in a PPU model)
and “performance-provider” (offering products in a PBC model), both of which
employ IoT for usage data, product performance.
With regard to related work, it can be concluded that the IoT-enabling of PSS to
support new business models has been widely discussed from various perspec-
tives in literature. However, to the best of our knowledge, we could not identify a
contribution regarding the logistics of PSS and their coordination with the help of
IoT-enabled information flows.
40
5 Case Analysis
5 Case Analysis
We use publicly available information as well as scientific publications on the

following cases of existing real-world IoT-enabled PSS to identify logistics require-
ments:
— Winterhalter provides industrial dishwashers in a “pay-per-wash” busi-

ness model. The fixed price includes all detergents, water treatment,
baskets, maintenance and service. There is also no minimum contract
duration, which means that all the installed equipment may have to
be transferred to another customer soon after the signup (Winterhalter,
2016).
— Canon offers document management solutions in pay-per-page contracts,

which include consumables replenishment, spare parts, maintenance
and technician labor (Canon Europe, 2017). Canon equipment can con-
nect and send to the Canon eMaintenance platform data related to meters
and machine status, ink levels, activity log and fault registry (Ardolino et
al., 2017).
— Kaeser Compressors offer is called Sigma Air Utility, which uses a PBC
business model with a fixed price per cubic meter compressed air. Users
get an individual solution concept, which is build, operated and opti-
mized by Kaeser. To enable remote monitoring of all compressors, the
Sigma Telecare is employed to gather data for both predictive mainte-
nance and energy efficient operation. Spare parts are delivered through
a worldwide network of logistics partners. These measures are aimed to
reduce downtimes and unnecessary onsite visits (Kaeser Kompressoren,
2017).
— Car2Go is a car sharing system, which is offered in various large cities,

e.g. Berlin, Vienna, Vancouver and New York City. Users are billed by the
minute, with prices depending on the type of car. Cars can be located,
booked and opened via a smartphone app. Car2Go will take care of
all required maintenance and cleaning, as well as paying for parking
lots, insurance and other vehicle related cost. If a car has to be refueled
by the user, he or she gets a credit of 10 extra driving minutes (car2go
Deutschland GmbH, 2017).
41
— Hewlett Packard offers the “Instant Ink” program for toner cartridges,
which are charged in a cost-per-page model. Different plans with various
numbers of included pages for a monthly subscription are offered for
occasional, moderate and frequent users. To participate, the user must
own an Instant Ink compatible HP printer, and register for the “HP con-
nected” service, which is facilitates the transmission of usage information
from the user to HP over the internet. The user receives new toner car-
tridges automatically via postal mail before the ink runs out. Also, empty
cartridges are returned to HP partners for recycling via postage-paid
shipping material (HP Development Company, 2017).
From these cases, we can derive the following main logistics processes: Delivery,
Maintenance, Replenishment and Recycling (see table 1). As Delivery is a process
that takes place before the actual operational logistics support begins, we do not
further discuss it in this research. However, it is obviously a process that is part of
the provider’s fleet management and can be also supported using IoT data.
42
Table 1: Logistic requirements of the selected cases
Replenish-
Case Model Delivery Maintenance Recycling
ment
Winter- PPU Machine
Included Detergents (unclear)
halter (per Wash) included
Labor, spare
PPU Setup by parts, mainte-
Canon Ink
(per Page) Canon nance kits,
print heads
Kaeser PBC (per
By Kaeser, Labor,
Com- m³ com- n/a n/a
incl. setup spare parts
pressors pressed air)
PPU (per Pickup Cleaning, Refueling

Car2Go n/a
Minute) by User maintenance by User
HP PPU Empty
Parcel n/a Ink
5 Case Analysis
Instant Ink (per Page) cartridges
43
6 A Framework for PSS Logistics
In this section, the results from the case analysis and related work are consolidated
into a general framework for PSS logistics. The main focus is the identification of
provision of information that is required to allow for a timely conduct of logistics
processes. First, we propose the following terms to generalize the logistics of a
PSS:
— Equipment: a product, machine or device that performs the desired

service
— User: the entity that requires the performance of an equipment
— Provider: the company that provides the PSS and takes care of its servic-
ing
— Service Level Agreement (SLA): an agreement between user and provider

on quantifiable properties of a service, e.g. availability.
— Consumables: any physical goods, which are used up by the equipment

during operation, such as ink in a printer, lubricants in machines etc.
— Spare Part: any component, which has to be replaced during the use of
the equipment
— Replenishment: any physical operation that is required to provide the

equipment with consumables.
— Maintenance: any physical operation that is required to improve the relia-

bility and longevity of equipment, which can also be determined through
safety laws and regulations, e.g. calibration of measurement devices.
Examples for maintenance are cleaning, inspection, and replacement of
wearing parts.
— Repair: any physical operation that is required to restore the correct

function of an equipment
— Recycling: the return of any unwanted physical material that is created

during the operation of the equipment.
44
To conduct the identified processes as efficiently as possible, IoT-enabling is

employed to create a CPS which provides automation for the individual steps from
physically identifiable demand, digitally detected demand, data transmission,
event processing, workflow initiation, physical provisioning of material, delivery
to operation site, and local operations to fulfill the demand.
For each of the operational logistics processes, the following characteristics are
thus required to propose a generalized CPS architecture: To perform the process,
one or multiple workflows have to be performed at the service provider. This
creates a demand for information, which helps to decide, when these workflows
have to be triggered. To fulfill these demands, data on usage, condition, location
etc. is needed from the equipment. This data is then processed using analytics,
which mainly fall in the category of predictive analytics or machine learning algo-
rithms (Ardolino et al., 2017). The consolidated characteristics for PSS logistics
process support are shown in Table 2.
45
46

Table 2: Generalized support PSS logistic processes
Information
Process Workflows Data Sources Analytics
Demand
Schedule service,
Reach of
Replenish- schedule transport, Stock levels, Consumption
remaining
ment allocate consuma- operational data patterns
consumables
bles
Remaining Equipment
Schedule service, Predictive
Repair lifetime of sensors,
allocate spare parts Maintenance
components operational data
Operational Determine
Main- Maintenance
Schedule service hours, inspec- maintenance
tenance demand
tion history timeframe
Determine
Pickup Current level
Recycling Schedule transport pickup
demand of residuals
timeframe
On this basis, we can deduct a generalized architecture of a CPS to support PSS

logistics. It is based on the framework for Cyber-physical Logistics System (CPLS),
as proposed by (Prasse, Nettstraeter and Hompel, 2014). It also relates to the
findings of Grubic (2014) with regard to the usage of real-time information and
remote monitoring in servitization. The concept of cloud-based analytics for
predictive maintenance has been previously described, e.g. by Johanson and
Karlsson (2016).
The main elements of the proposed architecture are:
1. The equipment, which contains sensors, embedded computing and com-

munication capabilities to record and transmit required data, such as
operational status, location, fill levels etc.
2. The data is sent to an IoT Cloud, where it is stored in a database. All ana-
lytics functions can be performed using this data. Based on the outcome
of the analytics functions, trigger events are created.
3. The providers’ transaction systems in the backend receive the events and
can initiate workflows accordingly.
4. The logistics execution within the triggered workflows can be conducted
with the help of technicians for local operations and logistics service
providers for delivering spare parts and consumables as well as picking
up any residuals for recycling.
The proposed architecture is presented in figure 1.
47
48

Figure 1: A generic CPS architecture to support PSS logistics
7 Discussion
7 Discussion
This research shows that there is currently a lack of generalized knowledge on

the principles and mechanisms of logistics for PSS. Given the high complexity of
PSS implemented on CPS, such knowledge is required to describe, explain, and
design these systems.
The results of this paper are meant to be a first step towards these goals. We
found that a common basis for PSS logistics can be established, even though the
analyzed cases show a wide spectrum of designs. Specifically, the processes of
replenishment, maintenance, and recycling could be identified to be of relevance.
Besides, a set of terms have been proposed to describe the main actors and
relations.
Furthermore, the employment of IoT technology to monitor equipment either

avoids error-prone and expensive manual data entry or makes the overall PSS
offer economically viable in the first place. The proposed generalized architecture
is a conceptual model of a CPS for PSS logistics, which shows how equipment,
IoT-technology, analytics, and workflows can be combined to coordinate logistics
processes for PSS. SLAs in performance-based contracted products create special
requirements on the logistics process, which must be considered as part of the
integrated product-service system (PSS). The use of IoT technology in physical
equipment enables efficient information flow to support the fulfillment of SLAs.
Practitioners and researcher alike may use our results to get a better overall un-
derstanding of PSS logistics. For the design of equipment targeted for use in a PSS,
it is advisable to consider the information demand and related data to prepare for
the integration of sensors, computation and communication capabilities. It might
also help logistic service providers to develop new services for PSS providers.
It should be noted, that in this research, only a relatively small number of cases
was analyzed. Furthermore, all case data was gained through existing public
information, mainly from the providers themselves. A larger sample of cases
and additional expert interviews with providers and users will probably create
a more detailed and differentiated result. However, we are convinced that the
main findings of this research still hold true, and can serve as a framework for
further research results.
For future research, we identified the three different directions fleet management,
SLA compliance and refinement for implementation:
49
1. Fleet management: Within this paper, only the operational logistics of

a PSS are covered. From a provider’s perspective, the provisioning and
operation of PSS is a fleet management task, which also includes logis-
tics processes. Examples are the provisioning to and the return of the
equipment from the customer to the provider. For these processes, the
current location and status is also important, e.g. to decide whether
outstanding requests can be already fulfilled with a particular equipment
in the fleet. Besides the actual transport, technicians might be required
to dismount and setup the equipment at the respective locations. Other
fleet management tasks include the use of monitoring data for optimized
operations, e.g. to reduce wear of components and energy consumption.
2. SLA compliance: While it has already been stated that SLAs can be part
of the PSS operation, the adherence to these SLAs was not in scope of
this research. For that, a timing perspective has to be established in
the proposed framework. Another improvement could be the inclusion
of additional data from the user, e.g. production schedules, shift plans
etc. which can be used as additional input in the scheduling to reduce
interruptions. Likewise, the process status, e.g. tracking and tracing
information could be relayed back to the equipment in order to optimize
its operations and provide more detailed information on the current
status to the user. Along these lines are proactive business processes,
which are triggered based on forecasts rather than actual events. This
might be of particular value in planning and dispatching processes, e.g.
to optimize management of product fleets, scheduling of technicians as
well as the optimization of stock levels and order sizes for replenishment.
3. Refinement for implementation: As the proposed architecture is on a con-
ceptual level, it needs further refinement with regard to both information
logistics and infrastructure. Information logistics should describe in de-
tail, which data has to be recorded, how it is pre-processed, when it has to
be transmitted, where it is stored, how it is analyzed, and how long it has
to be retained to fulfill the requirements of the PSS. Infrastructure refers
to a concrete technical system with concrete sensors, communication
technology, transmission protocols and IoT cloud architectures.
50
8 Conclusion and Outlook
8 Conclusion and Outlook
The increasing importance of PSS is recognized both in academia and practice.

With the emergence of reliable IoT-technology, the integration of physical and dig-
ital processes in CPS becomes a driver for further automation and thus efficiency
in the operation in such systems. When technical products are offered in PBC or
PPU business models, the product is combined with operational services into a
PSS. Such PSS were successfully established in practice, as shown by the cases
mentioned earlier. With that, the buyer of the PSS transfers the responsibility
and risk for these operations to the service provider, e.g. the manufacturer of the
product. While such scenarios can create a variety of positive effects for both user
and provider, we focused the effects of IoT-functionality in technical products on
the efficient logistics operations for PSS.
The results of this research fulfill the goal of identifying how logistics processes
in PSS can be coordinated through IoT-enabled products. More specifically, the
research links existing concepts from different research fields together, such as
PSS, CPS, IoT, PBC and PPU business models and relates them to real-world use
cases. With these results, we aim to contribute to the digitization of logistics with
regard to the operation of PSS. However, we acknowledge that this can only be
a starting point, as there are many open research issues, some of which were
presented in the previous section.
To follow these research directions is worthwhile from our point of view. Not
only because of the increasing relevance of PSS in general. Also the ongoing
digitization provides new opportunities to improve efficiency and customer value
of such systems. Various technological advancements within the broader concept
of “Industrie 4.0” can have an enormous impact on the operation of PSS. For
example, additive manufacturing (commonly known as “3D printing”) might allow
producing spare parts when they are needed nearby the place of operation, which
reduces stock keeping and delivery times. Another relevant are of innovations
is human-computer interaction. For example, augmented reality might help
less experienced local technicians to get context-based instructions for repair or
maintenance. Their interaction might be enriched with information from senior
engineers at the manufacturer, who are able to see both the view of the technician
as well as the current state of the product and its history. In summary, such
developments may reduce lead times in case of failures and improve efficiency
in the operations, but increase complexity of the overall system. Therefore, the
51
design and operation of PSS in PBC or PPU business models requires an integrated
approach of technical, commercial and behavioral aspects.
References
Alfathi, N., A. Lyhyaoui, and A. Sedqui (2015). “A New Product Driven System Concept for the
Replenishment Process”. In: Operational excellence in logistics and supply chains. Ed. by W.
Kersten, T. Blecker, and C. M. Ringle. Proceedings of the Hamburg International Conference of
Logistics (HICL). Berlin and Berlin: epubli GmbH.
Allmendinger, G. and R. Lombreglia (2005). “Four strategies for the age of smart services”. In:
Harvard Business Review 83.10, p. 131.
Ardolino, M., M. Rapaccini, N. Saccani, P. Gaiardelli, G. Crespi, and C. Ruggeri (2017). “The role of
digital technologies for the service transformation of industrial companies”. In: International
Journal of Production Research 83.10, pp. 1–17.
Bader, S., M. Vössing, C. Wolff, J. Walk, and M. Maleshkova (2017). “Supporting the Dispatch-
ing Process for Maintenance Technicians in Industry 4.0”. In: Tagungsband der 9. Konferenz
Professionelles Wissensmanagement. Ed. by S.-V. York, S. Zander, and A. Harth.
Baines, T. S., H. W. Lightfoot, S. Evans, A. Neely, R. Greenough, J. Peppard, R. Roy, E. Shehab,
A. Braganza, A. Tiwari, J. R. Alcock, J. P. Angus, M. Bastl, A. Cousens, P. Irving, M. Johnson, J.
Kingston, H. Lockett, V. Martinez, P. Michele, D. Tranfield, I. M. Walton, and H. Wilson (2007).
“State-of-the-art in product-service systems”. In: Proceedings of the Institution of Mechanical
Engineers, Part B: Journal of Engineering Manufacture 221.10, pp. 1543–1552.
Barbosa, J., P. Leitão, D. Trentesaux, and S. Karnouskos (2016). “Cross Benefits from Cyber-Physical
Systems and Intelligent Products for Future Smart Industries”. In: Proceedings of the 14th IEEE
International Conference on Industrial Informatics.
Broy, M., M. V. Cengarle, and E. Geisberger (2012). “Cyber-Physical Systems: Imminent Challenges”.
In: Large-scale complex IT systems. Ed. by R. Calinescu and D. Garlan. Vol. 7539. LNCS sublibrary.
SL 2, Programming and software engineering. Heidelberg and New York: Springer, pp. 1–28.
Canon Europe (2017). Total Service Care Large Format Solutions.
car2go Deutschland GmbH (2017). car2go Carsharing.
Cavalieri, S. and G. Pezzotta (2012). “Product–Service Systems Engineering: State of the art and
research challenges”. In: Product Service System Engineering: From Theory to Industrial Applica-
tions Product Service System Engineering: From Theory to Industrial Applications 63.4, pp. 278–
288.
Chaudhuri, A. and J. Ivcekno (2017). “Framework for improving spare parts availability using big
data analytics: case study of a mining equipment manufacturer”. In: 9th Service Operations
Management Forum.
Demirkan, H., C. Bess, J. Spohrer, A. Rayes, D. Allen, and Y. Moghaddam (2015). “Innovations
with Smart Service Systems: Analytics, Big Data, Cognitive Assistance, and the Internet of
Everything”. In: Communications of the Association for Information Systems 37.35.
Elmazoski, J., A. H. Glas, U. David, M. Eßig, and M. Seiter (2016). “Performance-based contract-
ing of PSS: Methodological aspects of an experimental investigation”. In: Proceedings of the
25th IPSERA Conference.
52
References
Essig, M., A. H. Glas, K. Selviaridis, and J. K. Roehrich (2016). “Performance-based contracting in

business markets”. In: Industrial Marketing Management 59, pp. 5–11.
Fleisch, E., M. Weinberger, and F. Wortmann (2015). “Business Models and the Internet of Things”.
In: Interoperability and Open-Source solutions for the Internet of Things. Ed. by I. Podnar Žarko,
K. Pripužić, and M. Serrano. Vol. 9001. LNCS. Cham: Springer, pp. 6–10.
Freund, S. and W. Stölzle (2016). “Configuring PBC for operating of capital goods – A case-based
typology”. In: Supply Management Research: Aktuelle Forschungsergebnisse 2015. Ed. by R.
Bogaschewsky, M. Eßig, R. Lasch, and W. Stölzle. Wiesbaden: Springer Fachmedien Wiesbaden,
pp. 177–200.
Gassmann, O., K. Frankenberger, and M. Csik (2014). The business model navigator: 55 models
that will revolutionise your business ; [discover the 55 business models responsible for 90% of
the world’s most successful companies]. 1. publ. Always learning. Harlow u.a.: Financial Times
Pearson Education.
Grubic, T. (2014). “Servitization and remote monitoring technology”. In: Journal of Manufacturing
Technology Management 25.1, pp. 100–124.
He, T., W. Ho, Y. Zhang, and P. K. Dey (2016). “Organising the business processes of a product
servitised supply chain: A value perspective”. In: Production Planning & Control 27.5, pp. 378–
393.
Herterich, M. M., W. Brenner, and F. Uebernickel (2016). “Stepwise Evolution of Capabilities for
Harnessing Digital Data Streams in Data-Driven Industrial Services”. In: MIS Quarterly Executive
15.4, pp. 297–318.
Herterich, M., F. Uebernickel, and W. Brenner (2015). “The Impact of Cyber-physical Systems on
Industrial Services in Manufacturing”. In: Procedia CIRP 30, pp. 323–328.
Holmbom, M., B. Bergquist, and E. Vanhatalo (2014). “Performance-based logistics – an illusive
panacea or a concept for the future?” In: Journal of Manufacturing Technology Management
25.7, pp. 958–979.
Hosoda, T. and S. M. Disney (2012). “On the replenishment policy when the market demand
information is lagged”. In: International Journal of Production Economics 135.1, pp. 458–467.
HP Development Company, L. P. (2017). HP Instant Ink.
Johanson, M. and L. Karlsson (2016). “Service architectures for product and production availability:
A system of systems approach”. In: 11th System of Systems Engineering Conference (SoSE), IEEE.
Piscataway, New Jersey: IEEE, pp. 1–6.
Johnson, M. and C. Mena (2008). “Supply chain management for servitised products: A multi-
industry case study”. In: International Journal of Production Economics 114.1, pp. 27–39.
Kaeser Kompressoren (2017). Operator model: SIGMA AIR UTILITY.
Kagermann, H., F. Riemensperger, and P. Weckesser (2016). Smart Service Welt: Digitale Service-
plattformen – Praxiserfahrungen aus der Industrie: Best Practices. Ed. by acatec. München.
Lockett, H., M. Johnson, S. Evans, and M. Bastl (2011). “Product Service Systems and supply
network relationships: An exploratory case study”. In: Journal of Manufacturing Technology
Management 22.3, pp. 293–313.
Medina-Borja, A. (2015). “Editorial Column—Smart Things as Service Providers: A Call for Con-
vergence of Disciplines to Build a Research Agenda for the Service Systems of the Future”. In:
Service Science 7.1, pp. ii–v.
Mont, O. (2002). “Clarifying the concept of product–service system”. In: Journal of Cleaner Produc-
tion 10.3, pp. 237–245.
Neely, A. (2007). “The Servitization of Manufacturing: An Analysis of Global Trends”. In:
53
Neely, A. (2008). “Exploring the financial consequences of the servitization of manufacturing”. In:
Operations Management Research 1.2, pp. 103–118.
Porter, M. E. and J. E. Heppelmann (2014). “How Smart, Connected Products Are Transforming
Competition”. In: Harvard Business Review.
Prasse, C., A. Nettstraeter, and M. ten Hompel (2014). “How IoT will change the design and operation
of logistics systems”. In: 2014 International Conference on the Internet of Things (IOT). Piscataway,
NJ: IEEE, pp. 55–60.
Sala, R., V. Zanetti, G. Pezzotta, and S. Cavalieri (2017). “The role of technology in designing and
delivering Product-service Systems”. In: International Conference on Engineering, Technology
and Innovation (ICE/ITMC).
Schief, M., C. Kuhn, P. Rösch, and T. Stoitsev (2011). “Enabling Business Process Integration of
IoT-Events to the Benefit of Sustainable Logistics”. In: 1st International Conference on Logistics,
Informatics and Service Science (LISS 2011).
Selviaridis, K. and F. Wynstra (2014). “Performance-based contracting: A literature review and
future research directions”. In: International Journal of Production Research 53.12, pp. 3505–
3540.
Trentesaux, D., T. Knothe, G. Branger, and K. Fischer (2015). “Planning and Control of Maintenance,
Repair and Overhaul Operations of a Fleet of Complex Transportation Systems: A Cyber-Physical
System Approach”. In: Service orientation in holonic and multi-agent manufacturing. Ed. by
T. Borangiu. Vol. 594. Studies in Computational Intelligence. Cham: Springer, pp. 175–186.
Tukker, A. (2004). “Eight types of product–service system: Eight ways to sustainability? Experiences
from SusProNet”. In: Business Strategy and the Environment 13.4, pp. 246–260.
Valencia, A., R. Mugge, J. Schoormans, and H. Schifferstein (2015). “The Design of Smart Product-
Service Systems (PSSs): An Exploration of Design Characteristics”. In: International Journal of
Design 9.1, pp. 13–28.
Weinberger, M., D. Bilgeri, and E. Fleisch (2016). “IoT business models in an industrial context”. In:
at - Automatisierungstechnik 64.9.
Winterhalter (2016). Pay-Per-Wash.
Xu, Z., X. G. Ming, T. Wu, and M. Zheng (2016). “Product Service Supply-Chain Design”. In: Handbook
of manufacturing engineering and technology. Ed. by A. Y. C. Nee. Springer reference. New York
[New York] and Boston, Massachusetts: Springer and Credo Reference, pp. 3219–3237.
54
Towards a Web Based Transportation
Infrastructure
Florian Kleedorfer1 , Christian Huemer 2
1 – Research Studios Austria
2 – Technical University of Vienna
The level of digitalization within transport companies is much higher than the
level of digitalization across organization boundaries. This fact suggests that
there is room for improvement. However, this situation is not likely to change
as long as there is no financial incentive for the whole sector to cooperate in
establishing a shared communication infrastructure. In this paper, we present
our approach for building such an infrastructure using the method of design
science. The goal is an open, Web based, de-centralized network operated by
transport organizations themselves. Based on expert interviews, we argue that
the current situation causes frictions that our approach may help reduce, thereby
providing the incentive to participate. The proposed system is described in terms
of its existing technologicial base, the Web of Needs, and the extensions needed
to provide the required functionality, giving an overview of the current state of
implementation.
Keywords: linked data; transport; decentralization; digitalization
55
Towards a Web Based Transportation Infrastructure
1 Introduction
Companies in the transportation domain have developed their IT infrastructure

in recent years, giving clients ever easier access to the services they provide.
Mobile apps and Websites allow users to create accounts and comfortably book
transports, make pick-up/delivery appointments, and pay their invoices. These
developments reduce the transaction cost of becoming and onboarding a new
client.
However, this style of digitalization is not without drawbacks. For one, there are
considerable costs associated with building and maintaining these tools. Because
of that, they are primarily designed to meet the requirements of the transporter,
not its clients. They tend to lock users in a streamlined one-stop shop for trans-
portation, whereas the opposite would be in the clients’ interest: to be able to
compare service offerings and prices across vendors, and to switch between ven-
dors easily. Instead, clients are confronted with a plethora of communication
channels, apps and on-line accounts each of which lets them communicate with
just one vendor.
Recent economic successes of matchmaking platforms like Uber or Airbnb (Uber
2017; Airbnb 2017) have proved the need for a vendor-independent medium
where supply and demand can meet. Such a medium for the management of
transactions between market participants obviously serves as a unifying force
leading to standardized interfaces between the participants, lowering the cost of
finding a partner and enabling trust between newfound partners.
While a remedy to some issues of the more traditional style of digitalization, this
approach has drawbacks, too. Most prominently, matchmaking platforms tend
to create winner-take-all (or winner-take-most) situations: For any given product
or service, it makes more sense for a participant to join the platform that already
hosts more of the clients the participant interested in. The result is a small number
of dominant platforms that have a de-facto monopoly.
The intrinsic logic of the platform business model is to allow transactions only
within the platform, and to allow only the transactions that are directly or indi-
rectly economically profitable for the platform. While the platform model thus
allows participants to reach a bigger pool of potential partners than the tradi-
tional model (each participant operates their own infrastructure), it introduces
new limits on that pool out of the platform’s self-interest: only members of the
platform are meant to interact.
56
2 Research Method
Our goal is the development of a de-centralized, open, standards-based infras-

tructure for transportation, providing the adequate technology for all market
participants without the limitations induced by the necessary self-interest of the
centralized platform model. The vision is to interconnect the existing IT systems
of transport companies on a protocol level instead of connecting each of them
to one central hub. This approach allows for diverse, inter-operable software
products that can be tailored to each participant’s needs; all market participants
can operate services of their own and thereby retain ownership of their data and
they are free to engage in interactions with any other participant that implements
the protocol.
In this paper, we report findings from expert interviews that support the relevance
of the problem we are addressing and give us a direction for further development.
We explain how we harness prior work and show which adaptations and exten-
sions have to be made in order to solve the problem, and finally, we describe the
current state of design and implementation.
2 Research Method
In our work, we follow the method of design science in information systems

research, proposed by Hevner (Hevner et al., 2004). According to this framework,
design science is primarily focused on the creation of technological artifacts.
The research topics are obtained from the environment of the artifact in the
form of relevant business needs. When finished, artifacts have to be applied
in the environment in order to assess their effects. This cyclic process is called
the relevance cycle. Using the available knowledge base, artifacts are iteratively
adapted and evaluated using rigorous scientific methods. Theoretical insights
are added to the knowledge base through scientific publication, completing the
rigor cycle (Hevner, 2007).
In the following, we describe our research process as an application of this method.

The first step is the assessment of problem relevance and the elicitation of busi-
ness needs, which we do by interviewing domain experts. As a second step we
propose a design for technological artifacts to address the business needs we
found.
57
3 Relevance: Qualitative Interviews
The problem we postulate is that there is considerable friction between market

participants in the domain of transportation in its current state of digitalization.
Our main hypothesis is that a unified open infrastructure that allows for matching
suitable business partners with each other can reduce friction.
3.1 Sample Selection and Data Collection
In order to asses the problem relevance and to find out which frictions are deemed
important by domain experts, we conducted qualitative expert interviews with
individuals who are active in the field of transportation (DiCicco-Bloom and Crab-
tree, 2006). The interview partners were initially identified by Web search for
keywords mainly from the domain of courier and parcel services in Austria. Inter-
views were conducted with those who showed interest in cooperating. The range
then was extended by a snowball technique (Varvasovszky and Brugha, 2000).
In the course of this analysis, the focus was widened to include the forwarding
sector as initial results pointed toward greater than expected problem relevance
in that area. A number of interviews were conducted at an agricultural trade fair,
which explains the influence of that sector in the results.
In total, 27 interviews were conducted with a prepared interview plan that evolved
a little over time; five interviews could be completed according to the plan, the
others had to be shortened due to limited availability of the participants. The
interview partners are active in the roles of trader(10), transporter(5), logistics(2),
broker(2), or others (8). Data collection was done by note-taking. The notes were
subsequently analyzed for commonalities that caught our interest.
3.2 Findings
In the following we highlight our findings, noting the number of participants

whose interviews support each one. For each finding, we argue how it influences
our assessment of problem relevance. While some findings will directly relate to
functionality currently under development, some findings can at this point only
serve as motivations for functionality that may be provided by our infrastructure
in the future.
58
1. Trust. Especially when freight is high-volume or high-value, established

contacts between consignor and transporter are highly valued. Trusting
new players is difficult and not deemed necessary once good business
relationships have been set up. This finding disconfirms the relevance
of our work because players prefer existing business relationships and
do not need matching with new partners. On the other hand, it would
motivate a feature that enables trust in new partners, and of one that
allows representing and leveraging one’s business network. Support: 9.
2. Digitalized communication. Transportation is digitalized only within

companies, and even within companies, the actual execution is often not
digitalized (communication with driver or for confirming handovers). If it
is digitalized across companies (other than via e-mail or messaging apps),
it is set up in such a way that the client company integrates the API or uses
the user interface of the transport company. The general view of market
participants was that the main means of communication is through phone
calls, using E-mail for written confirmation. Companies do have systems
using barcodes, matrix codes or electronic identification tokens to trace
their consignments, but they do not seem to work across companies.
This finding supports problem relevance because it shows there is room
for automation of communication across companies. Support: 4.
3. Documents. In some constellations, consignors and consignees demand

the original paper documents from the consignor be passed on to the
consignee, incurring additional complexity. One such docucment is the
CMR1 consignment note (UNECE, 1956) that needs to be signed by the
consignee and find its way back to the consignor - failing to obtain the
signed CMR consignment note can incur considerable costs. The con-
signee, on the other hand, has little to no incentive to send it back, which
can cause serious problems.2 Together with Finding 2 (digitalized com-
munication), this finding is a strong argument for further technological
harmonization of standards for transport documents and their seamless
integration in everyday business. Support: 4.
1
CMR is short for the United Nations’ Convention on the Contract for the International Carriage of
Goods by Road
2
Overall, delivery paperwork is being replaced by electronic data interchange. Concerning the
CMR document it should be noted that at the time of this writing, the digital e-CMR Waybill is
available in a number of European countries. (UNECE, 2008). It was not mentioned in any of
our interviews, though.
59
4. Privacy. Keeping trade secrects is an important issue for players in the

transportation domain. Several experts voiced concerns that their con-
tracts and bids may be inspected by other parties in an open system,
and therefore took quite a reserved stance toward our work. Multiple
experts stated that if requests for quotes and their bids could be revealed
to a selected audience only, it would be a system they would welcome.
Support: 4.
5. Appointments. Depending on the means of transportation and the situa-
tion of the consignee, the complexity of making appointments for pick-up
and delivery as well as the consequences of errors can vary greatly. The
main medium for making appointments and informing about changes
is telephone. Large logistics companies have dedicated communication
channels (apps or automated SMS). This finding generally supports the
relevance of our work if the developed technology supports the integra-
tion of processes and resources for coordinating appointments so as to
standardize the situation, removing the burden of last-minute improvisa-
tion and providing a reliable record of the events that led to a problem.
Support: 3.
6. Calls. Being able to talk to their counterpart in addition to written com-
munication is described as important by some experts. It allows for quick
problem solving, helps building trust quickly (video calls or meetings in
person even more so), and it is a good way of maintaining the customer
relationship. Ideally, it would be possible to reach the relevant person
(e.g. call the truck driver to inform them of an important change). This
finding provides support for an integration of voice or video chat in a
technology managing transport relationships. One broker noted that a
conference call with multiple participants would be ideal. Support: 3.
7. Certificates. Some experts stated that friction is caused by having to
prove that all parties in the logistics chain hold a certain certificate, such
as the GMP+ certificate for sustainable production. This finding supports
problem relevance and may warrant the use of blockchains. Support: 3.
8. Framework agreements. High-frequency consignors do not set up each
transport assignment individually. Rather, they select transportation
partners, often on a yearly basis, set up a framework contract, and then
execute many transports with easily calculated rates and quick plan-
ning. This finding seems to disconfirm the problem relevance as a system
that matches transportation partners for each individual transport is not
60
needed for this situation. However, if the solution allows for finding trans-
portation partners for framework agreements, and setting up individual
transports based on a framework agreement, possibly in a private call for
proposals among partners with whom a framework contract has been
set up, the described situation might even be improved. Support: 3.
9. Return freight. For a one-off transport, getting a transportation quote
can take a long time (one week is not uncommon for agricultural goods)
because of the need to find a freight for the return journey. This find-
ing supports problem relevance as the delay causes friction in the infor-
mation or negotiation phase that might be reduced by a more efficient
process of finding a transport request for the return journey. A solution
feature that could be especially useful here would be the option to offer
a transport only on the condition that a corresponding return contract is
found, and to represent and monitor the condition satisfaction automati-
cally. Support: 3.
10. Price sensitivity. Multiple experts stated that transportation is very price-
sensitive - price is the main decision factor for clients. Quality of service
(keep appointments, response speed and accuracy, no damages) is only
a negative decision factor in the case that it does not meet the expected
standard. A matchmaking technology like the one we are proposing
clearly addresses a market need of clients but it does so at the expense
of transporters, who seem to feel threatened by our vision. Support: 3.
11. Price negotiation. Some experts stated that price negotiation is a cor-
nerstone of the business. A system that does not allow for negotiating is
not acceptable. This finding does not affect our assesment of problem
relevance but points toward a negotiation feature. Support: 2.
12. Quality dimensions. Some experts stated they would not want to com-
pete world-wide with everyone because of price sensitivity, but they felt
they had a higher-quality offering than cheaper alternatives. It would be
necessary for open competition to be able to advertise verifiable quality
dimensions such that clients understand the higher price. This finding
shows that the technology should provide the possibility to differentiate
offerings from each other in multiple dimensions and to make verifi-
able claims about service/offer quality. The deeper problem this finding
touches upon is that apart from price, clients often do not know about
aspects of quality they require, either because these requirements are
implicit or because clients lack experience. Therefore, in addition to
61
providing a means for verifiable claims, it would seem beneficial if the

technology provided a means for conducting a conversation with the
client about service qualities with the goal of educating the user about
alternatives so as to allow them to make a more informed decision. Sup-
port: 2.
13. Wrong Assumptions. In some cases, important information is present
only in an implicit form. For example, the recipient of a shipment of
agricultural products may only be able to unload a dumper truck, not
any other kind of truck - but this fact is assumed to be known to the
transporter, which may lead to the wrong kind of truck being used, in-
curring delays and additional cost upon delivery. This finding provides
support for machine-interpretable, standards-based, automatic informa-
tion interchange that may uncover incompatibilities like the one above
automatically. Support: 1.
14. Freight exchanges. Transporters are reluctant to use freight exchanges
to find new assignments because the prices they can ask there are very
low, to the point that it is cheaper to make an empty return journey than to
obtain a contract there. This finding may be seen to disconfirm relevance
of our work, as the technology could have a similar effect for transporters
overall. It is possible, however, that this effect is related to systems that
focus mostly or solely on price and that in this case, the solution features
discussed in connection with Finding 12 (quality dimensions) could ease
the situation for transporters. Support: 1.
15. Contract negotiation. Contract negotiation may not be necessary, but
when it is, it can be a lengthy and error-prone phase. In some cases, one
side simply dictates the contract’s conditions. This is the case when a
transport partner is being searched for a longer term business relation-
ship, and transporters make bids, or when a transport company offers
standardized services together with standardized terms. In some situa-
tions, contracts have to be set up for individual transport assignments,
which is usually done using MS Word documents sent back and forth via
e-mail. At least one of our interview partners expressed the hope for a
better integrated solution that allows for more clarity during this phase.
Support: 1.
We judge Findings 2, 3, 5, 7, 9, and 13 as supporting problem relevance. Find-
ings 1, 4, 6, and 8 do not clearly refer to problems calling for a solution. Rather,
they describe current strategies evolved to avoid friction as much as possible.
62
4 Knowledge Base
Insofar, we would not say they contradict problem relevance, but we cannot in-
clude them in the set of supporting findings, either. Findings 10 and 14 disconfirm
problem relevance from the point of view of the transporters because the solution
may mean more pressure on prices. Findings 11, 12, and 15 are interpreted as
being neutral with respect to problem relevance.
In total, the interviews do seem to indicate the problem we are tackling is relevant.
This assessment is backed up by the fact that twelve experts said they would be
happy to try out our system and six welcomed our our plans explicitly. We were
able to collect information about expectations and requirements that may lead
to future developments, most of which need further refinement. In our design
process, we focus on fundamental functionalities first, which are generally taken
for granted in such interviews and therefore do not come up as important topics.
Therefore, the full consequences of the findings on concrete design decisions may
materialize at a later point in the process.
In the following, we present our approach to the technical design of the solution.
Any influence of our interview findings will be noted explicitly.
4 Knowledge Base
4.1 Web of Needs
The main building block for our solution is the Web of Needs (WoN) technology
(Kleedorfer, Busch, Pichler, et al., 2014). Its overall functionality is depicted in
Figure 1. Participants publish supply and demand on a de-centralized network
(1a, 1b), on servers (WoN nodes) of their choice, in a standardized form such
that they are machine-readable and can be found by dedicated, independent
matching services (2) that identify suitable (supply,demand) pairs and inform
them by sending a hint message (3). The participants controlling the supply and
demand objects can choose to establish a communication channel and exchange
messages.
WoN is domain-independent, focusing on providing a framework for common in-
teraction patterns while leaving space for creating domain-specific specializations.
In order to achieve this goal without introducing a break between the framework
and the content, WoN is entirely based on RDF (Manola and Miller, 2004) as a
data description language and the Web as its basic framework: the whole data
63
Figure 1: Interaction diagram showing interactions of two participants and a

matching service in Web of Needs. Image is based on Kleedorfer, Busch,
Huemer, et al., 2016, licensed under CC-BY-SA 4.0.
structure, comprising supply, demand, their connections and the complete data
exchange are represented as linked data (i.e., RDF that can be accessed on the
Web, Bizer, Heath, and Berners-Lee, 2009). The particpant’s privacy is protected by
representing needs (supply or demand), and not users, in the system; each need
can have a dedicated cryptographic key pair which is used to verify any message
it sends. The integrity of the communication history is automatically ensured
by iteratively signing a hash derived from the past message history (Kleedorfer,
Panchenko, et al., 2016). This approach allows for using the communication chan-
nel between two participants as a shared RDF database to which they can both
add data by sending messages, but not delete or otherwise change past data.
By representing the content domain adequately in RDF, this system allows for
cooperatively creating and manipulating a shared machine-interpretable model
of any business transaction.
The immutability property of WoN conversations makes it look similar to block-

chain systems at first glance (Nakamoto, 2008). The immutability property is not
quite as reliable in WoN as it is in a popular blockchain: signatures are only made
by participants of a conversation, as well as by the WoN nodes they use, so threat
model and trust model differ greatly between blockchain and WoN. It remains to
be seen if trust and threat models in WoN are acceptable for practical applications;
64
5 Artifact Design: Open Logistics Networks
it should be noted, however, that WoN provides functionality on a different level

than blockchains: expressing supply and demand, matching, and messaging are
not concerns of blockchain systems.
The reason for WoN to include a cryptograhpically assured message history is
the observation that in practice, the conversation that precedes a transaction
must be interpreted as the contract both parties agree to in the absence of any
more formally defined agreement. It is therefore important that they can rely on
an unchangeable message history in order to prove what has been agreed to in
the event of a dispute. In order to achive this, it is sufficent to have each others’
signatures on each message.
4.2 i-Cargo Ontologies
The EU project i-Cargo (Hofman et al., 2016; A. Garcia, 2015), finished in 2015, is a
highly relevant research endeavour we can build upon. One of the goals of i-Cargo
was the creation of an open freight management ecosystem spanning multiple
organizations and countries. The project produced a number of artifacts that can
be re-used for our work, most importantly a collection of ontologies specifically
developed for the transport sector (Daniele and Ferreira Pires, 2013). These
ontologies are the Logistics Core Ontology LogiCo(Daniele, 2013a), the Logistics
Services Ontology LogiServ(Daniele, 2013b), and the Transport ontology (Daniele,
2013c). They provide a basis for describing the entities relevant for the transport
domain using RDF. Using these ontologies, it is possible to describe such entities
as consignor and consignee, means of transport, consigments and transported
goods, packaging, delivery and pick-up options, transportation requests and
transport execution plans.
The application and extension of WoN to support transport and logistics creates
a virtual medium in which any number of real-world market participants can
be represented in self-organized open networks, hence the name of our project,
Open Logistics Networks.
The general idea is to describe the relevant entities according to the LogiCo,
LogiServ and Transport ontologies (or possibly according to simplified versions
65
thereof), to place them inside the needs according to the Web of Needs approach,
and to publish those needs on the Web.
Figure 2 shows a transport-related interaction diagram to illustrate this approach.

Matching services can use the domain specific descriptions for accurate matching
and allow the needs to establish an RDF-based communication channel. Each
such connection naturally gives rise to an RDF model that both parties have access
to via HTTP. It consists of the set of RDF triples defining the needs and the triples
exchanged in the communication channel. If additional resources are referenced
in any of these triples, they are added to the shared model as well.
For example, consider an example motivated by Finding 13 (wrong assumptions),

with one need describing the delivery event for a consignment of oil seed at an
oil mill, seen from the side of the oil mill. It references an RDF description of the
technical capabilities for unloading. The connected need describes the delivery
event as seen from the transporter, linking to an RDF description of the vehicle,
among other details. Both needs have access to the union of all these triples,
all of which have traceable provenance information. If information is missing or
needs to be changed, it can be requested and provided on the communication
channel in the form of RDF triples. Any state change in the shared transaction (e.g.,
a change to the expected time of arrival) is also represented as new triples that
are added to the channel. An actor organizing a bigger part of the transport chain
controls not just one, but a number of such needs, which are naturally combined
in one RDF model that just covers a larger part of the transport transaction chain
than the models accessible to each of the other partners.
In the remainder of this chapter, we describe adaptations required for applying

WoN in transportation use cases. The interested reader will notice that none of
these topics are uniquely specific to transportation. They must, however, be dealt
with before we can tackle domain-specific functionality that addresses problems
revealed in our interviews.
5.1 Service Descriptions
The basic element of the Web of Needs are entities representing concrete interest
in a transaction, for example, someone may offer a book for sale, and another user
wants to obtain a book. Both users could express their intention in the form of a
need. Both users would get a hint message from a matching service and engage in
66
Figure 2: Interaction diagram showing transport specific interactions of consignor,

consignee, and transporter in the Web of Needs. In the depicted situa-
tion, consignor and consignee have agreed to send the consignment (the
book), and publish needs representing the transport and the part they
each take in it. The transporter controls needs representing pick-up and
delivery. The diagram shows how messages are routed between trans-
porter, consignor and consignee. Image by Kleedorfer, Busch, Huemer,
et al., 2016 is licensed under CC-BY-SA 4.0.
67
a conversation. Both sides use the need object to maintain a reference to that spe-
cific business transaction. In the case of general service offerings, however, this
pattern is not sufficient. A company offering transport services in a given region
could publish an object in WoN, receive hints and establish connections with its
clients using that object. This would lead to this object collecting more and more
connections over time, possibly impacting data management and application
performance. Therefore, we decided to enable a new interaction pattern: a need
that creates new needs programmatically for each specific case. In analogy to
the concept of a factory in software development, an object that creates other
objects, we call these special needs factory needs.
General service offerings can be represented as factory needs that are informed of
any new, potential transaction partner. For example, if the service to be offered is
a courier service, the factory need is informed whenever there is a potential client
who wants a package delivered. The client, however, should not be informed
of just any existing service announcement. They already stated their need and
expect to be contacted with concrete offers, not links to services that they might
want to explore. Upon being informed of the existence of a potential client, the
logic behind the factory need is activated and makes a concrete offering (possibly
taking into account situational factors such as current traffic, concurrent requests,
or available resources). This concrete offering is represented by the newly created
need object, which initiates a connection with the client’s need.
Such a scenario requires different matching logic than the symmetric case that
had been possible in WoN, a case in which both needs in a match are notified of
the match. In the case of factory needs, only the factory should be notified. This
requirement caused us to extend the way matching works in WoN such that by
default, both parties are notified, but a party can choose to suppress notification
for the counterpart as well as the notification for itself. This change was made
to the prototypical implementation and tested functionally. Though not a full
solution, the fact that one can now define that others should not be notified of
one’s needs, may contribute to easing the issue raised in Finding 4 (Privacy).
5.2 Information Requirements
As described in Finding 13 (wrong assumptions), the information used for match-

ing is not necessarily identical with the information needed to execute a transac-
tion. In some cases, all the needed information is present when the match occurs.
68
In other cases, additional questions need to be asked. The fact that the whole
data exchanged between participants, including their initial need descriptions is
available to both sides as RDF, it is possible to query this data in order to see if
the required information has already been stated or not. Based on that property
of the system, it is possible for a participant to define information requirements
in a declarative way. When a connection is made, the information requirements
are automatically checked and missing information is identified. User interfaces
can use this information to generate interface components (e.g., forms) to elicit
the missing information. User agents may be able to fill in missing information
automatically from personal data stores.
In order to realize a sytem that allows participants to define their information
requirements, we need to be able to formulate and check them automatically.
We plan to use SHACL (Knublauch and Kontokostas, 2017) for this aspect of WoN.
SHACL is a formalism for defining shapes that an RDF graph can be validated
against, the shapes themselves are written in RDF. This allows a need author to
include SHACL in the need description, and any party with access to the con-
versation content can evaluate the content against the shapes. Transmitting
the information requirements (shapes) in declarative syntax instead showing
a service-defined user interface (e.g. a HTML form) allows for great flexibility;
user agents can choose the user interface technology for eliciting the required
information from the user. Moreover, SHACL allows for referencing shapes from
any Web location, which allows for re-use of shapes and thus an avenue for stan-
dardization of distinct use cases, as well as the evolution of such standards. A
complete prototypical implementation is to follow.
5.3 Message Retraction
The functionality of information requirements has to operate on a higher level

than the level of simple message exchange - one might call it the level of meaning
exchange. Each new message contributes its content (RDF triples, grouped in
named RDF graphs) to the meaning of the conversation. Sometimes, participants
need to change the meaning, for example, to correct a mistake. Therefore, the
communication protocol allows for a retraction message that links to a message
already contained in the conversation, which is thereby marked as retracted. Each
participant can only retract their own messages that were sent earlier. Retract
messages themselves cannot be retracted. The effect of retraction is that the pay-
load of each retracted message is disregarded in further operations. This planned
69
feature would contribute to functionality that expects users to make and correct
mistakes, such as the issues raised in Findings 2 (digitalized communication), 15
(contract negotiation), 5 (appointments), and 11 (price negotiation).
5.4 Distributed Transactions and Long-Running Work
Especially in the context of multi-modal transport, it is required to coordinate

more than two parties in order sucessfully to plan and execute the transport. In
addition to multiple parties being involved, such an endeavour can span a time
of weeks or even months from initial planning to the final delivery. Both aspects
make the concept of long-running work (Bocchi, Laneve, and Zavattaro, 2003)
suitable for representing such processes. We decided to use the protocol logic
defined by WS-BusinessActivity (WS-BA) to model interconnected and possibly
nested activities of multiple business participants. As part of such a plan, a need
assumes the role of the WS-BA coordinator vis-a-vis the needs that represent
possible solutions for the problem. Vis-a-vis a higher-level aggregation node in
this plan, that need assumes the role of a WS-BA particpiant.
In order to realize such a need in WoN, in either role specified by WS-BA, it is
possible to use server-side logic that is invoked for an incoming message after the
mandatory logic realizing the basic protocol has been executed. The need thus
implements the state machine prescribed by the WS-BusinessActivity protocol
specification.
This system of states and state-changing messages is shared between partici-
pants and can at all times be used to determine and influence the state of the
distrbuted work. The option to add sub-goals to an overall plan may be one of the
building blocks for an integrated appointment making sub-protocol, addressing
the situation mentioned in Finding 5 (appointments), and it could be used to
realize linked conditional needs that find a one way feight and the return freight
simultaneously as required by Finding 9 (return freight). There is an experimental
implementation of this feature that will be overhauled.
5.5 Expressing Agreement
In order for transactions or long-running work to function legally, it is required

that the two parties involved set up a contract. Theoretically, a spoken agreement
70
6 Conclusion and Future Work
is enough to fulfil this requirement; in practice, though, contracts are written and
their contents are an important phase during negotiation. Finding 15 (contract
negotiation) hints toward the fact that a unified method of contract negotiation
may be beneficial for all participants.
The communication channel established by our technology between the par-
ticipants retains all messages exchanged in a manner that makes the message
history unambiguous and unchangeable. It is thus a suitable medium for contract
negotiation as it can always be established who said what in which order.
To allow for negotiation, we introduce the possibility to send a special message
that marks earlier messages as proposed. The other participant has the option to
send an accept message, referencing the proposal, thereby creating an agreement.
The agreement is identifiable by the URI of the accept message. For calculating
the agreement’s content, only the message history before the accept message is
considered. Consequently, retraction has no effect on agreements. They can be
canceled by either side by sending a message that proposes to cancel an agree-
ment and the counterpart accepting that proposal. At time of this writing, the
protocol has been designed, but not yet implemented.
6 Conclusion and Future Work
In this work we describe the current state of our iterative design process with
the overall goal of a de-centralized, open, Web based transportation infrastruc-
ture. We use expert interviews to identify business needs and assess problem
relevance. Addressing these needs, we propose a design for the most fundamen-
tal functionalities needed to achieve our overall goal. Two aspects relevant for
practical application have been left aside so far, namely payment information
integration and reputation management. We identified both aspects as relevant
at the outset of the project; our interviews arguably support these features as well:
Findings 1 (trust) and 12 (quality dimensions) quite directly point toward reputa-
tion management, and a number of business interactions (e.g. high-frequency
transport described in Finding 8 (framework agreements)) may be greatly im-
proved if payment was integrated in the communication instead of being handled
in an independent channel.
In line with the iterative approach, we are striving for applying the technology for
simple transport problems at first, while gathering requirements for more complex
71
situations and gradually implementing solutions for them. Practical evaluation

of the designed artifacts in real-world applications and follow-up interviews and
workshops are planned for later development phases.
Acknowledgements
This work would not have been possible if it had not been for invaluable ideas and
criticism contributed by the research team at Studio SAT, which currently includes
Heiko Friedrich, Fabian Suda, Kevin Singer, and Soheil Human. We are indebted
to Laura Daniele for sharing her knowledge of the i-Cargo ontologies and to the
domain experts who provided valuable insights during our interviews. Special
thanks go to Dominik Kovacs of DoKo Agromarketing for introducing us to the
community of agricultural commerce. We are also grateful for all contributions to
the discussion on the topic of negotiation in the semantic Web mailing list.3 .
Financial Disclosure
This work is funded by the Austrian Research Promotion Agency (FFG) under the
project Open Logistics Networks (OLN) of the Program COIN - Cooperation and
Innovation.
References
Airbnb (2017). URL: https://www.airbnb.com/.
Bizer, C., T. Heath, and T. Berners-Lee (2009). “Linked data-the story so far”. In: International Journal
on Semantic Web and Information Systems (IJSWIS) 5.3, pp. 1–22.
Bocchi, L., C. Laneve, and G. Zavattaro (2003). “A calculus for long-running transactions”. In: Formal
Methods for Open Object-Based Distributed Systems. Springer, pp. 124–138.
Daniele, L. (2013a). Logistics Core Ontology. TNO. URL: http://ontology.tno.nl/logico/.
Daniele, L. (2013b). Logistics Services Ontology. TNO. URL: http://ontology.tno.nl/logiserv/.
Daniele, L. (2013c). Transport ontology. TNO. URL: http://ontology.tno.nl/transport/.
Daniele, L. and L. Ferreira Pires (2013). “An ontological approach to logistics”. In: Enterprise Inter-
operability, Research and Applications in the Service-oriented Ecosystem, IWEI’13 Proceedings.
Ed. by M. Zelm, M. van Sinderen, and G. Doumeingts. ISTE Ltd, John Wiley & Sons, Inc., pp. 199–
213.
3
See https://lists.w3.org/Archives/Public/semantic-web/2017Jul/0004.html [2017/07/07]
72
References
DiCicco-Bloom, B. and B. F. Crabtree (2006). “The qualitative research interview”. In: Medical
education 40.4, pp. 314–321.
Garcia, A. (2015). iCargo - Intelligent Cargo in Efficient and Sustainable Global Logistics Operations.
ATOS Spain SA. URL: http://cordis.europa.eu/project/rcn/100869_en.html.
Hevner, A. R. (2007). “A three cycle view of design science research”. In: Scandinavian journal of
information systems 19.2, p. 4.
Hevner, A. R., S. T. March, J. Park, and S. Ram (2004). “Design science in information systems
research”. In: MIS quarterly 28.1, pp. 75–105.
Hofman, W., M. Punter, H. Bastiaansen, E. Cornelisse, S. Dalmolen, Z. Palaskas, B. Karakostas,
J. Gato, J. Garcia, et al. (2016). “An interorganizational IT infrastructure for self-organization in
logistics: situation awareness and real-time chain composition”. In: International Journal of
Advanced Logistics 5.2, pp. 101–115.
Kleedorfer, F., C. M. Busch, C. Huemer, and C. Pichler (2016). “A Linked Data Based Messaging Ar-
chitecture for the Web of Needs”. In: Enterprise Modelling and Information Systems Architectures
11.3, pp. 1–18.
Kleedorfer, F., C. M. Busch, C. Pichler, and C. Huemer (2014). “The Case for the Web of Needs”. In:
Business Informatics (CBI), 2014 IEEE 16th Conference on. Vol. 1. IEEE, pp. 94–101.
Kleedorfer, F., Y. Panchenko, C. M. Busch, and C. Huemer (2016). “Verifiability and Traceability in a
Linked Data Based Messaging System”. In: Proceedings of the 12th International Conference on
Semantic Systems. SEMANTiCS 2016. Leipzig, Germany: ACM, pp. 97–100.
Knublauch, H. and D. Kontokostas (2017). Shapes Constraint Language (SHACL) - Proposed Recom-
mendation. Tech. rep. [Last accessed on 2017/07/06].
Manola, F. and E. Miller (2004). RDF Primer. Tech. rep. [Last accessed on 2016/06/06].
Nakamoto, S. (2008). Bitcoin: A peer-to-peer electronic cash system.
Uber (2017). URL: https://www.uber.com/.
UNECE (1956). Convention on the Contract for the International Carriage of Goods by Road (CMR).
Tech. rep. Geneva: United Nations Economic Commission for Europe (UNECE).
UNECE (2008). Additional Protocol to the CMR concerning the electronic consignment note (e-CMR).
Tech. rep. Geneva: United Nations Economic Commission for Europe (UNECE).
Varvasovszky, Z. and R. Brugha (2000). “A stakeholder analysis”. In: Health policy and planning
15.3, pp. 338–345.
73
Lead-time Optimization Potential of
Digitization in Air Cargo
Benjamin Bierwirth1 , Kai-Oliver Schocke2
1 – Hochschule RheinMain
2 – Frankfurt University of Applied Sciences
The air cargo supply chain consists of several parties: the forwarders collect air
cargo shipments and consolidate these shipments in regional warehouses all
over the continent. They decide upon the export airport and bring it there to
consolidate all shipments in their air cargo hub - warehouse according to the
chosen flight. Using local truckers, the shipments, consolidated by airline, are
forwarded to the handling agents to load air cargo containers or to build up air
cargo pallets. The containers or pallets are brought to the apron and loaded
into the planes. There is not a single chain existing; the parties involved are
part of a dynamic n * m * o network which is fixed with every single shipment.
The problem addressed in this paper is to document the waiting times at the
level of the handling agents. Own, long time research at Frankfurt International
Airport shows, that the waiting time for the trucker in front of handling agents
warehouses significantly influence the lead-time for shipments. We identify the
barriers to a faster adaption of digitization, which is currently around 30% in
Frankfurt. Quantifying the impact of waiting times and required warehouse space
due to long lead-times allows for a cost estimate.
Keywords: Air Cargo; Lead-Time Reduction; Digitization; Air Cargo Supply Chain
75
Lead-time Optimization Potential of Digitization in Air Cargo
1 Introduction
1.1 Problem description
In this paper, we address the optimization potential in air cargo handling by using
electronic data exchange and the barriers that hinder a faster development of
digitization. Digitization is expected to be an important enabler to optimize the
physical shipment flow between forwarder and cargo handling agent and reduce
existing waiting times.
As waiting time costs accumulate to millions of euros per year and the existing IT
systems provide interfaces to exchange information between the participants of
the supply chain and other solutions like cargo community systems exist, we try
to identify the reasons for the slow adaption of digitization in the air cargo supply
chain, mainly for forwarders and cargo handling agents.
Based on a survey within the members of the air cargo community of a major
hub airport we can show that the benefits of electronic data exchange in the
air cargo supply chain are well understood, but other barriers hinder further
development.
1.2 Air Cargo Supply Chain
Air cargo is the fastest mode of transport, but also the most expensive one. It is
therefore used in emergencies or in the global supply chain of high value goods.
General air cargo has to be differentiated from express and parcel services – e.g.
from DHL or UPS – as these companies mostly use dedicated air and road transport
networks and facilities. The air cargo supply chain on the other hand consists
of several parties: the forwarders collect the shipments from the shipper and
consolidate these shipments in regional warehouses all over the continent. They
decide upon the export airport and bring it there to consolidate in their air cargo
hub-warehouse all shipments according to the chosen flight. Using local truckers,
the shipments by airline are forwarded to the handling agents to load air cargo
containers or to build up air cargo pallets. The containers or pallets are brought
to the apron and loaded into the planes (see figure 1).
76
1 Introduction
Figure 1: Air Cargo Supply Chain at Frankfurt Airport (here: export process)
1.3 Legal Situation between Forwarder und Handling Agent
The transportation is based on a contract between shipper and consignee where

the respective party – according to the chosen INCOTERMS – orders the forwarder
to organize the transportation. Besides of having outsourced various steps in the
ground transportation. The forwarder is responsible for contracting an airline. The
airline has their own contracts with handling agents at the respective airports (see
figure 1). This leads to common supply chain management problem that at the
handling agent’s warehouse to parties meet which do not have a direct contractual
relationship and therefor show only limited interest in mutual optimized solution.
In difference to many production or distribution supply chains, a leading partner
that forces its partners into a hierarchical supply chain management is missing.
With many partners of various sizes and a volatile setup, also a consensus based
supply chain approach is difficult to realize.
1.4 Digitalization in Air Cargo
The air cargo shipment has to be accompanied by many shipment papers, which
are necessary for e.g. declaration or description of the content. This paperwork
77
Figure 2: Contracts in the air cargo supply chain
78
1 Introduction
Figure 3: IATA eAWB development (International Air Transport Association, 2017)
could be done electronically. The respective factors (papers transmitted electron-

ically/papers in total) in Frankfurt differs from airline to airline, from forwarder to
forwarder, but does not exceed 30% so far. Penetration rate of electronic airway
bills (eAWB) just reached the global 50% mark and is lagging behind the expected
development. Several activities have been conducted, e.g. by the International
Air Transport Association (IATA), or the local collaboration platform fair@link to
increase the digitalization share, but did not show fundamental success by now
(see figure 3).
1.5 Weekend Peak
Around half of all cargo is transported on passenger planes, which often offer
a daily (or better) service to major destinations. Nevertheless, cargo volume
typically peaks on weekends. As most businesses (shippers and consignees) do
not operate on weekends the shipments arrive on Friday or Saturday latest for
export and the incoming shipments will be picked up Monday morning. The
weekend peak is especially challenging for the handling agents, as air cargo
shipments are typically booked on short notice (see figure 4).
1.6 Key Performance Indicators in Air Cargo
In air cargo handling, competitiveness can be defined by lead-times and related

costs. Lead-times directly influence the related costs as shorter lead-times re-
79
Figure 4: Weekly cargo volume distribution (n = 102,793 Airway Bills of March 2015
= 40% of Volume of Frankfurt Airport)
duce the number of stored shipments in the warehouse and – besides being less
complex – allow for an increase in efficiency.
The lead-time concerning this paper is defined to be the difference between the
arrival time of the truck at the office of the handling agent (landside) and the
transportation of the shipments to the plane (airside) for the export process. The
lead-time in imports is measured between the arrival of the shipments at the
handling agent’s warehouse on the airside and the pickup by the forwarder.
1.7 Structure of the paper
This paper is organized as follows: The related literature from the areas of air
cargo, supply chain management and operational logistics is presented in section
2. A detailed problem description follows in section 3. It is based on other research
about truck waiting times and average lead-times in handling agent warehouses
before detailing the theoretical benefits of digitization and potential barriers or
risks.
80
2 Related Literature
Section 4 contains the relevant information about the setup and structure of the
survey, which is followed by the presentation of the most important results in
section 5. These are transformed into a benefits-barriers-model. Section 5 ends
with a critical appraisal of survey and model. The last section contains a summary
and directions for further research.
2 Related Literature
2.1 Aviation and Air cargo
Research in the Aviation industry is mostly focused on the passenger transport

(Wittmer, Bieger and Müller, 2013).
Specifically in air cargo networks only few current research papers are found
(Sieke, 2010; Vancroonenburg, et al., 2014; Feng, Li and Shen, 2015a; Zhang, et al.,
2017) which contain mostly pure theoretical considerations without containing
real data.
Chen et al consider the choice of optimal transshipment-hubs (e.g. that of the

export port or the airport) from the shippers and forwarders perspective (Chen, et
al., 2017). Besides quantitative factors such as costs and transportation distance,
location-specific qualitative elements are taken into account as well. Feng (Feng,
Li and Shen, 2015b) gives an up-to-date overview of the air cargo operation.
The examined questions are about the optimization of the layout or seperate
processes (Frye, 2003; Chan, et al., 2006; Yan, Shih and Shiao, 2008; Bierwirth,
2009; Lau, et al., 2009; Selinka, Franz and Stolletz, 2016). Other authors examine
the Revenue Management in air cargo (Hellermann, 2006; Pfeffer, 2015; Chao and
Li, 2017).
There are some, partially unpublished, preparatory works by the authors that
talk about the investigation of the processes at Frankfurt Airport (Wagschal and
Schocke, 2012; Schocke, 2013; Hertelendy, et al., 2015; Wegener and Schocke,
2015; Bierwirth, Höhl and Winning, 2016; Schocke and Koch, 2017).
81
2.2 Supply Chain Management
Beside Air cargo specific processes research and publications regarding the ben-
efits of information exchange in supply chain management is relevant for the
deduction of digitization benefits.
The majority of the publications focuses on a classic Supply Chain and the coop-
eration of supplier, producer, distributor and examines the effects of fluctuations
in demand and the resulting costs (Schulze, 2009; Barros, et al., 2015; Marinagi,
Trivellas and Reklitis, 2015; Rached, Bahroun and Campagne, 2015). Walter (Wal-
ter, 2015) examines the “Exchange of information in the maritime transport chain
– investigation of process performance in the data flow of capacity utilization” in
very similar and comparable aspects. Aspects of empirical investigations were
transferred to air cargo based on this publication.
Only in Walter (Walter, 2015), the relations of associates inside the transport
chain (forwarders, carriers, other logistics service providers) were object of a
publication. The quantification of use mostly occurred over simulation models
and calculations and was only seldom linked to empirical analyses.
Consisting approaches from the SCM-theory prove that a corporate planning pol-
icy such as e.g. a mutual optimal order- and production policy can be determined
in ways of minimizing the sum of total costs from costumer and producer (Sucky,
2004).
2.3 Operational bottlenecks
In theory and practice, loading ramps are known to be a bottleneck of transport

logistics. Many practice-oriented publications of the recent years in this context
make it clear that friction losses occur in the interfaces of processing (Semmann,
2012). The site operating companies, here Handling Agents, aim for maximal
capacity of their ramps and a continuous flow of cargo (Semmann, 2012). Ramp
management concepts aim for an optimization of the use of the ramps. In practice,
the active management is often not transcribed in consideration of all boundary
conditions, which leads to efficiency losses. There losses can affect the daily
planning substantially of the site operating and cargo carrying companies and are
not only complained about in practice but also talked about in a special report of
the federal office for freight transport (Semmann, 2012).
82
3 Problem Description
Firstly, we describe results of the research about the status, which give an indica-
tion about the optimization potential. Afterwards the benefits of digitization and
electronic data exchange will be detailed and potential risks discussed.
3.1 Current status and deficits
According to proven supply chain management theory, the exchange of informa-

tion between supply chain partners – in our case the forwarder, the airline and the
handling agent – could accelerate handling as all parties can prepare in advance
and optimize their use of resources. The minimum improvement would be the
time saving of entering shipment or airway bill data and avoidance of printing and
paper costs. Beyond these simple references, a more detailed approach regarding
the benefits of electronic data exchange is missing. Besides formulating, the
urgent need for digitization related risks as well as needed competencies and
capabilities necessary for successful digitization are unknown.
Earlier research showed that the parties involved (handling agents and forwarders)
use IT systems, which are capable of exchanging information in various formats
and to a centralized platform as well as over bilateral interfaces.
An on-site observation of the truck docks of several handling agent warehouses

over a period of 36 hours at 1st and 2nd of June, 2017, led to the calculation of
average truck lead-times.
The overall lead-time for handling agent 1 is 186 minutes, for handling agent 2 209
minutes and for handling agent 3 is 164 minutes. Supposing an identical mix of
shipments and defining the shortest length per segment as a benchmark (e.g. 39
minutes for handling agent 2 company arrival - office entrance), the benchmarking
lead-time is 88mins. Consequently, significant waiting times can be calculated:
at least 50% of the overall lead-time is waiting time (see figure 5).
As part of the observation, the truck drivers have been asked, if they announced
their arrival (e.g. using a truck dock scheduling system) or if they have been
informed about the current operating grade at the ramp. Just 25% of the truck
driver use existing ramp scheduling systems and 25% are informed about current
operating grade. This happens although 80% of the truck drivers are equipped
83
Figure 5: Truck waiting times as part of lead-time
with a smart phone. Exchange of information does not happen and leads to long
waiting times.
Beside the cost of waiting times, the missing information about the shipments
loaded onto each truck some shipments miss their booked flights which causes
extra costs. Expectation of several partners of the air cargo supply chain is that
with electronic data exchange and a truck slot time management system these
backlogs (see chapter 5).
Apart from observing truck drivers, we collected a set of data from four handling
agents and Fraport from March 2015, which covers 40% of the overall tonnage.
With this set of data, we are able to calculate the lead-time from office entrance
at handling agent until off block (plane leaves its parking position).
The analysis of the lead-times of these 102,793 shipments show that while the
mean is almost a day (from delivery to airside departure and airside arrival to
pick up) most cargo is delivered or picked up shortly before latest acceptance
time (LAT) or after time of availability (TOA). It can be logically deducted that
shipments which arrive early or stay long in the handling agents warehouse add
to the complexity and may block resources which are needed for time critical
shipments. As with the time management system the expectation is the increase
84
the efficiency of the use of resources and faster processes at this external interface
with digitization (import: see figure 6; export see figure 7).
Figure 6: Import lead-times
Figure 7: Export lead-times
85
3.2 Benefits of digitization
As mentioned in the beginning, the overall expectation would be cost saving.

However, beside this and the above-mentioned expected increase in resource
efficiency further aspects have to be considered as benefits (see figure 8).
Based on a better planning which is made possible by information transmitted
prior to truck arrival processes can be aligned and thereby reducing waiting times,
so a faster process can be achieved (time). At the same time, the amount of time
for re-entering the data can be eliminated. The planning leads to a more efficient
use of staff and less waiting times or buffers between various process steps and
thereby accelerates the handling.
Reduced lead-times increase the use of truck docks and storage area, so overall
capacity of the facility can be increased. The resources involved are used more
efficiently at the same time and thereby the marginal costs are reduced.
The benefits regarding the cost perspective is threefold: As already mentioned the
cost for re-entering data can be eliminated. This also eliminates possible mistakes
and the related effort to correct those. Less mistakes than lead to less effort for
searching and/or questions by customers. Additionally the more efficient use of
staff leads to reduced costs.
Not only less mistakes are made, but also mistakes could be identified earlier (e.g.
if number of packages differ with documentation).
Last factor to be considered should be a positive effect on the image of a handling
agent as less mistakes may lead to higher customer satisfaction.
3.3 Risks and required competencies
As the benefits are easy to deduct, although not quantifiable, additional barriers
or other missing elements are hindering a faster adoption.
Based on several expert interviews and students’ thesis as well as the authors
experience over the years, the authors identified risks and competencies that
might be crucial for a successful digitization.
A very important aspect is trust in the partners and service providers or fear
about the misuse of information provided electronically. This mainly defines
the willingness of exchanging data. This aspect than is of course the ability to
86
Figure 8: Detailing the benefits of digitization in air cargo
exchange data, which requires technical and management skills – from connecting
the systems to explanation of the transformation to the employees.
87
4 Survey
The survey within the air cargo community was designed to firstly verify the
understanding of the formulated benefits of digitization. The second part deals
with the other aspects, the risks and competencies such as the digital readiness
of partners in the air cargo supply chain and their supply chain thinking.
As electronic data exchange requires at least two parties (sender and receiver),
the survey should be relevant for all parties in the air cargo supply chain, and not
specified for one group.
The survey consisted of 40 questions which were sent to members and interested
parties of Air Cargo Community Frankfurt (ACCF). In total 136 persons were ad-
dressed. Prior to that, the questionnaire was pre-tested within in the project team
and with selected members of the community. A return rate of 26% could be
achieved. The returning questionnaires cover the entire air cargo supply chain
(see figure 9)
In the beginning, the participant had to select the company size and its role in the
supply chain.
To evaluate the importance of model sub-elements the participants were re-
quested to give their consent or opposition to statements on a scale from 1 to
10 where 1 was defined and indicated as “I totally disagree” and 10 as “I totally
agree”. Additionally we integrated open questions and multi-selection questions,
e.g. for the type of electronic data exchange that is currently used.
88
4 Survey
Figure 9: Position in air cargo supply chain of respondent (n=50)
89
5 Results
5.1 Benefits of digitalization
In part one of the survey, the respondents have been asked for their view on the
benefits of digitalization.
Concerning the digital exchange of information, 85% of the respondents reported
to already use digital data exchange (see figure 10).
In most cases IT-systems are connected via electronic data interchange (EDI),
however, 24% of respondents said they mainly use e-mails with attachments.
The tools being used are separated into one third each between the local cargo
community system fair@link, bidirectional interfaces as well as other media (here
especially e-mail clients)
The advantages of electronic data exchange are clear to the respondents. Digi-
tal data exchange improves resource utilization (mean 8,96; STD 1,33), reduces
costs (mean 8,55; STD 1,64), improves data quality (mean ,67; STD 1,99) and
makes it possible to increase capacity (mean 7,70; STD 2,03). Data exchange also
strengthens the image of the company (mean 7.98; SRD 1,58).
5.2 Maturity of Supply Chain Thinking
The importance of general thinking as well as cooperation is being recognized

cleary by the companies: The up- and downstream companies in the supply chain
are seen as part of the own network (mean 7,58; STD 2,45; see figure 11)
As mentioned before companies see big advantages in electronic data exchange.
But there is an obvious gap between desire and reality: Although practically all
companies use electronic data exchange not all partner companies are attached
(mean 7,58; STD 2,45; see figure 12).
5.3 Data quality
Reasons for the poor penetration of electronic cooperation are widely varied (see
table 1).
90
5 Results
Figure 10: Usage of digital data exchange (n = 40)
Figure 11: Downstream and upstream companies are part of companies supply
chain
91
Figure 12: Direct connections between the IT systems of all our partners do exist
Table 1: Reasons for poor electronic penetration
Information we get is … Mean STD Median

In time 5,61 2,38 5,00
Complete 5,74 2,41 6,00
Reliable 5,91 2,31 6,00
Clear and easy to understand 6,14 2,26 6,00
Up to date 6,19 2,32 7,00
92
5 Results
The information that is transmitted to the partners is unsatisfactory in several

dimensions. Respondents see immense improvement potential in both punctual
delivery and completeness and currency. Especially the aspect of reliability of
data is seen critically. In addition, the rework of electronically transmitted data
needs to be corrected by 26% of the companies.
The advantage of electronic transmission to avoid system breakdowns cannot

be verified for sure; IT processes need to be checked manually which can lead to
additional costs.
5.4 Risks
The contradiction between wanted and implemented electronic exchange of data

can be shown by the means of two additional dimensions. On the one hand,
digitization leads to transparency, on the other hand digitization is seen critically
due to stronger dependence on IT.
The operative processes in the air cargo industry nowadays are marked by random-
ness and lack of transparency. The decision which airline sends a consignment
is a case-by-case decision, which e.g. contains extensive requests for different
airlines (see chapter 2). Through electronic linking of IT systems cargo data can
be retrieved directly. At the same time, inefficiencies on individual value levels
become visible to the costumer. The companies however become more transpar-
ent and therefore better comparable in terms of the temporal component of the
processes (see figure 13).
Therefore market transparency will increase, which most of the companies sug-
gest (mean 7,24; STD 1,79). This transparency is seen more as a threat than a
chance (mean 4,76; STD 2,74).
These statements are influenced by the thought that many IT projects are obvi-
ously unpredictable (mean 5,59; STD 2,58), the confidentiality of data can only be
ensured conditionally (mean 5,76; STD 2,76) and one becomes dependent on one
service provider due to linkage (mean 6,03; STD 2,78) or the connection is very
complex.
In conclusion, a very heterogeneous picture is shown regarding the realization of

digitization in the air cargo industry.
93
Figure 13: The market becomes more transparent
5.5 Benefits-Barrier-Model for digitization in air cargo
Based on the results of the survey a benefits-barriers-model was developed to

reflect the potential benefits of exchanging data and the barriers the might hinder
the successful implementation. Therefore, an inverted triangle was chosen to
reflect the three perspectives: The five types of benefits can only be realized if
the four competencies are present with a basic knowledge and the risk side can
be calculated against the benefit (see figure 14).
As mentioned in section 3.3, the risk perspective mainly influences the willingness
of a participant to share information. Related to the exchange or distribution of
information on an exchange platform additional risks arise:
— The business might become more transparent

— The confidentiality of the information has to be ensured
— The partner may become dependent of the platform and the information
provided
The (IT) project risk involves the uncertainty and lacking experience with projects
– esp. of smaller partners. This could be solved with external partners providing
94
5 Results
Figure 14: Benefit-Barriers-Model
this experience or by the exchange service provider who includes the project
management.
On the competence perspective, a partner has to possess steering and process

competencies. The information acquired earlier in the process has to be used
to improve planning for example; otherwise, no benefits can be realized. The
partners have to understand their IT system and the platform and how it integrates
into the overall process and for example, which mobile equipment is needed.
This is meant with IT competence. As commonly known digitization changes the
working environment so an accompanying change management is essential for a
successful installation.
5.6 Critical Appraisal
The model was developed based on a limited sample size of answers from just
one airport community. The maturity of competencies is hard to measure and
the model so far does not indicate which maturity level has to be achieved to take
the competence barrier.
95
Quantified calculations of the benefits, e.g. from a best practice, are necessary
to calculate a business case which is necessary in most companies to justify
investments and projects.
6 Conclusion
The air cargo supply chain consists of many partners with individual goals, inter-
ests and perspectives. The grade of digitalization in information exchange is poor
compared to other industries. We presented lead-time calculations as well as
waiting time considerations which show the inefficiency of todays processes. We
describe the discrepancy between these inefficiencies and the expectations of
the protagonists of the air cargo supply chain.
We conducted a survey which shows that the benefits of digitization and electronic
data exchange are understood and agreed upon by all participants in the air cargo
supply chain. A positive impact on image and quality can be expected as well as
reductions in cost and lead times. Resources can be used more efficiently.
The survey revealed a maturity in supply chain thinking although there is a huge
gap between the perception of supply chain thinking and the existing electronic
data exchange between the partners.
Beside predicting benefits from digitization and having a supply chain thinking
other barriers hinder the development of electronic data exchange. Managers
mostly fear the transparency of a digitized supply chain and lack some competen-
cies for making the digital transition.
The results of the survey led to the development of the benefits-barriers-model.

The benefits of digitization can only be realized if risks are low and other compe-
tencies exist.
Acknowledgements
The authors would like to thank the Air Cargo Community Frankfurt e.V., which
coordinated the contact to and the communication with the companies involved
into this research.
96
References
The present study was financially supported by the research-funding program

”HOLM Innovationsfonds” of Hesse’s Ministry of Higher Education, Research, and
Arts (HO 2014/40).
References
Barros, A. P. d., C. S. Ishikiriyama, R. C. Peres, and C. F. S. Gomes (2015). “Processes and Benefits of
the Application of Information Technology in Supply Chain Management: An Analysis of the
Literature”. In: Procedia Computer Science 55, pp. 698–705.
Bierwirth, B. (2009). Optimierung der Sicherheitskontrolle in der Luftfrachttransportkette aus logis-
tischer Sicht: Vortrag gehalten auf dem Kolloquium Luftfracht und Luftsicherheit, 13. November
2009 Frankfurt/Main. Praxiswissen.
Bierwirth, B., S. Höhl, and J. v. Winning (2016). Electronic Air Cargo Supply Chain. Frankfurt.
Chan, F. T., R. Bhagwat, N. Kumar, M. K. Tiwari, and P. Lam (2006). “Development of a decision
support system for air-cargo pallets loading problem: A case study”. In: Expert Systems with
Applications 31.3, pp. 472–485.
Chao, C.-C. and R.-G. Li (2017). “Effects of cargo types and load efficiency on airline cargo revenues”.
In: Journal of Air Transport Management 61, pp. 26–33.
Chen, G., W. Cheung, S.-C. Chu, and L. Xu (2017). “Transshipment hub selection from a shipper’s
and freight forwarder’s perspective”. In: Expert Systems with Applications 83, pp. 396–404.
Feng, B., Y. Li, and H. Shen (2015). “Tying mechanism for airlines’ air cargo capacity allocation”. In:
European Journal of Operational Research 244.1, pp. 322–330.
Feng, B., Y. Li, and Z.-J. M. Shen (2015). “Air cargo operations: Literature review and comparison
with practices”. In: Transportation Research Part C: Emerging Technologies 56, pp. 263–280.
Frye, H. (2003). Flächenbezogene Optimierung von Luftfrachtterminals: Zugl.: Dortmund, Univ., Diss.,
2002. Logistik für die Praxis. Dortmund: Verl. Praxiswissen.
Hellermann, R. (2006). Capacity options for revenue management: Theory and applications in the
air cargo industry ; with 17 tables. Vol. Vol. 575. Lecture notes in economics and mathematical
systems. Berlin, Heidelberg, and New York: Springer.
Hertelendy, T., T. Gerhardt, H. Frye, and K.-O. Schocke (2015). Datenmodell für Monitoring und
Benchmarking der Cargo City Frankfurt.
International Air Transport Association (2017). e-AWB international monthly report May 2017.
Lau, H., T. M. Chan, W. T. Tsui, G. Ho, and K. L. Choy (2009). “An AI approach for optimizing multi-
pallet loading operations”. In: Expert Systems with Applications 36.3, pp. 4296–4312.
Marinagi, C., P. Trivellas, and P. Reklitis (2015). “Information Quality and Supply Chain Performance:
The Mediating Role of Information Sharing”. In: Procedia - Social and Behavioral Sciences 175,
pp. 473–479.
Pfeffer, S. (2015). “Revenue Management in der Luftfracht”. Dissertation. Stuttgart: Universität
Stuttgart.
97
Rached, M., Z. Bahroun, and J.-P. Campagne (2015). “Assessing the value of information sharing
and its impact on the performance of the various partners in supply chains”. In: Computers &
Industrial Engineering 88, pp. 237–253.
Schocke, K.-O. (2013). Cargo City 2020. Frankfurt.
Schocke, K.-O. and S. Koch (2017). Analyse der Verkehrssituation in der CargoCity unter spezieller
Berücksichtigung der Tor 26/31/32. Frankfurt.
Schulze, U. (2009). Informationstechnologieeinsatz im Supply-chain-Management: Eine konzep-
tionelle und empirische Untersuchung zu Nutzenwirkungen und Nutzenmessung: Zugl.: Vallendar,
WHU - Otto Beisheim School of Management, Diss., 2009. 1. Aufl. Vol. Bd. 10. Gabler Research.
Wiesbaden: Gabler.
Selinka, G., A. Franz, and R. Stolletz (2016). “Time-dependent performance approximation of truck
handling operations at an air cargo terminal”. In: Computers & operations research : and their
applications to problems of world concern : an international journal 65 (2016), pp. 164–173.
Semmann, C. (2012). “Eiszeit an der Rampe”. In: DVZ, pp. 1–2.
Sieke, H. (2010). Ressourcenallokation im Luftfrachttransportnetz mit Hilfe einer standortüber-
greifenden Frachtflusssteuerung: Zugl.: Diss., Technische Universität Dortmund, 2010. Luftfahrt.
München: Verlag Dr. Hut.
Sucky, E. (2004). Koordination in Supply Chains: Spieltheoretische Ansätze zur Ermittlung integrierter
Bestell- und Produktionspolitiken. Produktion und Logistik. Wiesbaden: Dt. Univ.-Verl.
Vancroonenburg, W., J. Verstichel, K. Tavernier, and G. Vanden Berghe (2014). “Automatic air cargo
selection and weight balancing: A mixed integer programming approach”. In: Transportation
research / E 65 (2014), pp. 70–83.
Wagschal, H. and K.-O. Schocke (2012). Logistische Verkehrsstrukturanalyse am Frankfurter Flughafen.
Frankfurt.
Walter, F. (2015). Informationsaustausch in der maritimen Transportkette: Untersuchung der Prozessleis-
tung im Datenfluss und der Kapazitätsauslastung. Wiesbaden Germany: Springer Gabler.
Wegener, K. and K.-O. Schocke (2015). Verkehrsanalyse Cargo City Süd. Frankfurt.
Wittmer, A., T. Bieger, and R. Müller, eds. (2013). Aviation Systems: Management of the Integrated
Aviation Value Chain. Springer Texts in Business and Economics. Berlin: Springer Berlin.
Yan, S., Y.-L. Shih, and F.-Y. Shiao (2008). “Optimal cargo container loading plans under stochastic
demands for air express carriers”. In: Transportation Research Part E: Logistics and Transporta-
tion Review 44.3, pp. 555–575.
Zhang, C., F. Xie, K. Huang, T. Wu, and Z. Liang (2017). “MIP models and a hybrid method for the
capacitated air-cargo network planning and scheduling problems”. In: Transportation Research
Part E: Logistics and Transportation Review 103, pp. 158–173.
98
Investigating the Factors Influencing
the Acceptance of Fully Autonomous
Cars
Ahmed Ziad Benleulmi1 , Thorsten Blecker1
Once thought of as a product of science fiction, self-driving cars are discussed

today as an unavoidable means towards improving transportation systems. In
fact, many car manufacturers have announced their plans to deploy highly au-
tonomous cars as soon as 2020; according to the Society of Automotive Engineers
(SAE) these vehicles are capable of reacting “even if the human driver does not
respond appropriately to the request to intervene (SAE level 4)”. There is however
a long way to go before fully autonomous cars (SAE level 5) - where pedals and
steering wheels are forgone and limitations to driving during severe weather or
in unmapped areas are surmounted - are produced. Herein, the overall aim is to
study the drivers and inhibitors of autonomous cars’ acceptance across cultures
with a special focus on the different risks that might deter consumers from using
highly and/or fully autonomous cars. After an extensive reviewing of previous
works, a research model based on UTAUT2 was developed and accordingly an
online survey was conducted in the US and in Germany; 313 valid answers were
collected and analyzed. The findings presented here have serious implications
both on the academic field as well as the industry, especially in regards to the
roles that risks, culture and gender play in the acceptance of fully autonomous
cars.
Keywords: Fully Autonomous Cars; Technology Acceptance Models; PLS SEM
99
Investigating the Factors Influencing the Acceptance of Fully Autonomous Cars
1 Introduction
According to the World Health Organization (WHO), car accidents rank within
the top causes of death of people worldwide and rank first if we only consider
young people aged 15 to 29 (WHO, 2015); In 94% of the cases, the fault did not
lie with the vehicle but with the human drivers, be it drunkenness, drowsiness or
distraction (NHTSA, 2015); Sadly, even though the price paid every year is too high,
the number of casualties is expected to rise as a result to more people embracing
car ownership (Lipson and Kurman, 2016).
Supporters of a rapid adoption and quick mainstreaming of autonomous cars
believe that their benefits far outweigh their disadvantages; these beliefs are far
from baseless as many facts seem to support them; for instance, according to the
Eno Center for Transportation (Eno), if 90% of the driven cars in the USA were
autonomous “the number of driving related deaths would fall from 32400 per
year to 11300“ (Eno, 2013); Other major benefits of autonomous cars are their
convenience, ease of use and the freedom they offer to consumers, especially to
the old, disabled and those incapable of driving.
However, benefits of autonomous cars notwithstanding, they are far from being
perfect or at the very least they present many issues that make it hard for con-
sumers to accept them wholeheartedly. In fact many are reluctant to use them
for a multitude of reasons; some do not trust them to be safe, secure or private,
others avoid them because of social pressure and some consumers have trouble
accepting them simply because they love driving.
In order to have a deep understanding of the drivers and inhibitors governing the
consumers‘ attitudes towards autonomous cars‘ acceptance, we investigated the
main probable drivers; the proposed research model comprises constructs from
UTAUT2 as well as other relevant constructs that are rooted in the literature and
strongly relevant to the context of this research.
2 Theoretical Background and Hypotheses
Technology acceptance models have been around since the early days of infor-
mation system research, their aim –unchanging over the years- is to investigate
the factors influencing the adoption of a technology or its rejection; Several such
models emerged over time: TRA (1975), TPB (1985), TAM (1989) and UTAUT (2003)
100
to mention some of them (Venkatesh et al., 2003). UTAUT2, an extended ver-

sion of UTAUT proposed by Venkatesh in 2012, presents a fitting basis for our
present research; some of its constructs, pertaining to usefulness (Performance
Expectancy), ease of use (Effort Expectancy), social pressure (Social influence)
and enjoyment (Hedonic Motivation), should play a major role considering the
nature of the present research (Venkatesh, Thong and Xu, 2012). Certainly, the
next step is to adapt the theory to the current context; Hong et al (2014) clearly
defined the approaches for the contextualization of a theory (Hong et al., 2014);
the first level of this process is to add or remove core constructs; Next, contextual
factors such as antecedents are incorporated in the model. Following these guide-
lines we assimilated some key constructs –rooted in the literature- into the model
(see figure 1); these constructs are Desirability of Control (DEC) defined as “the
fear of losing control over the vehicle” (Planing, 2014), Perceived Convenience
(PC) which is “the level of convenience toward time, place and execution that
one feels when driving an autonomous car” (Hsu and Chang, 2013), Personal
Innovativeness in IT (PIIT) defined as an “individual trait reflecting a willingness
to try out any new technology” (Agarwal and Prasad, 1998) and the Intention to
Prefer an autonomous car over a conventional car (IP). The hypotheses regarding
the influence of the previously mentioned constructs on the Behavioral Intention
(BI), i.e. the intention of the consumer to use fully autonomous cars, are the
following:
H1: PIIT has a positive influence on BI

H1: DEC has a negative influence on BI
H3: HM has a positive influence on BI
H4: PE has a positive influence on BI
H5: EE has a positive influence on BI
H6: SI has a positive influence on BI
H7: PC has a positive influence on BI
H8: BI has a positive influence on IP
Additionally, risks are expected to play a major role inhibiting the acceptance of
autonomous cars. we singled out five relevant types of risks, these risks are (1)
Privacy Risk (PRIV) linked to a “possible loss of privacy as a result of a voluntary
or surreptitious information disclosure to the autonomous car” (Dinev and Hart,
2006; Liao, Liu and Chen, 2011), (2) Performance Risk (PERR) associated with “The
101
Figure 1: Proposed research model
102
possibility of the autonomous car malfunctioning and not performing as it was

designed and advertised and therefore failing to deliver the desired benefits.”
(Grewal, Gotlieb and Marmorstein, 1994), (3) Safety Risk (SAFE) which is the risk
of the user’s safety being endangered through his use of an autonomous car, (4)
Financial Risk (FINR) pertaining to “the potential monetary outlay associated
with the initial purchase price as well as the subsequent maintenance cost of
autonomous cars” and finally (5) Socio-psychological Risk (SPR) defined as the
“Potential loss of status in one’s social group as a result of using fully autonomous
cars, looking foolish or untrendy and risking to lower the consumer’s self image”
(Kim, Lee and Jung, 2005). Understanding the influence risks have on trust is
a major point in this research as it will not only show which risks are relevant
but also which ones have more impact. Trust (TRUST) is defined by Mayer et al.
(2011) as “the willingness of a party to be vulnerable to the actions of another
party based on the expectation that the other will perform a particular action
important to the truster, irrespective of the ability to monitor or control that
other party”. We believe the disposition of a person to trust might also play a role
in him/her trusting autonomous cars, hence the DTRUST construct, defined as
“how a person sees himself/herself in regards to his/her interactions with other
people”(Srivastava, Singh and Srivastava, 2013).
The hypotheses pertaining to trust are the following:
H9: PERR has a negative influence on TRUST
H10: FINR has a negative influence on TRUST
H11: PRIV has a negative influence on TRUST
H12: SAFE has a negative influence on TRUST
H13: SPR has negative influence on TRUST
H14: DTRUST has a positive influence on TRUST
H15: TRUST has a positive influence on BI
103
3 Methodology
3.1 Sample Description
An online study was conducted in Germany and the USA to collect the necessary
data for the present study. The questionnaire was designed following the well-
established principles for survey design by Dillman, Tortora and Bowker (1998).
The survey took approximately 10 minutes to complete and was accessible for
three weeks starting from May 31st, 2017.
313 participants answered the survey of which 160 reside in the USA and 153 in
Germany. The survey was designed in a way that participants had to answer all
questions before they were able to submit the questionnaire.
Age distribution shows the mean age of the respondents to be 35; it also shows
41.9% of them to be male and 58.1% to be female (see table 1).
The majority of the participants have an annual income inferior to €50.000. 95.5%
of the respondents have had some experience driving cars and 84.3% of them
currently own one.
104
Table 1: Participant demographics
Variable Category Freq In% Mode

Gender Male 131 41.9
Female
Female 182 58.1
Age Younger than 35 175 55,9
29
35 and older 138 44.1
Education 8th grade or less 2 0.6
Some high school (Grade 9-11) 14 4.5
Graduated from college,
Graduated from high school 81 25.9
graduate or post-graduate
1-3 years of college/university 74 23.6
school
Graduate or postgraduate 138 44.1
No answer 1 0.3
Annual Income Less than €50.000 157 50.2
€50.000 to €100.000 88 28.1
€100.000 to €150.000 14 4.5 Less than €50.000
€150.000 and more 3 1
I would rather not say 51 16.3
Car Owner- ship Yes 264 84.3
Yes
No 49 15.7
3 Methodology
Driving Experience Yes 299 95.5
Yes
No 14 4.5
105
3.2 Data Analysis
For the purpose of data analysis, we opted for the PLS SEM approach. PLS SEM
also known as PLS Path Modeling presents a good choice when the purpose of the
study is prediction rather than confirmation and when the analyzed samples are
small (Reinartz, Haenlein and Henseler, 2009). There are two parts to PLS SEM;
the first part is where the measurement (outer) model is assessed to determine its
reliability and validity. The structural (inner) model assessment is performed in
the second part of the analysis, the aim of which is to determine the significance
of the paths and R2 values (Anderson and Gerbing, 1988). For the purpose of this
analysis SmartPLS 3.0 Professional was used.
3.3 Measures
All questionnaire items were measured on a seven-point Likert scale ranging from
“strongly disagree” to “strongly agree”. The DEC, HM, SI, BI, DTRUST, PRIV and
SPR constructs were measured with three items; EE, PE, TRUST, PERR and FINR
with two items each, SAFE was the only construct measured with four indicators
and IP was the only single item construct in the model.
4 Results
4.1 Assessment of the Reliability and the Validity of the

Measurement Model
First we examine the indicators‘ reliability, according to Chin (1998) and Hulland
(1999) indicator reliability is established if each indicator presents a loading value
of 0.70 or higher. Our results demonstrated that the items’ loadings were generally
satisfactory.
Next, internal consistency reliability is in turn assessed, Hair et al. (2014) advise to
use composite reliability (CR) as a measure for it; a CR of 0.70 or higher is generally
regarded as acceptable for research (Bagozzi and Yi, 1988; Chin, 2010; Henseler,
Ringle and Sinkovics, 2009; Vinzi, Trinchera and Amato, 2010); our results fully
satisfy the required threshold as all estimated values were recorded to be higher
than 0.70 (see table 2).
106
4 Results
As a result to the first two assessments, the models’ reliability is established.

Turning to convergent validity, we assessed the AVE values for all our datasets;
the values we registered were all higher than 0.50, which is the required threshold
for this measure.
Finally, we used the Heterotrait-Monotrait Ratio of Correlations (HTMT) criterion
(see table 3) to detect possible discriminant validity issues in the model. According
to Hair et al. (2014) HTMT values lower than 0.90 signify the model to be clear of
such issues; this condition is fulfilled in all datasets as shown in table 3, hence,
discriminant validity is satisfied.
Table 2: Assessment results of the measurement model
Construct Item Loading CR AVE

Bi1 0.95
BI Bi2 0.94 0.96 0.89
Bi3 0.94
Dec1 0.88
DEC Dec2 0.66 0.84 0.63
Dec3 0.83
Dtrust1 0.86
DTRUST Dtrust2 0.87 0.89 0.73
Dtrust3 0.83
Ee1 0.91
EE 0.91 0.84
Ee2 0.92
Finr1 0.78
FINR 0.86 0.76
Finr2 0.95
Hm1 0.93
HM Hm2 0.94 0.95 0.86
Hm3 0.91
Pc1 0.91
PC 0.93 0.86
Pc2 0.95
Continued on next page
107
Table 2 – continued from previous page

Construct Item Loading CR AVE
Pe1 0.91
PE 0.92 0.86
Pe2 0.95
IP Ip1 1.00 1.00 1.00
Perr1 0.87
PERR 0.89 0.81
Perr2 0.93
Piit1 0.92
PIIT 0.92 0.84
Piit2 0.92
Priv1 0.87
PRIV Priv2 0.95 0.94 0.84
Priv3 0.93
Safe1 0.87
Safe2 0.88
SAFE 0.90 0.68
Safe3 0.77
Safe4 0.78
Si1 0.93
SI Si2 0.96 0.96 0.89
Si3 0.94
Spr1 0.69
SPR Spr2 0.81 0.84 0.64
Spr3 0.89
Trust1 0.90
TRUST 0.85 0.74
Trust2 0.82
108
Table 3: Discriminant validity assessment – HTMT
DEC DTRUST EE FINR HM IP PC PE PERR PIIT PRIV SAFE SI SPR TRUST

SPR 0.08
SI 0.37 0.50
SAFE 0.19 0.08 0.51
PRIV 0.62 0.05 0.12 0.25
PIIT 0.12 0.18 0.31 0.23 0.61
PERR 0.24 0.36 0.72 0.20 0.31 0.57
PE 0.31 0.39 0.04 0.19 0.42 0.07 0.67
PC 0.76 0.41 0.33 0.06 0.19 0.36 0.09 0.60
IP 0.59 0.61 0.53 0.41 0.17 0.40 0.52 0.15 0.67
HM 0.62 0.62 0.60 0.36 0.38 0.11 0.25 0.44 0.16 0.67
FINR 0.20 0.37 0.31 0.17 0.65 0.24 0.28 0.71 0.26 0.21 0.41
EE 0.33 0.58 0.63 0.69 0.74 0.46 0.44 0.14 0.34 0.39 0.05 0.76
DTRUST 0.30 0.20 0.24 0.20 0.18 0.23 0.14 0.40 0.10 0.14 0.31 0.09 0.46
DEC 0.09 0.42 0.56 0.34 0.63 0.42 0.34 0.69 0.16 0.35 0.64 0.19 0.11 0.41
BI 0.51 0.33 0.70 0.47 0.69 0.83 0.63 0.62 0.55 0.58 0.19 0.42 0.60 0.14 0.80
4 Results
109
4.2 Assessment of the Reliability and the Validity of the Structural

Model
Moving on to the assessment of the validity and reliability of the structural model,
first and foremost, we assessed the Variance Inflation Factor (VIF); our findings
were that all VIF estimates were smaller than 5; indicating the absence of collinear-
ity issues.
Next, we assessed the significance of path coefficients by calculating t-values; we
performed a bootstrapping procedure and used the settings recommended by
Hair et al. (2014) (5000 bootstraps, two-tailed tests). The significance thresholds
for the t-values are 2.58 for 99% confidence level, 1.96 for 95% confidence level
and 1.75 for a 90% confidence level (Hair et al. 2014).
The results of our analysis (displayed in figure 2) show that P IIT > BI, DEC >
BI, HM > BI, SI > BI, EE > BI, T RU ST > BI, DT RU ST >
T RU ST , F IN R > T RU ST , P ERR > T RU ST and BI > IP are signifi-
cant at 99% confidence level (t-value > 2.58); they also show SAF E > T RU ST
to be significant at 95% confidence level (t-value > 1.96) and SPR > TRUST at
90% confidence level (t-value > 1.75). The relationships that did not satisfy
the minimum requirements for significance were P E > BI, P C > BI, FINR
> TRUST and PRIV > TRUST. As a result Hypotheses H4, H7, H10 and H11 are
rejected.
The results for the R2 calculations (displayed in table 4) show that all endogenous
variables are well explained by their relationships. The highest R2 was estimated
at 0.71 for BI, followed by IP (0.64) and finally TRUST (0.32). In the field of consumer
behavior, a R2 of 0.20 is usually regarded as high (Hair, Ringle and Sarstedt, 2011;
Hair et al., 2014).
In order to perform a group moderation, we used the Multi Group Analysis algo-
rithm available in Smartpls3.0; it allowed us to have an better understanding of
whether people of different cultures (USA and Germany), ages (<=35 and >35)
and genders (male and female) would approach autonomous cars‘ acceptance
differently, the results are visible in table 4.
Gender and culture were found to play an active role as moderators contrary to
age. Our results show that women are more likely to prefer using an autonomous
car if they choose to use it; they also show women to be more influenced by their
social environment than men; these latter’s dispositions to trust were found to
play a more active role in them trusting autonomous cars. In regards to culture as
110
4 Results
Figure 2: Path coefficients and significance level of the relationships
111
Table 4: Group moderation - results of the parametric test
Path/Moderator Gender Age Country

BI ->IP 2.03** 0.78 0.29
DEC ->BI 0.63 1.39 0.29
DTRUST ->TRUST 2.24** 0.99 0.70
EE ->BI 1.06 1.14 0.10
FINR ->TRUST 1.65 0.18 0.49
HM ->BI 1.94* 0.27 0.82
PC ->BI 0.33 1.02 0.00
PE ->BI 0.53 0.82 0.33
PERR ->TRUST 0.55 0.96 0.47
PIIT ->BI 0.59 0.07 0.34
PRIV ->TRUST 0.4 0.29 1.36
SAFE ->TRUST 0.34 0.06 0.74
SI ->BI 2.24** 0.12 0.81
SPR ->TRUST 1.22 1.55 2.59***
TRUST ->BI 0.49 0.53 0.25
Significance level: *p<.10 **p<.05 ***p<.01 Confidence level: *90% **95% ***99%
112
5 Discussion and Conclusion
a moderator, findings show that socio-psychological risks are strong in Germany

contrary to the USA where they were found to be insignificant.
It is worth noting that in regards to effect sizes (f2 values), a strong effect size was
recorded for BI > IP (1.79); SI > BI (0.17) had the only registered medium effect
and all the remaining effects were small, with the exception of FINR > TRUST, PC
> BI, PE > BI and PRIV > TRUST (the rejected hypotheses) where the recorded
effect sizes were inferior to 0.02 and therefore deemed insignificant.
5 Discussion and Conclusion
The main aim of this study was to identify the drivers and inhibitors governing the
consumer‘s behavior in regards to the acceptance of fully autonomous cars.
For that purpose, we conducted an online survey in two countries namely Ger-
many and the USA; the collected data was then analyzed using PLS SEM. The
results of the analysis shed light on the main drivers influencing consumers’ in-
tention to use fully autonomous cars; they also allowed to gauge the role these
drivers play in terms of impact.
Our findings showed that many factors positively influence the user’s intention
to use an autonomous car; and while some of these factors are related to the
technology itself such as it being “easily used” or “enjoyable”, others are associ-
ated with the user’s own personality such as “personal innovativeness” or the
influence of the environment on him/her; The key inhibitors to autonomous cars’
acceptance were found to be the consumers’ thirst for control and risks; Some
risks were recorded to have a great influence on a consumer trusting a car, these
risks are: the risk of the car being unsafe and the risk of the car malfunctioning.
Surprisingly, the performance expectancy of the autonomous cars as well as their
convenience were found not to be significant, this can be explained as consumers
being more influenced by the negative aspect i.e. the car malfunctioning as well
as the unavailability of the car in the market at the present time. Risks that were
ascertained to be extraneous are the financial risk as well as the privacy risk as
many people did not find them to be “deal breakers”.
In terms of implications for theory, the contributed research model comprises
all the key elements that play a major role in the acceptance or rejection of fully
autonomous cars; as such it can be used in future research as a reference for a con-
textualized model of technology acceptance in the automobile industry. In terms
113
of managerial and practical implications, the current study presented findings

that clarified which elements have an impact -and to which level- making it easier
for stakeholders to make a better advised decision when dealing the inhibitors to
the acceptance of autonomous cars or when promoting the attractive qualities of
these vehicles. Advertizing the benefits of the car should help people have a better
understanding of what to expect in terms of productivity and comfort. Decision
makers should find ways to promote trust in fully autonomous vehicles, and that
through dealing with the consumer concerns’ linked to the performance of the car
and to the safety of the user, as these were found to be the strongest inhibiting risk
factors; they should also find ways to soften and implicate the people who prefer
to have constant control while commuting. Findings related to gender, culture
and personal innovativeness should ultimately help devise improved strategies
to target these specific population segments in a more effective way.
References
Agarwal, R. and J. Prasad (1998). “A Conceptual and Operational Definition of Personal Innova-
tiveness in the Domain of Information Technology”. In: Info. Sys. Research 9.2, pp. 204–215.
Anderson, J. C. and D. W. Gerbing (1988). “Structural equation modeling in practice: A review and
recommended two-step approach.” In: Psychological bulletin 103.3, p. 411.
Bagozzi, R. P. and Y. Yi (1988). “On the evaluation of structural equation models”. In: Journal of the
academy of marketing science 16.1, pp. 74–94.
Chin, W. W. (1998). “The partial least squares approach for structural equation modeling.” In:
George A. Marcoulides. Chap. 10, pp. 295–336.
Chin, W. W. (2010). “How to Write Up and Report PLS Analyses”. en. In: Handbook of Partial Least
Squares. Ed. by V. E. Vinzi, W. W. Chin, J. Henseler, and H. Wang. Springer Handbooks of Compu-
tational Statistics. Springer Berlin Heidelberg, pp. 655–690.
Dillman, D. A., R. D. Tortora, and D. Bowker (1998). “Principles for constructing web surveys”. In:
Joint Meetings of the American Statistical Association.
Dinev, T. and P. Hart (2006). “An Extended Privacy Calculus Model for E-Commerce Transactions”.
en. In: Information Systems Research 17.1, pp. 61–80.
Eno (2013). Preparing a Nation for Autonomous Vehicles. Tech. rep.
Grewal, D., J. Gotlieb, and H. Marmorstein (1994). “The Moderating Effects of Message Framing
and Source Credibility on the Price-Perceived Risk Relationship”. en. In: Journal of Consumer
Research 21.1, p. 145.
Hair, J. F., C. M. Ringle, and M. Sarstedt (2011). “PLS-SEM: Indeed a Silver Bullet”. In: The Journal of
Marketing Theory and Practice 19.2, pp. 139–152.
Hair, J. F., G. T. M. Hult, C. Ringle, and M. Sarstedt (2014). A primer on partial least squares structural
equations modeling (PLS-SEM). English. Los Angeles: SAGE.
114
References
Hair, J. F., M. Sarstedt, L. Hopkins, and V. G. Kuppelwieser (2014). “Partial least squares structural
equation modeling (PLS-SEM): An emerging tool in business research”. en. In: European Business
Review 26.2, pp. 106–121.
Henseler, J., C. M. Ringle, and R. R. Sinkovics (2009). “Advances in International Marketing”. en. In:
Advances in International Marketing 20. Ed. by J. Henseler, C. M. Ringle, and R. R. Sinkovics.
Hong, W., F. K. Y. Chan, J. Y. L. Thong, L. C. Chasalow, and G. Dhillon (2014). “A Framework and
Guidelines for Context-Specific Theorizing in Information Systems Research”. en. In: Information
Systems Research 25.1, pp. 111–136.
Hsu, H.-h. and Y.-y. Chang (2013). “Extended TAM Model: Impacts of Convenience on Acceptance
and Use of Moodle.” In: Online Submission 3.4, pp. 211–218.
Hulland, J. (1999). “Use of partial least squares (PLS) in strategic management research: a review
of four recent studies”. en. In: Strategic Management Journal 20.2, pp. 195–204.
Kim, H. K., M. Lee, and M. Jung (2005). “Perceived Risk and Risk-Reduction Strategies For High-
Technology Services”. In: 6, pp. 171–179.
Liao, C., C.-C. Liu, and K. Chen (2011). “Examining the impact of privacy, trust and risk perceptions
beyond monetary transactions: An integrated model”. en. In: Electronic Commerce Research
and Applications 10.6, pp. 702–715.
Lipson, H. and M. Kurman (2016). Driverless: intelligent cars and the road ahead. Cambridge,
Massachusetts: The MIT Press.
Mayer, P., D. Volland, F. Thiesse, and E. Fleisch (2011). “User Acceptance of’Smart Products’: An
Empirical Investigation”. In:
NHTSA (2015). Critical Reasons for Crashes Investigated in the National Motor Vehicle Crash Causa-
tion Survey.pdf. Tech. rep. NHTSA’s National Center for Statistics and Analysis.
Planing, P. (2014). Innovation Acceptance. en. DOI: 10.1007/978-3-658-05005-4. Wiesbaden: Springer
Fachmedien Wiesbaden.
Reinartz, W., M. Haenlein, and J. Henseler (2009). “An empirical comparison of the efficacy of
covariance-based and variance-based SEM”. en. In: International Journal of Research in Market-
ing 26.4, pp. 332–344.
Srivastava, R. K., S. Singh, and V. Srivastava (2013). “The Role of Trust and Technology Acceptance
Model(TAM) on Customer Acceptance of Mobile Banking: An Integrated Model to Improve Mobile
Banking in Banking Sector-A Conceptual Paper”. en. In: International Journal of Asian Business
and Information Management 4.1, pp. 31–43.
Venkatesh, V., M. G. Morris, G. B. Davis, and F. D. Davis (2003). “User acceptance of information
technology: Toward a unified view”. In: MIS quarterly, pp. 425–478.
Venkatesh, V., J. Y. Thong, and X. Xu (2012). “Consumer acceptance and use of information tech-
nology: extending the unified theory of acceptance and use of technology”. In: MIS quarterly
36.1, pp. 157–178.
Vinzi, V. E., L. Trinchera, and S. Amato (2010). “PLS Path Modeling: From Foundations to Recent
Developments and Open Issues for Model Assessment and Improvement”. en. In: Handbook
of Partial Least Squares. Ed. by V. E. Vinzi, W. W. Chin, J. Henseler, and H. Wang. Springer
Handbooks of Computational Statistics. Springer Berlin Heidelberg, pp. 47–82.
WHO (2015). Global status report on road safety 2015. English. OCLC: 985423312.
115
Data Source Taxonomy for Supply
Network Structure Visibility
Johannes Zrenner1 , Ahmad Pajam Hassan2 , Boris Otto1 , Jorge Marx Gómez2
1 – Technical University of Dortmund
2 – Carl von Ossietzky University of Oldenburg
The supply network structure of manufacturers is complex and non-transparent.

In order to achieve a higher visibility and consequently increase the performance,
the existing lack of data has to be closed. This paper answers the questions, how
to identify, describe and compare suitable data sources for an end-to-end visibility.
Following the design science research process, two artifacts are developed based
on conceptual-to-empirical approaches. The initial conceptualizations result
from literature reviews. The conceptual representation of supply network struc-
ture data sources clarifies the relevant data entities and attributes. It supports
the identification process of relevant data sources. The data source taxonomy
(i.e. classification scheme) describes data sources using fourteen dimensions
and up to four potential characteristics. It assists a standardized description.
Both artifacts are demonstrated in case studies with German automotive Original
Equipment Manufacturers. The findings add to the knowledge base of supply
network visibility with a focus on the network structure. A large part of the existing
literature about supply chain visibility is too vague on the data perspective. There-
fore, this paper closes an important gap regarding the supply chain digitalization
by introducing two applicable results, which enable a new course of action for
practitioners and researchers.
Keywords: Data Source Taxonomy; Supply Chain Visibility; Supply Network

Structure; Design Science Research
117
Data Source Taxonomy for Supply Network Structure Visibility
1 Introduction
In the age of globalization, cost pressure and the customizing requirement, com-
panies are constrained to be focused on their main capabilities. Therefore, make
or buy decisions are followed by outsourcing business processes, which makes
supply networks more complex and dynamic (Tang and Tomlin, 2008). For han-
dling the dynamic of supply networks and reduce the uncertainty, it is necessary
to make supply network relations visible (Christopher and Lee, 2004). Thus, visibil-
ity becomes a key topic in Supply Chain Management research (Yu, Yan and Edwin
Cheng, 2001). With more visibility in supply chains, the performance improves
(Pidun and Felden, 2012) and decision making processes get supported (Kulp,
Lee and Ofek, 2004).
From the perspective of a manufacturer, visibility of all supply network members
is only given for the first tier supplier (Basole and Bellamy, 2014), because of
direct business relationships. Data about them is internally stored in databases.
After the first tier supplier, the upstream supply chain gets increasingly unknown
(Christopher and Lee, 2004). Missing interconnections between data silos among
business areas, supply chain members and other external data sources are the
reason for the lack of end-to-end visibility. This issue can be solved by linking
disparate data sources. The various available data sources are highly heteroge-
neous, own different characteristics and contents (Rozados and Tjahjono, 2014).
For that reasons, it is difficult to identify and select the suitable data sources
for linking processes. Furthermore, a general understanding of the data sources
is missing, which is needed to negotiate about data source sharing with supply
chain members. For example, potential data sources are owned by a first tier
supplier, logistics contractors or data providers.
The goal of this paper is to enhance the knowledge about data sources for supply
network structures and thus support linking processes for practitioners and re-
searchers. According to that, the first research question (RQ) addresses a general
data model, which helps to identify data sources with a suitable content. The
data model clarifies the entities and attributes, a data source about the supply
network structure has. The first research question is:
RQ1: How does a conceptual representation for supply network structure data
sources look like?
Following the identification of available data sources, a comparison has to be
conducted. The goal is to find out, which are the most suitable data sources
118
2 Related Work
from an effort-benefit point of view, for bridging data lacks within the supply
network. For that reason, a taxonomy (i.e. classification scheme) is needed, which
standardizes the description of the data sources and makes them comparable.
Accordingly, the second research question is:
RQ2: How does a taxonomy of data sources for supply network structures look
like?
This investigation contributes to the research of information systems. It follows
the design science research (DSR) process from Peffer et al. (2007). It is an iterative
research and design process. The process combines scientific and practitioner’s
knowledge for designing the artifact.
The remainder of the paper is structured as follows. Section 2 lays out the re-
lated work of concepts for supply chains and data source taxonomies. Section 3
presents a conceptual representation for supply network structure data sources.
In section 4 the data source taxonomy is established. Section 5 describes the
application of the conceptual representation and the taxonomy. Section 6 demon-
strates the conceptual representation and the taxonomy through two case studies
with German Original Equipment Manufacturers (OEMs). Section 7 summarizes
the contributions to research and practice as well as showing fields for future
research.
2 Related Work
2.1 Concepts for Supply Chains
Supply Chain Management is focused on managing the whole supply chain from
the raw material producer to the end customer (Harland, 1996). The intentions
of supply chain integration are improvements on performance and operational
figures (Ramdas and Spekman, 2000; Frohlich and Westbrook, 2001; Rosenzweig,
Roth and Dean, 2003; Cagliano, Caniato and Spina, 2004).
This investigation addresses the supply network from n-tier supplier to its Original
Equipment Manufacturer in a different way from most other researches. They
are focused on processing data from the interaction between supply network
members (Levary, 2000; Zhao, Xie and Zhang, 2002) and data sharing (Bowersox,
Stank and Closs, 2000). The object of this investigation is to analyze different data
119
sources about business relations in a manner, that it can be used for increasing
the visibility of supply network structures. In this context, Mukaddes et al. (2010)
propose a conceptual information model for Supply Chain Management and
integration. The model deals with the integration of supply chain members and
focuses on the flow of information between the chain members. Grubic and
Fan (2010) present a study of state-of-the-art research in supply chain ontology.
They identify outstanding research gaps and six supply chain ontology models
are identified from a systematic review of literature. In the context of decision
support in global supply chain, Wang, Wong and Fan (2013) build an ontology
for steel manufactures, which represent eleven main classes and the associated
attributes.
Despite existing contributions, there remains a need to specify a dedicated ontol-
ogy as a model for conceptualizing supply network structure data sources.
2.2 Data Source Taxonomies
In the generation of big data, there are a lot of data sources relevant for Supply
Chain Management. Leveling, Edelbrock and Otto(2014) propose an overarching
model of data categorization. It classifies data in nucleus data, community data
and open big data. For this, the following characteristics are considered: fuzziness,
volume and change frequency.
A review from Rozados and Tjahjono (2014) identifies 52 data sources and clas-
sifies them to a taxonomy. The main characteristics for this classification are
volume, velocity and variety. The characteristic variety is divided into structured
data, semi-structured data and unstructured data. For example, core transac-
tional data like transportation costs, origin and destination are structured data
with small volume and velocity. Semi or unstructured data like weather data or
machine-generated data have a high volume and velocity.
Otto, Abraham and Schlosser (2014) propose a morphology, which makes the
relevant characteristics of the data resource in networked industries transparent.
The morphology is based on four case studies and has eleven dimensions with
two to six characteristics. Even though this morphology is pretty detailed, it is
specific to the data resource. Data resources entail all databases in a company.
Given this scarce knowledge base about data source taxonomies, there is a clear
need to investigate this topic further.
120
3 Conceptual Representation of Supply Network Structure Data Sources
3 Conceptual Representation of Supply Network

Structure Data Sources
Since there is no common standard for the vision of an integrated supply network
through data flow integration (Loh, Koh and Simpson, 2006; Patnayakuni, Rai and
Seth, 2006), it is necessary to find new approaches to make the supply network
structure visible. For this purpose, the first step is to define the relevant aspects
of supply network structure data sources. Therefore, a conceptual representation
based on domain specific literature is developed. The authors identify relevant
literature, extract fundamental supply network data features and create an entity
relationship model following the framework of Chen (1976). By space limitations,
only the most relevant components of the model are explained in detail.
“The membership in a network and its organization vary for a given product and
over time” (Choi and Hong, 2002), the attribute “valid time” is considering this
fact and makes relations time sensitive. It is assigned to the relation “deliver”
and “demand”. In respect to the data and product flow, it is also necessary to
differentiate between several kinds of functional areas within an organization
and between supply network member organizations (Basole and Bellamy, 2014).
Following that, the entity company owns an attribute representing the type of
location. Examples for locations types are supplier, manufacturer, warehouse,
distributor and retailer. Moreover, to the entity transferred product an attribute is
assigned, which classifies its type, for example raw material, parts and compo-
nents. Companies within a supply chain are using applications which are based
on different ontologies. That leads to inconsistent terms and semantics which
have a negative influence on the interoperability for supply chain integration (Ye,
et al., 2008). According to the eCl@ss standard (eCl@ss, 2017), the entity product
could own different designations, for example identifier, preferred name or short
name. Those designations could even differ between languages. Considering
the variety in semantics, both entities “product” and “company” own an “alias”
feature. This feature makes it possible to express different semantics, which are
used by their data source.
Figure 1 shows the designed representation. It provides the relevant aspects of
supply network structure data sources in a formal way. Furthermore, it is used in
the taxonomy application (section 5) to identify potential data sources.
121
Figure 1: Conceptual representation of supply network structure data sources
4 Data Source Taxonomy
4.1 Background
A taxonomy is a classification scheme of objects from a specific research area. It

supports scientists and practitioners to understand and analyze complex domains
(Nickerson, Varshney and Muntermann, 2013). A morphological box structures the
identified dimensions and characteristics of the investigated objects. Therefore,
the morphological box is a suitable presentation of a taxonomy.
For the taxonomy development in information systems, Nickerson et al. (2013)
propose a method with two different approaches. The empirical-to-conceptual
approach starts with identifying a subset of objects and determining the char-
acteristics. After that, the characteristics are grouped into dimensions and the
taxonomy is created. The first step of the conceptual-to-empirical approach is
the conceptualization of the characteristics and dimensions. Step two is about
the examination of objects with these characteristics and dimensions. The third
step is to create the taxonomy.
122
4.2 Taxonomy Development Process
The authors follow the method of Nickerson et al. (2013) and choose the concep-
tual-to-empirical approach. For researchers with significant knowledge about
the domain, it is the recommended approach (Nickerson, Varshney and Munter-
mann, 2013). The initial conceptualization of the dimensions and characteristics
is based on the researcher notions about data sources as well as a review of the
literature. The authors scour scientific databases (IEEE Xplore, Web of Science,
ACM DL, Emerald Insights, ScienceDirect) for the following search terms: “data
taxonomy”, “big data in Supply Chain Management” and “Supply Chain Visibility”.
Two demonstration iterations in a business environment (section 6) are used to
examine data sources with the identified dimensions and characteristics.
The results of the described process lead to the taxonomy of data sources for
supply network structures (figure 2).
123
Figure 2: Taxonomy of Data Sources
124
4.3 Taxonomy Details
Each of the fourteen dimensions has two to four potential characteristics, which
are not exclusive but listed in the same row. That means, one dimension can have
one or more applying characteristics.
Data Source Availability: If a data source is internal or external, is the main differ-
entiation between data sources. From an enterprise perspective, an internal data
source is available in the company’s IT infrastructure for example, a database
of the ERP system. External data sources are not available in the company’s IT
infrastructure and can be divided into closed and open data. The difference is
the charged or free access to the data source. When the data has to be either
purchased or licensed, the data source is called closed (Leveling, Edelbrock and
Otto, 2014). Open Data is freely available on the public internet and can be used
without technical, financial or legal barriers (Murray-Rust, et al., 2010).
Data Source Interface: For end-to-end visibility of supply network structures,

data from different sources has to be accessible from a single point of access.
Therefore, an interconnection of data sources is required. Internally available
data sources can be distributed across business areas (Rozados and Tjahjono,
2014) and thus an internal interconnection is sufficient. Between the OEM and the
supplier normally exists a traditional Electronic Data Interchange (EDI) connection
for transactional data (Zilbert, 2000), which can be the connection for those kinds
of external sources. Web Services are a cross-platform way to connect external
data sources (Li, Sun and Tian, 2015) from a supplier, data provider, or Open Data.
A Data dump is a snapshot of data and is often used when IT infrastructure is
limited for a single or prototypical data transfer, like for example an Excel file sent
by email.
Data Source Pricing Model: For each of the four used characteristics, one example
of a potential supply network structure data source is named. The pricing model
from Achilles (2017) is based on the number of used product codes. Therefore,
Achilles is an example for a volume-driven pricing model. Panjivia (2017) has a
time-driven model with a monthly and annually option. IHS (2014) offers reports
for a unique payment. When linking data sources within a division of a company,
normally no pricing model is used.
Data Aggregation: Data aggregation describes the aggregation level of the data
source. An instantiation of an attribute of a data object is called data item. A data
record is the instantiation of a data object. Data records constitute database tables.
125
The highest level of aggregation is data resource. The data resource includes all
databases of a company (Otto, Abraham and Schlosser, 2014).
Data Occurrence/Updates: Data occurrence is distinguished in batches and streams.
Examples for streams are the social media streams or the shop-floor data streams
(Otto, Abraham and Schlosser, 2014). Batches are sets of data records and no
continuous data flow like streams. Therefore, data updates are usually for batches.
There can be an event or time-driven trigger for updates.
Data Ownership: If data is a private good, it is owned by a legal entity like a
company. Data as a club good is owned by a community and public data goods
like addresses are available for the public (Otto, Abraham and Schlosser, 2014).
Data Structure: Data without an identifiable structure or a data model is called
unstructured data. For example, photos or narrative text are unstructured data
(Kubler, et al., 2015). Semi-structured data like XML or machine-generated log
files have an irregular and flexible structure, but cannot be processed in relational
databases efficiently. However, ERP transaction data or origination and destina-
tion are structured data and can be straightforwardly transferred to a relational
database (Rozados and Tjahjono, 2014).
Data Format: Open data formats, such as HTML, XML, JSON, RDF or CSV can be
used by anyone, unlike proprietary formats. Binary data from machine to machine
communication is one example for proprietary formats (Kubler, et al., 2015). PDF
or DOC are other common proprietary data formats.
Data Standardization: Data can be standardized on the syntactic level, on the
semantic level and on value level. Standardization on a value level standardizes
the possible values of a data item (Otto, Abraham and Schlosser, 2014). Semantic
heterogeneity through homonyms or synonyms leads to different meanings or
interpretation problems. A homonym denotes different terms and a synonym
denotes the same term as another notation (Bergamaschi, et al., 2011). The
demonstration of the taxonomy (section 6) leads to additional characteristics.
If the data source is not standardized at any level, the characteristics “No” is
necessary. Moreover, the demonstration results lead to a differentiation between
“Intra Data Standardization” and “Inter Data Standardization”. Intra refers to
the standardization inside of a data source. On which level the standardization
of the data fits from one data source to other ones, is covered by “Inter Data
Standardization”.
Data Quality and Data Value: Data quality is essential for data analytics because
the accuracy of the analytics methods depends on the data which they are based
126
5 Application
on. That means the results can only be as good as the input data. The intrinsic
data quality dimensions are (Hazen, et al., 2014):
— Accuracy: Does the data equivalent to their corresponding real values?
— Timeliness: Currency (length of time since the record’s last update) and
Volatility (frequency of updates)
— Consistency: Matches the data regarding format and structure?
— Completeness: Is the data complete or is there data missing?
Extended literature considers the value of information and neglects the value of
data. Data is used to generate information and therefore scholars do not distin-
guish between information resources and data resources. Data quality affects
data value and the study of Ahituv (1989) identifies influencing factors for the
value of information (Otto, 2015). The influencing factors for the value of informa-
tion are classified into four categories (timeliness, contents, format, cost) (Ahituv,
1989). Except “Currency”, “Completeness” and “Accuracy”, all named data quality
and data value attributes are already covered in the taxonomy. The currency of
data can be distinguished in forecast, up-to-date and outdated. Completeness
and accuracy are very situational attributes. Therefore, a distinction between
low, medium and high is appropriate. For assessing the completeness, the con-
ceptual representation of supply network structure data sources (figure 1) can be
helpful.
Data Sharing: The shareability of data can be distinguished in proprietary, free
and open. Free and open data are allowed to being shared. The difference is that
the source of free data always has to be disclosed. However, it is not allowed to
share proprietary data (Otto, Abraham and Schlosser, 2014).
5 Application
Prerequisite for the application of the taxonomy is the identification of data

sources, which should be described or compared for linking processes. The con-
ceptual representation (figure 1) is used for identifying potential data sources, by
demonstrating the relevant data entities and attributes.
Through applying the taxonomy on the identified data source, the taxonomy is
instantiated. Therefore, each dimension of the taxonomy is analyzed and the
127
corresponding characteristics of the data source are identified. As a result of the

application process, the identified characteristics are highlighted in the taxonomy
or written in a new table like in section 6. An important factor for an effective
application process are the involved roles. Very suitable roles from a knowledge
and capability perspective are data stewards, data owners, data architects and
data scientists (Otto, Abraham and Schlosser, 2014).
6 Demonstration
According to Hevner, March and Park (2004), case studies are a suitable demon-
stration method for artifacts. In order to demonstrate the applicability of the
developed conceptual representation of supply network structure data sources
and the taxonomy, the authors make case studies with two German OEMs. The
data source availability is the superior and most selective data source dimension
in the taxonomy. Therefore, each case study covers one data source scenario for
each characteristic of that dimension (internal, external-closed, external-open).
The authors perform one workshop with each OEM. The workshops are divided
into several steps. The first step is an introduction to the paper‘s topic and its goals.
After that, the workshop participants discuss the dimensions and characteristics
of the taxonomy. The results of that discussion are already considered in the
taxonomy (figure 2). Step three is the collective decision of the data sources
for the three data source scenarios. The selected data sources should have the
potential for reconstructing the supply network structure of the OEM, which is
determined by using the conceptual representation of supply network structure
data sources (figure 1). Step four is the application of the taxonomy for each
scenario. The last step is a brainstorming session in order to obtain feedback to
the data model and the taxonomy. The participants and results of both workshops
are described below.
6.1 Case 1 - OEM 1
The participants of the workshop are the manager of the department “Informa-
tion Processes Logistics”, one data architect and one data scientist. The results of
the discussion about the dimensions and characteristics of the taxonomy (step
128
6 Demonstration
Table 1: Scenario I - SAP MM from the OEM
Dimension Characteristics
Data Source Availability Internal
Data Source Interface Internal Interconnection
Data Source Pricing Model No
Data Aggregation Database, Records, Items
Data Occurence/Update Event-Driven Batch, Time-Driven Batch
Data Ownership One Legal Entity
Data Structure Structured, Semi-Structured
Data Format Proprietary
Intra Data Standardization Semantics, Syntax, Values
Inter Data Standardization Semantics
Data Currency Forecast, Up-To-Date, Outdated
Data Completeness High
Data Accuracy High
Data Sharing Proprietary
two) are two changes in the taxonomy. The name of the dimension “Data Organi-
zation” changes to “Data Aggregation”. Data Organization does not describe the
corresponding characteristics well. Organization is a very general term and leads
to a lack of clarity. The OEM has future supplier relations in his ERP system. Thus,
the characteristic “Forecast” becomes part of the dimension “Data Currency”.
In step three, the participants decide about the IT systems or rather data sources
for the three scenarios, by using the conceptual representation. The first scenario
is about the OEM’s internal SAP MM (Materials Management) system. The Mate-
rial Management module is part of the SAP ERP (Enterprise Resource Planning).
It contains material, supplier and other master data. Furthermore, it contains
transaction data about material flows from the supplier to the OEM. The scenario
about the internal SAP MM data source is summarized in table 1.
Scenario II is about the SAP MM system of a 1-tier supplier, which is an external

and closed data source. In contrast to the OEM’s SAP MM system, there is master
data and transactional data about 2-tier supplier, which are part of the supply
network from the OEM. Table 2 includes the characteristics of the data source
“SAP MM from a 1-tier supplier”.
129
Table 2: Scenario II - SAP MM from a 1-tier supplier

Data Source Availability External-Closed
Data Source Interface Traditional EDI, Web Services, Offline Data Dump
Data Source Pricing No
Model
Data Aggregation Records, Items
Data Ownership One Lega Entity
Data Structure Structured, Semi-Structured
Intra Data Standardiza- Semantics, Syntax, Values
tion
Inter Data Standardiza- Semantics
tion
Data Currency Forecast, Up-To-Date, Outdated
Data Accuracy High
For scenario III the open data source “Eurostat” (2017) is used, which provides
databases like statistics about international sourcing activities, subsidiaries and
air freight transports (table 3).
In the feedback session, the participants report a good level of satisfaction with
the workshop results. In their opinion, the results provide clear insights into
the different characteristics of the available data sources. The conceptual rep-
resentation and the taxonomy are a useful support to identify the data sources
with a high potential for increasing the supply network structure visibility. The
determination of the data source characteristics presupposes knowledge and
experience in the field of data science. That is the critique of the participants
because data scientists are rare in this company. In conclusion, the department
“Information Processes Logistics” will use the taxonomy to compare data sources
for future linking projects. They have a number of project requests from opera-
tional logistics departments, which need more data about the supply network for
a proactive risk management.
130
6 Demonstration
Table 3: Scenario III - Eurostat

Data Source Availability External-Open
Data Source Interface Web Services, Offline Data Dump
Data Aggregation Records
Data Ownership Public
Data Structure Structured
Data Format Open
Intra Data Standardization Semantics, Values
Inter Data Standardization No
Data Currency Outdated
Data Completeness Low
Data Accuracy Medium
Data Sharing Open
6.2 Case 2 - OEM 2
The participants of the workshop are the head of the department “Information
management of Procurement” and two data scientists.
In step two the participants discuss and review the taxonomy. The outcomes are
two changes. The first one is adding the characteristic “No” to the dimension
“Data Standardization”. This value is rarely needed in a business application
focus, but it becomes important when external data sources are investigated. The
second change is the partition of the dimension “Data Standardization” into “Intra
Data Standardization” and “Inter Data Standardization”. The intra standardization
refers to standardization inside a data source. The inter standardization describes
the degree of standardization over various data sources.
In step three, the participants choose three scenarios to demonstrate the tax-
onomy. Scenario IV is about the OEM’s internal and customized visual basic
application, which is based on a spreadsheet for the purpose of capacity man-
agement. It contains material, supplier and other master data. Furthermore, it
contains transaction data about material flows between sub-supplier, supplier
131
Table 4: Scenario IV - Spreadsheet for Capacity Management
Data Source Availability Internal
Data Source Interface Offline Data Dump
Data Aggregation Records
Data Occurence/Update Event-Driven Batch
Data Currency Up-To-Date
Data Accuracy High
and the OEM. Table 4 summarizes the characteristic of the Spreadsheet for Capac-
ity Management.
The fifth scenario (table 5) is about an external data source. The market research
firm “IHS” provides a collection of automotive supplier relationships called “Who
Supplies Whom”. It contains data about 700 supplier and their automotive com-
ponents, modules and systems. The online accessible demo data source does not
provide a full access (IHS, 2017). The whole data set is just available for registered
customers. This is the reason why the data source is characterized as external
and closed.
The sixth demonstration scenario is about an external and open data source.
Online news and press releases are available on the internet and provide data
about relations between companies. An example for supply network structure
data in online news is the product press release on a supplier website (AGC, 2017).
Table 6 summarizes the scenario about online news and press releases.
The brainstorming session in the last step of the workshop brings up the follow-
ing feedback from the participants. The taxonomy clarifies the differences and
similarities of public data sources and data sources, which are liable to costs.
132
6 Demonstration
Table 5: Scenario V - IHS Who Supplies Whom
Data Source Availability External-Closed (only demo is open)
Data Source Interface Web Services, Offline Data Dump
Data Source Pricing Model Volume-Driven, Time-Driven
Data Aggregation Database
Data Occurence/Update Time-Driven Batch
Inter Data Standardization Semantics
Data Currency Up-To-Date, Forecast
Data Accuracy High
Table 6: Scenario VI - Online news and press releases
Data Source Availability External-Open
Data Source Interface Web Services
Data Aggregation Item
Data Occurence/Update Event-Driven Batch
Data Ownership One Legal Entity, Community, Public
Data Structure Unstructured
Data Format Open
Intra Data Standardization No
Data Currency Up-To-Date
Data Completeness Medium
Data Accuracy Medium
Data Sharing Proprietary, Free, Open
133
Therefore, it is simpler to identify weak spots of a data source and make a decision
based on objective criteria. The conceptual representation helps to focus on
the essential data source content and supports the selection processes of the
large number of available data sources. A way to quantify the suitability of data
sources for increasing the supply network structure visibility, would be a field for
future research. The quantification helps to identify the most suitable data source
for a linking project with current data sources. The department “Information
management of Procurement” expects a strategic advantage for sourcing and
supplying processes by increasing the supply network structure visibility.
7 Conclusion
Following the design science research process, this research addresses two ques-
tions to implement a higher supply network structure visibility through the linking
of different data sources. The first question, how a conceptual representation
for automotive supply network structure data sources looks like, is answered
with an entity relationship model. It defines the essential content, a data source
about automotive supply network structure has to have. More particularly, the
model clarifies the necessary entities and attributes of a data source for improv-
ing the supply network structure visibility. The second question deals with a
taxonomy of data sources for supply network structures. The taxonomy describes
data sources using fourteen dimensions and up to four potential characteristics.
The demonstration of both research results uses a case study with two German
OEMs. The case study proves the utility of the conceptual representation and of
the taxonomy, as a support to identify and describe data sources about supply
network structures in a systematic way.
The paper contributes to practice and research. Practitioners in charge of the

data management can use the conceptual representation in a first step to identify
data sources for an initial selection. In the second step, the taxonomy enables
them to describe and compare the data sources, with the goal to identify the most
suitable data sources for the linking process. Furthermore, they get a general
understanding of the data sources, which is needed to negotiate about data
source sharing and evaluate the data source portfolio. For example, potential
data sources are owned by data providers or other supply chain members like
supplier and logistics contractors. For researchers, the conceptual representation
134
References
and the taxonomy add to the growing knowledge base of supply network visibility
and are a basis for future research in the field of linking data sources.
Even with the promising results of the demonstration in a business environment,
the authors cannot ensure that the taxonomy is complete. Both companies from
the case studies are German automotive Original Equipment Manufacturers. Thus,
more review and evaluation iterations in other industries are a valid field for future
research.
References
Achilles (2017). Helping automotive companies manage supply chain compliance and meet regula-
tory and legislative demands.
AGC (2017). Press release. http://www.agc-glass.eu/English/Homepage/News/Press-room/Press-
Detail-Page/page.aspx/979?pressitemid=3612.
Ahituv, N. (1989). “Assessing the value of information”. In: Proceedings of the tenth international
conference on Information Systems - ICIS ’89. Ed. by J. I. DeGross, J. C. Henderson, and B. R.
Konsynski. New York, New York, USA: ACM Press, pp. 315–325.
Awan, I., ed. (2015). 3rd International Conference on Future Internet of Things and Cloud: 24 - 26
Aug. 2015, Rome, Italy. Piscataway, NJ: IEEE.
Basole, R. C. and M. A. Bellamy (2014). “Visual analysis of supply network risks: Insights from the
electronics industry”. In: Decision Support Systems 67, pp. 109–120.
Bergamaschi, S., D. Beneventano, F. Guerr, and M. Orsini (2011). “Data Integration”. In: Handbook
of conceptual modeling. Ed. by B. Thalheim and D. W. Embley. Berlin and New York: Springer,
pp. 441–476.
Bowersox, D. J., T. Stank, and D. Closs (2000). “Ten Mega-Trends That Will Revolutionize Supply
Chain Logistics”. In: Journal of Business Logistics 21, pp. 1–16.
Bundesvereinigung Logistik e.V., ed. (2014). 7th International Scientific Symposium on Logistics:
Germany, 4 - 5 Jun. 2014.
Cagliano, R., F. Caniato, and G. Spina (2004). “Lean, Agile and traditional supply: How do they
impact manufacturing performance?” In: Journal of Purchasing and Supply Management 10.4-5,
pp. 151–164.
Chen, P. P.-S. (1976). “The entity-relationship model—toward a unified view of data”. In: ACM
Transactions on Database Systems 1.1, pp. 9–36.
Choi, T. Y. and Y. Hong (2002). “Unveiling the structure of supply networks: Case studies in Honda,
Acura, and DaimlerChrysler”. In: Journal of Operations Management 20.5, pp. 469–493.
Christopher, M. and H. Lee (2004). “Mitigating supply chain risk through improved confidence”. In:
International Journal of Physical Distribution & Logistics Management 34.5.
DeGross, J. I., J. C. Henderson, and B. R. Konsynski, eds. (1989). Proceedings of the tenth interna-
tional conference on Information Systems - ICIS ’89. New York, New York, USA: ACM Press.
eCl@ss (2017). Classification and Product Description. http://wiki.eclass.eu/w/index.php?title=File:
ReleaseNotesBasicClassificationClass.JPG&filetimestamp=20120227160627.
Education Special Interest Group of AITP, ed. (2000). Proceedings ISECON 2000.
135
Eurostat (2017). Your key to European statistics. http://ec.europa.eu/eurostat/data/database.

Frohlich, M. T. and R. Westbrook (2001). “Arcs of integration: an international study of supply chain
strategies”. In: Journal of Operations Management 19, pp. 185–200.
Grubic, T. and I.-S. Fan (2010). “Supply chain ontology: Review, analysis and synthesis”. In: Com-
puters in Industry 61.8, pp. 776–786.
Harland, C. M. (1996). “Supply Chain Management, Purchasing and Supply Management, Logis-
tics, Vertical Integration, Materials Management and Supply Chain Dynamics”. In: Blackwell
Encyclopedic dictionary of operations management 15.
Hazen, B. T., C. A. Boone, J. D. Ezell, and L. A. Jones-Farmer (2014). “Data quality for data science,
predictive analytics, and big data in supply chain management: An introduction to the prob-
lem and suggestions for research and applications”. In: International Journal of Production
Economics 154, pp. 72–80.
Hevner, A. R., S. T. March, and J. Park (2004). “Design Science in Information Systems Research”.
In: MIS Quarterly 28.75-105.
ICISCE, ed. (2015). 2nd International Conference on Information Science and Control Engineering
(ICISCE), 2015: 24 - 26 April 2015, Shanghai, China. Piscataway, NJ: IEEE.
IEEM, ed. (2014). IEEE International Conference on Industrial Engineering and Engineering Manage-
ment (IEEM): 9 - 12 Dec. 2014, Malaysia. Piscataway, NJ: IEEE.
IHS (2014). Auto Tech Report: Infotainment Market Overview. http://shop.ihs.com/buy/en/ihs/auto-
tech-report-infotainment-market-overview-3000006155.
IHS (2017). Who Supplies Whom. http://www.ihssupplierinsight.com/shop/product/261/who-
supplies-whom-complete-database.
Kubler, S., M.-J. Yoo, C. Cassagnes, K. Framling, D. Kiritsis, and M. Skilton (2015). “Opportunity to
Leverage Information-as-an-Asset in the IoT”. In: 3rd International Conference on Future Internet
of Things and Cloud. Ed. by I. Awan. Piscataway, NJ: IEEE, pp. 64–71.
Kulp, S. C., H. L. Lee, and E. Ofek (2004). “Manufacturer Benefits from Information Integration with
Retail Customers”. In: Management Science 50.4, pp. 431–444.
Levary, R. R. (2000). “Better supply chains through information technology”. In: Industrial Manage-
ment 42.3.
Leveling, J., M. Edelbrock, and B. Otto (2014). “Big data analytics for supply chain management”.
In: IEEE International Conference on Industrial Engineering and Engineering Management (IEEM).
Ed. by IEEM. Piscataway, NJ: IEEE, pp. 918–922.
Li, B., L. Sun, and R. Tian (2015). “Multi-source Heterogeneous Data Exchange System for Cloud
Services Platform Based on SaaS”. In: 2nd International Conference on Information Science and
Control Engineering (ICISCE), 2015. Ed. by ICISCE. Piscataway, NJ: IEEE, pp. 327–331.
Loh, T. C., S. C. L. Koh, and M. Simpson (2006). “An investigation of the value of becoming an
extended enterprise”. In: International Journal of Computer Integrated Manufacturing 19.1,
pp. 49–58.
Mukaddes, A. M. M., Choudhury, Abdul Anam Rashed, Abdul Malek, A. B. M., and J. Kaiser (2010).
“Mukaddes 2010 - Developing and Information Model for Supply Chain Information Flow and
its Management”. In: International Journal of Innovation, Management and Technology 1.2,
pp. 226–231.
Murray-Rust, P., Neylon C., Pollock R., and J. Wilbanks (2010). Panton Principles, Principles for open
data in science.
136
References
Nickerson, R. C., U. Varshney, and J. Muntermann (2013). “A method for taxonomy development
and its application in information systems”. In: European Journal of Information Systems 22.3,
pp. 336–359.
OSCM, ed. (2014). 6th International Conference on Operations and Supply Chain Management: 10-12
Dez. 2014, Bali.
Otto, B. (2015). “Quality and Value of the Data Resource in Large Enterprises”. In: Information
Systems Management 32.3, pp. 234–251.
Otto, B., R. Abraham, and S. Schlosser (2014). “Toward a Taxonomy of the Data Resource in the
Networked Industry”. In: 7th International Scientific Symposium on Logistics. Ed. by Bundesvere-
inigung Logistik e.V., pp. 382–421.
Panjivia (2017). Powerful Insights with U.S. Customs Data.
Patnayakuni, R., A. Rai, and N. Seth (2006). “Relational Antecedents of Information Flow Integration
for Supply Chain Coordination”. In: Journal of Management Information Systems 23.1, pp. 13–49.
Peffers, K., T. Tuunanen, M. A. Rothenberger, and S. Chatterjee (2007). “A Design Science Research
Methodology for Information Systems Research”. In: Journal of Management Information Sys-
tems 24.3, pp. 45–77.
Pidun, T. and C. Felden (2012). “Two Cases on How to Improve the Visibility of Business Process
Performance”. In: 45th Hawaii International Conference on System Science. Ed. by R. H. Sprague.
Piscataway, NJ: IEEE, pp. 4396–4405.
Ramdas, K. and R. E. Spekman (2000). “Chain or Shackles: Understanding What Drives Supply-
Chain Performance”. In: Interfaces 30.4, pp. 3–21.
Rosenzweig, E. D., A. V. Roth, and J. W. Dean (2003). “The influence of an integration strategy on
competitive capabilities and business performance: An exploratory study of consumer products
manufacturers”. In: Journal of Operations Management 21.4, pp. 437–456.
Rozados, I. V. and B. Tjahjono (2014). “Big Data Analytics in Supply Chain Management: Trends and
Related Research”. In: 6th International Conference on Operations and Supply Chain Management.
Ed. by OSCM.
Sprague, R. H., ed. (2012). 45th Hawaii International Conference on System Science: (HICSS) ; USA, 4
- 7 Jan. 2012. Piscataway, NJ: IEEE.
Tang, C. and B. Tomlin (2008). “The power of flexibility for mitigating supply chain risks”. In:
International Journal of Production Economics 116.1, pp. 12–27.
Thalheim, B. and D. W. Embley, eds. (2011). Handbook of conceptual modeling: Theory, practice,
and research challenges. Berlin and New York: Springer.
Wang, X., T. N. Wong, and Z.-P. Fan (2013). “Ontology-based supply chain decision support for
steel manufacturers in China”. In: Expert Systems with Applications 40.18, pp. 7519–7533.
Ye, Y., D. Yang, Z. Jiang, and L. Tong (2008). “Ontology-based semantic models for supply chain
management”. In: The International Journal of Advanced Manufacturing Technology 37.11-12,
pp. 1250–1260.
Yu, Z., H. Yan, and T. C. Edwin Cheng (2001). “Benefits of information sharing with supply chain
partnerships”. In: Industrial Management & Data Systems 101.3, pp. 114–121.
Zhao, X., J. Xie, and W. J. Zhang (2002). “The impact of information sharing and ordering co–ordination
on supply chain performance”. In: Supply Chain Management: An International Journal 7.1,
pp. 24–40.
Zilbert, A. B. (2000). “A Comparative Study of Traditional Electronic Data Interchange versus
Internet Electronic Data Interchange”. In: Proceedings ISECON 2000. Ed. by Education Special
Interest Group of AITP.
137
Transforming Blood Supply Chain
Management with Internet of Things
Paradigm
Wijai Boonyanusith1 , Phongchai Jittamai1
1 – Suranaree University of Technology
Internet of Things (IoT) paradigm has transformed the traditional healthcare sys-
tem into the smart one. The IoT-based healthcare system is able to connect all
available resources to operate healthcare services via the internet network. It
would maximize the utilization of healthcare resources such as blood and blood
products, which are scarce resources. Blood supply chain is a challenging sys-
tem to manage, which researchers usually study for improved methodologies
to boost its operation efficiency. Thus, this paper applied a content analysis in
order to propose guidelines for enhancing and transforming the blood supply
chain management based on the theoretical considerations in the IoT paradigm.
The implementation strategies are provided to point out the issues that blood
utilization and blood operational decision makings can be improved. Finally, the
IoT-based framework of blood supply chain management is proposed to conceptu-
alize the interactions between things, humans, processes, and service operations
in the blood supply chain management.
Keywords: Blood Supply Chain; Internet of Things; Healthcare Services; Logis-

tics Management
139
Transforming Blood Supply Chain Management with Internet of Things Paradigm
1 Introduction
Internet of Things (IoT) has been highlighted as a new intelligent tool for the
ICT industry. IoT incorporates advance technologies that facilitate applications,
devices, and objects to communicate among themselves through the connected
networks (Tarouco et al., 2012). The IoT technologies mainly consist of infor-
mation generation, collection and sharing as well as big data management and
service application. Initially, the IoT has been used in the transportation track-
ing, information pricing of products, and map navigation. However, the advance
technologies such as wireless, cloud based network, smart phones, wearable
devices, and high speed connection have incorporated these IoT services into
daily life activities. The IoT paradigm has applied multidiscipline knowledge and
transformed the service platform to be more intelligent. It has become a useful
tool in the market and its popularity has been on the rise after the introduction of
many new concepts such as Smart City, Smart Home, and Smart Health (Kang et
al., 2015). According to the health context, the IoT has been widely used in various
healthcare services (Fernandez and Pallis, 2014). In the IoT-based healthcare,
there are interactions between things (applications) and humans (practitioners
and patients) in order to obtain real-time data and support decision-making ac-
tivities for care deliveries and other associated services. Several works have been
conducted with the use of IoT technologies to the healthcare domain including
pervasive healthcare (Schreier, 2010; Doukas and Maglogiannis, 2010), and drug
interaction checking (Jara et al., 20106). Many researchers have raised the issue
of the IoT capabilities in healthcare systems in order to develop medical appli-
cations based on the IoT technologies (Bui and Zorzi, 2011). Moreover, the IoT
would increase the utilization of resources in the healthcare environment, for
example, managing limited resources for elderly care in the community more
effectively (Feki et al., 2013; Li et al., 2011). The IoT-based healthcare system
is able to connect to available resources in order to operate healthcare actions
via the internet network. According to the benefits of the resource mobility, this
IoT concept could also be applied to manage other resources in the healthcare
domain such as specimen, vaccines, and blood products.
Blood is one of the most important resources in the healthcare system (Nagurney,
Masoumi and Yu, 2012). Demand for blood usage caused by the illness of human
is uncontrollable. Blood supply is also collected from human beings who have
willingness to donate. Uncertainties in blood demand and supply are inevitable.
Blood is perishable product, and its lifetime is within a short time period after
collecting. The balance of blood demand and supply in the system has direct
140
2 IoT Applications in Healthcare Domain
impact on the survival of the patient. This makes the task of managing the blood
supply chain to be a challenging one (Dobbin, Wilding and Cotton, 2009). Due
to an increasing blood demand, new methods for improving blood supply chain
management are necessary (Rytila and Spens, 2006). Consequently, the moti-
vation of this paper is to incorporate the IoT paradigm in managing the blood
supply chain in order to propose guidelines for elevating blood demand and
supply fulfillment services.
2 IoT Applications in Healthcare Domain
The IoT-based applications in the healthcare domain can be classified into four
main groups which are tracking, identification and authentication, automatic
data collection, and sensing (Vilamovska et al., 2009).
2.1 Tracking
Tracking is to identify the flow of people or objects while transferring. The func-
tion involves the real-time tracking such as patient-flow monitoring, workflows in
hospital departments, and tracking of motions through bottleneck points. Track-
ing is mostly applied to real-time inventory location tracing in order to maintain
the available resources and usage monitoring. Particularly, it is used for materials
tracking during operations, such as specimen and blood products.
2.2 Identification and Authentication
Mainly, this step assures accurate patient identification with matching medical
record, proper treatment and medication in order to prevent any damage to any
patient. Authentication is fundamental security protocol for medical practition-
ers to access patients’ information. Moreover, the function is used as security
protection to prevent the losses of medical resources.
141
2.3 Automatic Data Collection
Automatic data collection is functioned to collect data form various operational

processes in order to minimize processing time, enhance automated process, im-
prove care and procedure traceability, and heighten inventory management. This
function is integrated with the radio-frequency identification (RFID) technology
to collect comprehensive real-time health information of the patient with the
medical applications through the connected networks.
2.4 Sensing
Sensors can be applied both for in-patient and out-patient care operations. Its
functions have ability to drive the patient into the center of processes such as
diagnosing conditions, providing patient health information, monitoring patient
behaviors, and alerting patient drug prescriptions. Wireless access-based network
systems can be used to reach the patient everywhere, with the support of bio-
signal monitoring of wearable devices.
In the healthcare domain, the IoT-based applications consist of these four afore-
mentioned main functions. This paper has incorporated this paradigm to propose
the guidelines to improve performances in blood service operations.
3 Blood Supply Chain Management
Managing the blood supply chain is one of the challenging operations research
problems in the supply chain literature (Pierskalla, 2005). Practically, blood ser-
vice operations involve blood collection, processing, inventory management,
distribution, blood-banking management, and transfusion. Blood Center collects
whole blood from donors, processes it into blood products at a Regional Blood
Center and distributes them to hospitals in the network in order to transfuse to
the patient. The main objective is to maximize blood utilization by minimizing
both blood shortage and outdate rates in the system. The diagram of blood supply
chain processes is shown in Figure 1.
142
3 Blood Supply Chain Management
Figure 1: Blood supply chain diagram
143
3.1 Regional Blood Center
Regional Blood Center (RBC) has principal operations for blood collection, pro-
cessing and testing, production, inventory management, and distribution to
hospitals. Blood collection is to acquire the whole blood from donors, includ-
ing donor recruitment and promotion. Collected whole blood units are tested
in the laboratory for infectious agents and negative blood units. Then they will
be processed into blood products and kept in the RBC storage facility. RBC is
responsible for managing this inventory and allocating available blood units to
hospitals according to requisitions from their local blood banks.
3.2 Hospital Blood Bank
Each hospital blood bank (HBB) has to manage blood inventory and its operations
within the hospital. Doctors put requests for the certain amount and groups of
blood products for patients’ treatments. When available blood products are as-
signed for any patient, these units will be crossmatched to verify the compatibility
with each particular patient. Crossmatched units will be stored in the assigned
inventory and these units may not always be used because either doctors may
postpone operations or there are blood units left after patients’ treatments. These
un-transfused units will be returned to the unassigned inventory if they do not
expire. The period between blood crossmatching and releasing any unused units
is called “a period for reserving blood”. The longer this period is, the higher the
probability that blood will expire before its actual use. Finally, the crossmatched
units will be transfused to the patient for treatment.
4 Transforming Blood Supply Chain Management with

IoT Paradigm
Blood supply chain is a challenging system to manage, which researchers usually

study for improved methodologies to boost its operations. Most of the related
works were involved with the use of operations research to study the associated
problems in the context. There was also a literature review paper in the supply
chain management of blood products (Beliën and Forcé, 2012) which presented
144
4 Transforming Blood Supply Chain Management with IoT Paradigm
a structured review in different perspectives for managing blood service opera-

tions. However, this paper investigated a few number of papers related to the
technological issues and it did not provide trends in the technology aspect.
Moreover, there are some works related to the study of using RFID in the blood op-
erations for tracking (Dvais et al., 2009), inventory management (Xu, Lian and Yao,
2013), and quality and safety (Briggs et al., 2009; Hohberger et al., 2012). These
previous works pointed out the benefits of using this technology in managing
blood. According to the arising trends in technologies, it would be beneficial to
consider the possible advantages from using the IoT paradigm in the blood supply
chain management. Thus, this paper applied a content analysis method to review
the related works as well as investigating the blood service organizations in order
to identify the opportunities for enhancing the blood supply chain performance
based on the theoretical considerations in the IoT paradigm.
In practical, there is an implementation of RFID for blood management in each
blood bag labeled in order to prevent medical errors in blood transfusion. It also
provides blood tracking to the origin and to ensure the quality of blood products
during the storage and transportation. However, the capabilities of IoT-based
applications can be extended by the implementation of the RFID and the other
functions to enhance the blood service operations more efficiently. This paper
provides conceptual guidelines in transforming blood service operations manage-
ment to enhance blood utilization with the IoT-based applications paradigm.
4.1 Tracking
4.1.1 Inventory count and location tracking
There are two main blood inventory storages in the supply chain, RBC and HBBs.
RBC has to operate its inventory in order to fulfill demands from the HBBs in
its network. In practical, there is a collaboration system between the RBC and
the HBBs that helps managing blood supply throughout the network. However,
in some developing and underdeveloped countries, there are not any efficient
systems to track down the blood units from the donors to the patient. This issue
could lead to the shortage of blood units in some hospitals where their HBBs’
systems do not connect to the system of the RBC and the systems of other hos-
pitals in the network. Real-time inventory count would be applied to check the
collaborative inventory in order to automated count the overall blood stock levels
145
in the HBBs. Hence, the RBC can obtain these counted blood stock data for future
blood collection and replenishment planning. Moreover, the location tracking of
each blood unit would enhance blood transshipment among the HBBs in case
there are any blood emergency demand requests made from the HBBs and the
RBC cannot supply enough blood units to responsd to those requests.
4.1.2 Safety and traceability
The blood traceability system is an online platform which is connected into the
RFID-blood bag and barcode. It is able to integrate the sharing information and
coordination among blood service organizations in the supply chain in order to
mitigate blood transfusion risks to the patient, such as human error, incorrect
blood products, and errors in medication administration (Dzik, 2007).
4.2 Identification and Authentication
4.2.1 Auto ID/bar code enabled transfusion administration
The identity check between each patient and particular blood product is the
crucial task to avoid mistake in the transfusion. The RFID-blood bags and the
traceability system can ensure that the right blood type with the correct quantity is
always delivered to the right patient when there is a request (Koshio and Akiyama,
2009). The right information and the right process of transfusion are to be secured
for the safety of the patient by using auto ID at the transfusion. RFID is used to
guarantee that the right process is conducted according to the medical standard
of transfusion. In case there are any complications after blood transfusion, this
function is able to trace back information to the donor of that particular blood
unit/product. Hence, RFID and barcode have capabilities of improving patient
safety through managing transfusion activities and traceability of blood products
to authenticate the right process and information.
146
4.3 Automatic Data Collection
4.3.1 Blood inventory management
Blood inventory is a crucial task in the supply chain. The main objective is to
maximize blood utilization in such a way that blood shortage and outdated rates
are at minimum levels. Using RFID and barcode to connect each blood bag to the
blood inventory system can lead to the ability to manage blood information more
efficiently. Real-time data of blood stock levels throughout the network would
support blood collection planning for the RBC. The actual demand visibility can
enhance an effective blood component production planning at the RBC in order
to satisfy demands at each HBB. The blood operational costs are reduced because
the RBC does not have to supply very large amount of blood units. Consequently,
the blood outdated rate is minimized. Furthermore, using RFID and barcode to
connect among the HBBs in the network can enhance the blood transshipment
between the hospitals more rapidly when there are emergency demands in any
specific groups needed and the RBC cannot response to these events.
4.3.2 Blood tracking and tracing
According to the use of RFID and barcode in the blood data connection, their
functions completely support the blood tracking and tracing for expiration dates,
blood shortage, and blood restocking requests at the HBBs. This smart network is
capable of storing records of the dates and times of the blood expiration data so
the RBC has the ability to track down the amount of blood units that are soon to
expire. This information is beneficial for the RBC to set its blood rotation policy in
order to minimize the number of expired blood units. Moreover, real-time blood
tracking and tracing system will help the RBC to replenish its own inventory and
allocate blood products to the hospital blood banks according to the specific
demand from each hospital more efficiently.
4.3.3 Logistics management
Logistics activities in the blood supply chain management include blood col-
lection, blood processing, inventory management, distribution, blood-banking
management, and transfusion. The decision-making done in each process has di-
rect impact on the subsequent processes. RFID and barcode can transmit relevant
147
data within the blood operations management and they will provide real-time
information for supporting various types of operational decisions in the blood
service organizations. This function is also applicable to manage the stochastic
data of blood demand and supply in order to reduce the uncertainties in the blood
supply chain management.
4.4 Sensing
4.4.1 Transportation monitoring
Temperature control during transportation is vital for maintaining blood quality.

Using sensing technology and real-time temperature tracking in blood transporta-
tion can ensure the quality of blood products. The sensing devices implemented
for temperature tracking and measuring during the blood transportation could
help ensuring the blood quality control during transportation in real-time. More-
over, the integration of the RFID system and the location tracking technology
can provide information on the routes that tend to have oscillated temperature,
which may be harmful to blood quality.
4.4.2 Patient monitoring
It is necessary for the medical practitioners to be alerted if the patient suffers any
complication symptoms after receiving blood. A wearable bio-signal device that
binds the patient to the system can be used to monitor and transmit data automat-
ically back to the medical staff for any prompt treatments during postoperative
blood transfusions.
In summary, blood service operations can be enhanced by using IoT-based appli-
cations. Tracking is used to operate the inventory counting and location tracking
as well as to prevent the potential mistakes in blood transfusion. Identification
and authentication function is integrated with the RFID technology in order to
ensure the standard of the transfusion administration. Moreover, the IoT-based
technologies allow the data collection to be conducted automatically, particu-
larly the real-time blood inventory information at each storage location. Blood
inventory tracking and tracing data in real-time could systematically enhance the
collaboration processes between the RBC and the HBBs. Finally, sensing is aimed
to improve the quality of the blood supply chain management by monitoring the
148
temperature and locations of the blood products during transportation as well as

monitoring post-transfusion reactions of the patient. The conceptual guidelines
of transforming blood supply chain management with the IoT-based applications
can be summarized in Table 1.
Table 1: Transforming blood supply chain management with IoT paradigm
IoT Applications Blood Service Utilization

Operations
Inventory count Shortage minimization

Tracking Location tracking Transshipment
utilization
Safety and traceability Safety and quality
issues
Identification and Auto ID/bar code Safety of patients

authentication enabled transfusion
administration
Inventory Performance
Automatic data
management improving
collection
Blood tracking and Outdated
tracing minimization
Logistics management Performance
improving
Transportation Quality of blood

Sensing
monitoring
Patient monitoring Safety of patients
149
5 Implementation Strategies
According to the aforementioned guidelines of transforming the blood supply

chain management, this paper can briefly provide the implementation strategies
in two issues, resource management and big data and knowledge management.
5.1 Resource Management to maximize blood utilization
Applying the IoT-based healthcare applications in the blood supply chain can
enhance the performances in managing blood and blood products more system-
atically. The tracking function can maintain and control real-time blood stock
levels in the hospitals, leading to the accurate blood collection planning. It allows
the RBC to replenish the blood units and fulfill demands from the local blood
banks more appropriately. This contributes to blood outdated rate reduction
in the system. The connection of RFID and barcode can be used to manage the
unused blood units in the hospitals by tracking down the remaining shelf-lives
of blood units in the inventory. This could help reducing the number of blood
shortage situations in the area. These proposed IoT-based applications could
lead to more efficient blood rotation policy, where any unexpired blood units can
be sent back to the RBC for further transshipping to the other hospitals. Moreover,
it also yields the decreasing rate of blood expiration in the system.
5.2 Big Data and Knowledge Management to support decision

makings in blood supply chain
Using automatic data collection and transferring enables the blood management
system to store a large amount of useful data in the blood service organizations. In
developed countries, a blood-banking system which links between blood centers
and hospitals can be performed automatically to communicate or make decisions
on blood service operations. However, in some countries where blood services
still rely on personnel judgments, it is necessary to have information stored in
the blood management system to use as a knowledge tool for supporting future
decision-makings in any levels such as blood allocation policy, blood inventory
management, and blood crossmatching policy. The extended network system is
used to connect between blood service organizations and IoT-based blood service
applications in order to enhance collaborative information sharing.
150
6 Challenging Issues
Figure 2: The IoT-based framework of the blood supply chain management
IoT-based technologies can be used to manage blood and blood products in order
to minimize blood shortage and outdates rates. Blood service operations are
systematically driven by the information together with the big data and knowledge
management in order to support the operational decision makings in the blood
supply chain management. The IoT-based framework of the blood supply chain
management is presented in Figure 2.
The IoT-based blood supply chain management emphasizes the interactions be-
tween the blood service organizations. Blood service operations can be enhanced
through the advantages of using these intelligent technologies which directly
increase the blood utilization in the system. However, there are some challenging
issues to take into consideration in order to transform the blood supply chain
management with the IoT paradigm. These issues are briefly introduced as fol-
lows.
151
6.1 Barriers
Blood supply chain is a challenging system to manage. Its overall operations are
complex which are associated with human, blood centers, hospitals, and patients.
In theory, there are several studies that focused on the use of operations research
in the blood operations problems. However, these works are too complex for ap-
plication in the real context. Currently, there are advanced medical technologies
that support various aspects of health services in order to improve the health
outcomes of patients as well as enhance performances of care delivers, including
Telemedicine, E-health, and IoT. Despite these benefits, Llewellyn et al. (2014)
addressed barriers to the adoption and implementation of the advanced medical
technologies. These barriers involve with the lack of encouragement from the
Department of Health, change issues in new operational procedures, skills of
practitioners, threats perception, leadership, infrastructure, project management,
lack of knowledge toward new technologies, and costs. However, blood service
operations depend largely on staff decisions which are also a main barrier in the
adoption of the new advanced technologies in the healthcare context. In order
to overcome these barriers, the empirical study on the advantages of the IoT
application in the blood operations system is required to guarantee that such
technologies are trustworthiness for investment in the real context.
6.2 Stakeholders’ Complexities
Although the structures of the blood supply chain networks of different countries
are not the same. But the main stakeholders commonly consist of regional blood
center (RBC) and hospital blood banks (HBBs). Practically, the RBC is responsible
for the blood collection, processing, production, and inventory management in
order to allocate blood units to the HBBs. Each HBB has to manage its blood bank
in order to provide blood units for the patients’ treatments. The complexities
among all stakeholders are involved with inter-organizational collaboration of the
associated blood service operations from the human to the patient. The regula-
tions and procedures in each stakeholder are different which make the IoT-based
technologies difficult to implement in the entire system. System components, de-
ployed technologies, and service applications should be well developed in order
to connect among the stakeholders with the restrictions of privacy information in
each organization.
152
Moreover, data sharing among the stakeholders is one of the most complexity
issues in the blood supply chain system. In practical, the information systems
among the stakeholders are not fully linked, because of the regulations and pro-
cedures in each organization. This problem causes the data transmission to be
discontinuous among the networks. However, the HBBs have to provide actual
demand data to the RBC so that the RBC in order to manage blood collection
properly. The proposed IoT-based applications will allow the data transmission
among the stakeholders to be real-time and more automatic. The visibility of
blood information will reduce the uncertainty in blood service operations as well
as improve blood management throughout the network.
The IoT-based applications can be applied to connect between the RBC and the
HBBs for real-time data sharing in order to support the decision making in the
blood service operations. Moreover, the IoT technologies can be used to manage
blood inventory more systematically. The IoT-based system will automatically
transmit the necessary information in order to mitigate the uncertainties in blood
operations management as well. Thus, the consolidation of the blood service or-
ganizations is a key success towards the implementation of IoT-based applications
in the blood supply chain management.
6.3 Architecture
The generic IoT-based system architecture consists of four layers which are sens-
ing, network, service, and interface. There are brief details of its components
to be addressed (Li, Xu and Zhao, 2015) as well as the issues in developing the
architecture of the IoT-based applications in managing the blood supply chain.
Sensing layer is able to track the environment and exchange information among
devices by using tags or sensors. In the sensing layer, there are necessary issues
to be concerned such as cost, size, resource, energy, deployment, heterogeneity,
network communication, and protocol. Network layer is aimed to aggregate data
from information technology infrastructures to transmit data in the environment.
Also, there are primary issues to be addressed such as network management
technologies, energy efficiency, quality of service, data and signal processing,
and security. Moreover, all of the service activities are performed at the service
layer such as information exchanging, data management, database, searching,
and communication. There are main components to take into considerations,
including service discovery and composition, trustworthiness management, and
application programming interfaces. In interface layer, many devices are involved
153
in the IoT-based applications. The compatibility issue of the various things should
be addressed for the connected interactions in the environment.
6.4 Standardization
Standardizations are primary key to the success in the implementation of the

IoT-based technologies in the system, including the blood supply chain. The
standardizations must deliver interoperability, compatibility, reliability, and ef-
fectiveness of the operations in the environment. The RBC and the HBBs have to
closely collaborate in determining policies and system architecture, ensuring the
privacy, acceptability, security of networks, and developing joint standards in the
IoT-based blood supply chain management system.
6.5 Security and Privacy
The architecture of the IoT system has an impact on the security and privacy of the
stakeholders. The security requirements in the IoT system are addressed (Fabian
and Gunther, 2007), including resilience to attacks, data authentication, access
control, and information privacy. Moreover, information privacy is one of the
most sensitive topics in the IoT area. Privacy Enhancing Technologies have been
developed to provide information privacy such as Virtual Private Networks, Trans-
port Layer Security, and Domain Name System Security Extensions. Stakeholders
should contain these security and privacy requirements in their implementation
plans for managing the blood supply chain with the IoT-based technologies.
7 Conclusions
IoT-based healthcare applications paradigm consists of four main functions, track-

ing, identification and authentication, automatic data collection, and sensing.
This study has incorporated the IoT paradigm to provide the guidelines to trans-
form the blood supply chain management in order to improve blood service
operations. Tracking is used to monitor blood inventory counting and storage
location. Identification and authentication is the function aimed to ensure that
the blood transfusion standard is followed. Automatic data collection is used as
a mechanism to drive the information flows in blood logistics operations. The
154
References
use of sensing can improve the quality of blood during transportation as well as
monitor the patient for irregularities during postoperative blood transfusion.
Furthermore, the implementation strategies are provided to point out the issues
that blood utilization and blood operational decision makings can be improved.
The IoT-based framework of the blood supply chain management is proposed to
conceptualize the interactions between blood products, donors, patients, prac-
titioners, processes, and service operation applications through the network
connection. The challenging issues are discussed to provide further details in
related aspects. The consolidation of the blood service organizations is a key
success towards the implementation of IoT-based applications in the blood supply
chain management in the real context.
References
Bui, N. and M. Zorzi (2011). “Health care applications: a solution based on the Internet of Things”.
In: International symposium on applied sciences in biomedical and communication technologies.
Ed. by ACM. New York, p. 131.
Dobbin, J., R. Wilding, and S. Cotton (2009). “True blood: challenges of the blood supply chain in
England”. In: Logistics & transport focus 11.11, pp. 32–36.
Doukas, C. and I. Maglogiannis (2012). “Bringing IoT and cloud computing towards pervasive
healthcare”. In: International conference on innovative mobile and internet services in ubiquitous
computing (IMIS), pp. 922–926.
Dzik, W. H. (2007). “New technology for transfusion safety”. In: British journal of haematology 136.2,
pp. 181–190.
Feki, M. A., F. Kawsar, M. Boussard, and L. Trappeniers (2013). “The Internet of Things: the next
technological revolution”. In: Computer 46.2, pp. 24–25.
Fernandez, F. and G. C. Pallis (2014). “Opportunities and challenges of the Internet of Things for
healthcare: systems engineering perspective”. In: International con-ference on wireless mobile
communication and healthcare - transforming healthcare through innovations in mobile and
wireless technologies (MOBI-HEALTH), pp. 263–266.
Jara, A. J., A. F. Alcolea, M. A. Zamora, A. F. G. Skarmeta, and M. Alsaedy (2010). “Drugs interaction
checker based on IoT”. In: Internet of Things (IOT), pp. 1–8.
Kang, Y., M. Han, K. Han, and J. B. Kim (2015). “A study on the Internet of Things (IoT) applications”.
In: International journal of software engineering and its applications 9.9, pp. 117–126.
Koshio, A. and M. Akiyama (2009). “Blood transfusion and patient safety with IT - minimizing risk
of transfusion with Point-of-Act-System”. In: International conference of the Asia-Pacific Medical
Informatics Association, pp. 46–53.
Li, X., R. X. Lu, X. H. Liang, X. M. Shen, J. M. Chen, and X. D. Lin (2011). “Smart community: an
Internet of Things application”. In: IEEE Communication magazine 49.11, pp. 68–75.
155
Nagurney, A., A. H. Masoumi, and M. Yu (2012). “Supply chain network operations management
of a blood banking system with cost and risk minimization”. In: Computational management
science 9, pp. 205–231.
Pierskalla, W. P. (2005). “Supply chain management of blood banks”. In: Operations research and
health care, pp. 103–145.
Rytila., J. S. and K. M. Spens (2006). “Using simulation to increase efficiency in blood supply chains”.
In: Management research news 29.12, pp. 801–819.
Schreier, G. (2010). Pervasive healthcare via the Internet of Medical Things. Graz, Austria: Austrian
Institute of Technology GmbH.
Tarouco, L. M. R., L. M. Bertholdo, L. Z. Granville, L. M. R. Arbiza, F. Carbone, M. Marotta, and J. J. C.
de Santanna (2012). “Internet of Things in healthcare: interoperability and security issues”. In:
conference on communications (ICC). Ed. by I. International. Ottawa, pp. 6121–6125.
Vilamovska, A. M., E. Hattziandreu, R. Schindler, C. Van Oranje, H. De Vries, and J. Krapelse (2009).
RFID application in healthcare - scoping and identifying areas for RFID deployment in healthcare
delivery. Tech. rep. RAND.
156
E-Cash & E-Vouchers: the Digitalization
of the Humanitarian Aid and Logistics
Loïc Cohen1 , Vincent Salaun1
1 – CRET-Log Université Aix Marseille
In the recent years, the mobile phone coverage expansion and the more widespread
use of cash cards have eased the implementation at scale of Cash and Vouchers
(C&V) programmes in the humanitarian sector. Furthermore, the development of
web-based interfaces, designed to respond to the specifics needs of the sector,
supports the implementation of C&V programmes. Due to the technological in-
novations, C&V assistance is growing rapidly and the sector seriously considers
the ‘’e-assistance” as an alternative to the in-kind assistance. This paper aims to
investigate the potential impact of this digitalization on the humanitarian supply
chain. If the literature states that C&V programming reduces cost, brings speed
and serves better the needs of the beneficiaries, no real reference is made on
the impact on the humanitarian supply chain. To explore this gap, the research
uses the punctuated equilibria theory proposed by Eldredge and Gould (1972)
and adapted for Management Sciences by several researchers such as Trush-
man and Romanelli (1985) or Rowe and Besson (2011). Having closes similarities
with the humanitarian supply chain, a parallel is drawn with the event supply
chain, which stands at a stage further in the integration of new technologies and
has consequently re-shaped its structures and strategies. Through a qualitative
methodology based on participant observations, institutional communications
and secondary data, this paper highlights how C&V supported by digitalization has
created a rupture in the role of the humanitarian logistics and how humanitarian
organisations search to find a new equilibrium between strategy of intervention
and supply chain strategy.
Keywords: Humanitarian logistics; Event logistics;Cash & Vouchers (C&V)
157
E-Cash & E-Vouchers: the Digitalization of the Humanitarian Aid and Logistics
1 Introduction
Today, progress in new technologies raises a strong interest from donors, practi-
tioners and governments as to how the technology can best serve humanitarian
responses. In the recent years, Cash and Voucher (C&V) assistance is growing
rapidly and many humanitarian agencies start to implement extensively this type
of project as an alternative to in-kind assistance (Kovács, 2014). If there is cur-
rently no systematic tracking of the volume of humanitarian assistance delivered
in the form of C&V, a report from ODI, (2016) suggests that in 2015, out of a total
spending of 24 billion on humanitarian aid at least $1.9 billion was spent in the
form of cash-based responses (51% cash and 49% vouchers). United Nations
agencies accounted for around two-thirds of the total and non-governmental
organisations (NGOs) for just under a third. At the World Humanitarian Summit in
2016, many of the world’s largest humanitarian donors and agencies made a set
of commitments to increase the use of cash-based programming (ODI, 2016).
If for a non-expert of humanitarian assistance, the C&V programming sounds like

an anecdote, it is important to show how the digitalization of aid through C&V
implies a change of paradigm for the humanitarian logistics. Since a long time,
the traditional humanitarian aid is based on an ultime supply chain, as described
by Mentzer et al. (2001), covering from the needs assessment to the distribution
to the recipients. The C&V programming creates a significant break with this
traditional model because it delegates all the downstream logistics operations
to local traders. In this way, the C&V programming implies for the NGOs to move
from a traditional in-kind assistance to a new financial and digitalized assistance,
this is a move from physical flows management in the past, to information flows
management in the future.
In this context of potential revolution for the NGOs, this research tries to un-
derstand what sort of impacts could have the C&V programmes by posing the
following research question: How the C&V programmes can modify the logistics
strategy of humanitarian NGOs?
This interrogation was raised by observations, from one of the authors involved
in the sector as a logistician, of fast-paced changes in the humanitarian sector.
Furthermore, this topic is often discussed in the different logistics coordination
platforms by practitioners, as they feel that they have to redefine the logistics
function because their role is challenged by a new type of assistance (Logistics
cluster 2014).
158
2 From humanitarian logistics to event logistics
To reply to this research question, this communication is divided in three sections.

The first one highlights the principles of C&V programmes and humanitarian logis-
tics, and draws a parallel with the event logistics using the punctuated equilibrium
theory (Eldredge and Gould, 1972; Trushman et Romanelli, 1985), to show from a
theoretical point of view how C&V programmes can change the NGOs logistics
strategy. Subsequently, the second point focus on the qualitative methodology
used for this research. The last point exposes a discussion of the results.
2 From humanitarian logistics to event logistics: a

theoretical reading with the punctuated equilibrium
In order to establish a theoretical reading of the impact of C&V programmes on the

logistics strategy of aid agencies, this part is segmented in four points: a review
of the principles of C&V assistance (2.1), a literature review on the characteristics
of humanitarian logistics (2.2), a parallel with another temporary logistics set-up,
the event management and specifically on music festivals (2.3), to finish a reading
by the punctuated equilibrium theory of the conclusions (2.4).
2.1 Cash & Vouchers programmes: the principles and operating
ECHO (2013), defined Cash-based responses as mechanisms to provide resources

to a population in two main ways:
— Vouchers provide access to pre-defined commodities or services. They
may be denominated in cash, commodities or service value. Vouchers
come with some restrictions and must be exchanged for a given com-
modity or service. Vouchers are often grouped under the heading of
‘cash-based responses’, but they differ from cash transfers.
— Cash transfers are the provision of money to individuals, households,
either as emergency relief intended to meet their basic needs for food,
non-food items or services essential for the recovery of their livelihood.
Beneficiaries decide how to use the cash received.
According to Doocy and Tappis (2016), in the different studies reviewed, C&V
assistance appears to be more efficient to deliver when compared to in-kind
assistances. However, the overall cost-efficiency of C&V depends on the prices
159
paid by beneficiaries for the commodities in local markets compared to the cost
for the humanitarian organisations to procure, transport, store and distribute
the In-kind assistance (Venton et al, 2015).) The advantages of C&V have been
segmented in three categories by WFP (2017). The first one, it’s fast, efficient, and
generally secure, by reducing the cost and logistical complexity, C&V shortens the
path to deliver assistance. The second advantage is the C&V programme’s ability
to offer to beneficiaries a greater choice to purchase what they need most. And
the last one concerns the potentiality to stimulate trade by injecting cash into the
local economy, and to create a virtuous circle of production and consumption.
As stated by ECHO (2013), ”In certain humanitarian disasters, the supply of food
and non-food items to markets is organized and sustained. In such context, the
cash and vouchers project gives the aid organisations the mean to reach timely
and directly the affected population”. However, despite the rapid growth and the
clear advantages, C&V programmes are not suitable in all contexts. C&V assistance,
or the combination with the In-kind assistance, should be defined by a context-
specific assessment which analyses the availability of goods and services, the
possible impacts on markets, the cost effectiveness, the technologies to support
the transfer, the processes for targeting beneficiaries, the security and corruption
risks.
Providing assistance to the peoples in need through cash and vouchers is not new;
This mean of aid exists for over a century (Devereux 2006), but in the past, the C&V
assistance implementation was hampered by the lack of technological solutions.
In the recent years, the evolutions of the mobile telephony and electronic payment
systems have allowed to switch from physical distributions to e-transfers and have
facilitated the implementation of C&V programmes. There are two different forms
of e-transfers used most frequently: The e-cash and the e-vouchers. A synthesis
of differences and commons between e-cash and e-vouchers is suggested in the
table 1.
As shown in this point, the e-cash and e-vouchers programmes are based on a
transition from in-kind aid to a financial, direct or indirect, assistance allowed by
the development of web-based interfaces and by adapting information technol-
ogy to the crisis contexts. This kind of programme potentially implies for the NGOs
to modify their deep organisation and more specifically their logistics strategy
to move from a physical-based flow management to an information-based flow
management. To continue our research, it seems important to return to the basics
of humanitarian logistics to identify the potential impacts of e-transfers on the
NGOs logistics strategies.
160
Table 1: Differences and commons between e-cash and e-voucher
e-Cash e-Voucher
Program Uses Like cash, e-cash allows Like paper vouchers,
beneficiaries to buy which evouchers are used in
commodities are most, crit- programmes designed
ical to them, and where and to increase access to a
when they want to spend particular set of goods or
their money. services.
Service Requires a local service Can be selected from a list
Providers provider (bank, mobile of pre-qualified e-voucher
network, etc.) that can service providers
issue and accept e-cash.
Agent/Vendor Relies on a network of trans- Requires to establish a lo-
Networks fer agents or ATMs that ac- cal vendors network. Also
cept e-cash in exchange for requires to equip vendors
goods, services or physical with hardware.
cash.
Availability Systems are local. Sub- Systems are global and can
and ject to banking and anti- be used anywhere with
Regulations money laundering regula- minimum infrastructure
tions, some of which can re- requirements. Does not
quire official forms of ID for rely on, or link to, local
programme participants. financial services. Not
subject to local banking or
financial regulations.
Adapted from Mercy Corps e-Transfer Guide (2014)
161
2.2 Some basics about humanitarian logistics
When it comes to “humanitarian logistics”, the most commonly used definition is:
“Humanitarian logistics is the process of planning, implementing and controlling
the efficient, cost-effective flow and storage of goods and materials, as well as
related information, from point of origin, to point of consumption for the purpose
of meeting the end beneficiary’s requirements” (Thomas and Mizushima, 2005).
To perform this task, the humanitarian logistics conducts needs assessments and
logistics planning, manages the procurement cycles, organises the transports,
performs customs clearance, manages warehousing and distributes the assistance
to the beneficiaries.
As highlighted by Pettit and Beresford (2009), the supply chain management is
often the most complex element of a humanitarian operation. The success or
failure of such operations depends on understanding and addressing the issues
of the supply chain. The difficulty of an emergency response is how to get the
right quantities of relief supplies, to a determined location, in a minimum of time.
This is the responsibility of the humanitarian logistics and to do so, humanitarian
organisations have to put in place a complex and temporary network (Salaun,
2016) with multiples actors in charge of different nodes.
Many authors (Beamon 2004; Van Wassenhove, 2006; Kovacs and Spens, 2007;
Day et al., 2012) have highlighted the Humanitarian Supply Chain Management
characteristics and have identified five major components:
1. The unpredictability of the demand, which makes forecasting difficult, in
terms of timing, location and volume.
2. The operational context, which makes very complex the effective flow of
goods and materials.
3. The imperatives associated with the timeliness for a wide variety of sup-
plies
4. The recurrent lack of resources in terms of supplies, expertise, technology,
funding.
5. The supply chain is dynamic, temporary and short term.
The humanitarian sector regroups a wide panel of different types of organisations
which range from NGOs, United Nations agencies, International organisations, etc.
Although, if they operate in the same contexts, the place of the logistics within
the organisation and the level of logistical capabilities vary from one organisation
to another (Thomas & Kopczak, 2005). Some organisations consider the logistics
162
as a central function, while some other organisations often neglect its role in the
strategic planning (Kovacs & Spens, 2009). Various factors, such as their size, man-
date, structure, funding and area of expertise determine their deep organisational
structure and operational boundaries, which impact on their logistics strategy
(Cohen, 2016).
Humanitarian logistics is caught in a vicious circle (see figure 1) where a lack of

understanding and recognition of the importance of the function precludes it
from planning and funding processes, leading logistics not to be able to fulfill its
objectives and obligations.
In a sector where the logistics has been suffering for long from a lack of investment
and has shown some chronical deficiencies, humanitarian logistics researchers
are concerned with the implications and the complexity of the physical flows
management. This includes the number of points at which the products are
handled, moved and stored, regardless of the number of actors, transfers of
responsibility and liability in the logistics process.
Moreover, Humanitarian organisations compete to access funding, and donors

are demanding greater accountability and performance of the logistics activities.
Since logistics accounts for about 60% of a relief operation (Van Wassenhove,
2006), improving its performance would surely give a ‘’competitive advantage” to
organisations to access funding. As the basic principles of managing the flows of
goods, information and finances, which are valid for the commercial logistics, are
also valid for the humanitarian logistics (Kovacs and Spens, 2007), one envisaged
way by certain organisations to improve the performance of their supply chain, is
to delegate the management to Logistics Services Providers (Balcik et al., 2010).
But from a contract management standpoint and given the unpredictable nature
Figure 1: The vicious circle of the humanitarian logistics from Van Wassen-
hove(2006)
163
of relief operations, it could be complex to delegate the management of the supply

chain to logistics services providers (Cohen, 2016).
A closer look to this problematic reveals that the solution of improving the ef-
fectiveness of these supply chains may not be in the hands of these logistics
operators, but maybe in the shift on how the assistance is delivered. Indeed, the
e-assistance dematerialized and transformed the in kind-assistance from physical,
informational and financial flows to informational and financial flows only. The
complex physical flow management and the final distribution are delegated to
local traders, who become de facto a new type of services providers. By the elimi-
nating or transferring many logistical activities to third parties, C&V assistance
modifies the humanitarian supply chain. It simplifies the complex procurement
cycles, although there is still contracts management to do. But most of all, it is an
opportunity to disengage from the transport and warehousing management and
to reduce distribution implications. This shift on aid delivery mechanisms poten-
tially leads to a reconfiguration of the logistics function (Heaslip 2013). However,
this thinking stays an assumption and the lack of research about the impacts of
C&V programmes on NGOs logistics strategy requires an extension of this research
to strengthen our theoretical conclusions. For this, it’s possible to establish a
parallel between the humanitarian logistics and another specific logistics, which
is also engaged in the digitalization: the music festivals logistics.
2.3 Toward another digitalized logistics: the event management
The e-cash and e-vouchers programmes are new for practitioners and researchers,
so it is difficult to identify the potential impacts of C&V programmes on the hu-
manitarian logistics. However, it is possible to use the example of a near industry,
which currently lives the same case of digitalization than the humanitarian aid.
For few years, the event management sector is characterized for using more and
more information technology. To illustrate this fact, we can look specifically on
the music festivals industry which uses the cashless technology since the early
2010’s (Jackson, 2014).
At first glance, humanitarian and event logistics can be seen as very different
industries: different aims, different stakeholders, different ”customers”, etc. How-
ever, as shown by Salaun (2016), if the stakeholders and the aims of these two
industries are distinct, their deep structures are very similar: a permanent struc-
ture with a large network and a wide variety of actors to implement temporary
164
projects (relief for humanitarian and festival for event). In fact, if we look at the
major components of the humanitarian logistics highlighted previously, we can
identify a lot in common with the event logistics. As shown by Lexhagen et al.
(2005) and Locatelli and Mancini (2014), in event management, the ‘’operation
excellence” is one of the biggest factors of success for any events. Even if these
authors don’t give any exact definition of ”operation excellence”, it looks possible
to associate this notion to the logistics performance. Like in the humanitarian
context where the logistics is seen as essential, the logistics is crucial in event
management. Moreover, according to Kerzner (2013), the two major reasons of
an event failure are, the incapacity to respect the time restriction of the event,
and the difficulties to address consumers’ demands due to a lack of resources
or goods. Also, we can find two additional similarities between event and hu-
manitarian industry: the short term cycle of the supply chain and the lack of
resources. In addition, O’Toole (2000) reports one other characteristic for event
logistics : the operation management complexity due to a wide variety of supplies,
which is another similitude with the humanitarian logistics. According to these
authors, event and humanitarian logistics are very close and are based on the
same principles. This proximity between the two logistics permits to build these
basic inferences:
— Humanitarian and event logistics are very similar.

— Event logistics has been impacted and continues to be impacted by an
innovation, the cashless.
— Given the similarities with the event logistics, humanitarian logistics can
also be impacted by innovations.
Before developing this idea, it is important to explain what is the cashless and
how this innovation has strongly impacted the event logistics in music festivals.
Used for a decade (Jackson, 2014), the cashless constitutes a personal and virtual
wallet assigned to each festival-goers. This virtual wallet is linked to the bank
account of the festival-goers who can put at any moment credits on it and which
is materialized by a RFID or NFC tag included on a card or a bracelet. When a
festival-goer wants to buy something, he presents his RFID tag to a terminal to
pay (Dowson and Bassett, 2015). To our knowledge, the impacts of cashless
on event logistics have been addressed only on two studies. The first one by
Jackson (2014) shows how the cashless produces new information flows. The
second one, leaded by Salaun (2017) highlights the capacity of the cashless to
impact festivals logistics strategy on three points: the logistics processes (1), the
165
warehousing (2), and the flows management (3). According to Jackson (2014)
and Salaun (2017), in music festivals management, the cashless has impacted
strongly the logistics strategy of the event and encouraged new logistics practices,
such as subcontracting logistics services providers (Salaun, 2017). To explore
the potential impacts of C&V programmes on NGOs logistics strategy, it could
be pertinent to use the example of the cashless on event logistics: From a small
digitalization to a big logistics revolution.
2.4 A reading of digitalization through the punctuated equilibrium

theory
As shown in the previous point, the humanitarian and event logistics, at least
in music festivals, present similarities. Therefore, it is possible to imagine that
the impacts of the cashless on event logistics can potentially be transposed as
the future impacts of C&V programmes on humanitarian logistics. In both sit-
uations, the opportunities permitted by the information technology evolution,
lead to rethink the deep structure of organisations and to develop a new logistics
focus on information flow management instead of physical flow management.
In order to propose a framework to study the case of humanitarian logistics and
C&V programmes, we need to embed our research in a stable theoretical frame
in order to develop cross-sectorial extrapolations from event to humanitarian
management.
The rapid development of cashless and C&V programmes which appears to be the
consequences of fast evolutions of information technology used gradually by the
actors of these two industries, suggests to use an evolutionist point of view, and
more specifically the punctuated equilibrium theory. Historically, developed in
the biology field by Eldredge and Gould (1972), the punctuated equilibrium theory
proposes a gradual evolution and rejects the idea of a permanent adaptation to
the environment. The evolution of organisms is based on the research of homeo-
static equilibrium during long periods, punctuated by short and sudden periods
of radical modifications linked to important changes of the environment. Adapted
to management by Trushman and O’Reilly (1996), the punctuated equilibrium
theory highlights how organisations quest for stability and homeostatic equilib-
rium, punctuated by swift periods of rupture and the re-reshape of the structure.
This theory looks relevant to study the potential impacts of C&V programmes on
humanitarian logistics strategies.
166
Even if this theory was developed in biology, the academic community of re-
searchers on management uses frequently the punctuated equilibrium. For ex-
ample, we can point out the works of Day (2014) and Stevens and Johnson (2016)
who referred to the punctuated equilibrium on humanitarian logistics researches.
One of the most important contribution of this theory is to show how the evolu-
tion cycles based on fast periods of rupture can modify the organisation deep
structure (Gersick, 1991). According to Silva and Hirschheim (2007), the organisa-
tion deep structure can be splitted in four major components: the organisational
culture (1), the distribution of power (2), the organisation (3), and the control and
management system of performance (4). By comparing different models from
the literature, including the model of Trushman and O’Reilly (1996), Besson and
Rowe (2011, 2012) describe the evolution cycle in four steps:
1. Unfreeze, which is a rupture of the homeostatic equilibrium and a de-

crease of the organisation inertia.
2. Move, is a step of exploration and the emergence of a new deep structure.
3. Refreeze, corresponding to the stabilisation and the anchoring of the new

deep structure.
4. Convergence, is the last step during which the organisation searches the
optimisation and the routinisation of the new equilibrium.
According to the model of Besson and Rowe (2011, 2012), and with the information
given by the literature about cashless in music festivals, it is possible to consider
that music festivals are currently in the third step of the cycle of evolution, the
refreeze. Indeed, in music festivals the organisational culture is modified to adopt
the cashless technology. Some activities are subcontracted, as a power delegation.
The organisations have been reconfigured to include the new information flow,
and new ways of control and performance management have been developed
and implemented. As per Silva and Hirschheim (2007), this description shows a
modification of the deep structure of the music festivals organisations in order to
find a fit with the new environment and opportunities. Concerning the impact
of C&V programmes on humanitarian logistics, at this stage there is not enough
information provided by the literature. Therefore, It would require a new empirical
exploratory study to identify in which step of the cycle the humanitarian logistics
is positioned, and then to use the analogy with the cashless to think ahead the
evolution of the humanitarian logistics in the coming years.
167
3 Methodology of empirical study
In order to position the humanitarian logistics on the evolution cycle shown by

Besson and Rowe (2011, 2012 and to try to foresee the impacts of C&V programmes
on humanitarian logistics, this research is based on a qualitative methodology.
This includes participant observations with the direct involvement of one of the
researchers as practitioner in the humanitarian sector, who has analysed the im-
pact of C&V programming on the logistics through the lens of the researcher /actor.
As a member of a NGO involved in C&V programmes, the author has conducted
discussions on this topic with different stakeholders in November 2016 in Nigeria,
to understand how the actors perceive the impact of C&V on the logistics. Also, to
support the analyse of empirical data, the results of a survey conducted in April
2014 by the Global Logistics Cluster (Logistics Cluster, 2014) to senior logisticians
from different organisations (87 respondents from 24 organisations) have been
used to capture more widely the perception from practitioners. In addition, to
increase the validity of this research, two other ways of data collection were used:
a collection of secondary data from academic and professional researches, and
the exploitation of institutional and commercial communications from major C&V
actors. In line with the exploratory nature of this research, the data were trian-
gulated and exploited with the preconisation of Miles and Huberman (1984), in
particular with the different matrix of their methodology. According to the target
of the authors to highlight new hypothesis for future works, this research doesn’t
pretend to show an objective reality but just the perception from practitioners.
4 Findings and discussion
To expose the results, this last part is segmented in three points representing
the evolution of the findings. A first point positions the humanitarian logistics in
between the first and the second step of the cycle shown by Besson and Rowe
(2011, 2012) (4.1), a second point develops an analogy with the impact of cashless
in event logistics (4.2), a last point proposes a prospective view of possible impacts
of C&V programmes on the humanitarian logistics (4.3).
168
4 Findings and discussion
4.1 The humanitarian logistics in between the first and the second
step of the evolution
For diverse reasons, the humanitarian logistics suffers from chronical deficiencies
and this represents a challenge for the sector. In addition, the rapid technology
evolution offers the opportunity to digitalize the relief assistance and to deliver
it by another mean, which has multiple advantages over the traditional in-kind
assistance. The combination of these two factors pushes the organisations to
break a long lasting equilibrium that they have built for decades, with the In-kind
assistance model. Since, the sector has integrated this break and organisations
realise the opportunities offered by the technology. Numerous aid agencies
are developing new internal systems, capacities and ways of working for C&V
programmes implementation (CaLP 2014).
This rupture engaged by humanitarian organisations corresponds to the stage
One of the evolution cycle, the Unfreeze period.
In this move, the logistics sector tries to follow the pace and to clarify its role
toward C&V programmes. The survey from the Global Logistics Cluster (Logistics
Cluster, 2014), shows that humanitarian logisticians realise that their role will be
transformed or diminished, and they expressed concerns about how their skills
will be used in the e-assistance supply chain.
All these changes and the search from the actors to take position in the emergence
of a new deep structure, clearly mark the entrance of the sector into the second
stage which is the Move period defined by Besson and Rowe (2011, 2012). What we
need to discuss now, is how this exploration/evolution phase will transform the
humanitarian supply chain, which despite its deficiencies remains a key function
in traditional in-kind assistance.
4.2 Using the experience of the event logistics to imagine the future
of the humanitarian logistics
If the humanitarian logistics is currently in between the first and the second step
of the evolution cycle from Besson and Rowe (2011, 2012), studies from Jackson,
(2014) and Salaun, (2017) show how the event logistics can be positioned on
the third step of cycle, which marks the evolution of the deep structure and
the institutionalisation of new logistics practices. To illustrate the modification
169
of the deep structure in event logistics due to the cashless technology, we can
analyse information about one of the major music festivals in France (herein
called Alpha Festival to conserve its anonymity). After only two editions with the
cashless technology, the Alpha Festival has changed a large part of its logistics
structure, including new access systems for the festival-goers, new management
of food and beverage supplies, reduction of the waiting time to buy something
during the event for the festival-goers, etc. All of this comes from the cashless
technology and the possibilities offered by the new information flow. This new
approach leads the organisation to modify at least three components of the
deep structure: the organisation (new repartition of resources and competencies
between the services), the management system of performance (new data for
real-time management), and the delegation of power inside the organisation (new
cashless technology suppliers and new information flow management specialists).
A study of these two last points highlights the emergence of new actors inside the
network. A new actor started with a ”simple” solution of cashless few years ago,
but nowadays this same firm offers the possibility for the events to deploy cashless
solutions with the back office software for the information flow management. As a
logistics services provider who built and manage a full supply chain for a customer,
the cashless provider offers a full package of solutions for event supply chain
management. Based on this example of the cashless impact on event logistics, we
can imagine a similar trajectory for the humanitarian logistics with the apparition
of new actors in the future step of the evolution.
4.3 What humanitarian logistics in few years?
By creating a concept of e-assistance, C&V programmes can be considered as a

major break in the humanitarian sector. To reach beneficiaries, C&V responses
will increasingly rely on financial services providers, technology infrastructures
and private sector capacities, where these exist. The growing demand for these
services will in turn lead to an increasing number and diversity of enterprises
seeking to engage with humanitarian agencies. As such, the traditional humani-
tarian actors will progressively forge new working relationships and the private
sector will likely create further demand for its expertise by developing products
and services designed to humanitarian C&V responses.
Thus, the aid agencies will have to rethink their strategy and to re-shape their
organisational structure. This will strongly impact the supply chain and will lead to
a redefinition of the logistics function. By this move to digitalization, organisations
170
5 Conclusion
are transferring many logistical activities to third parties and they rely on the local
market to transform the e-assistance into in-kind assistance. The technology
interface virtually pilots the supply chain and the physical flows are handled by
local traders who become de facto a new type of logistics services providers.
As well as in the music festivals sector, the humanitarian organisations will modify
three main components of the deep structure. The organisational structure with
a new repartition of resources and competencies within the organisation (less
logistics, more C&V specialists), the performance evaluation (data available for a
real-time monitoring), and the delegation of power to new actors (C&V technology
suppliers and information management specialists). Consequently, as stated
in the Global Logistics Cluster survey (Logistics Cluster, 2014), the logisticians
perceive their role narrowed to market assessments, evaluation of local traders’
supply chains capacity and contracts management with these new actors. But if
the logisticians realise that their traditional role is challenged, they can also see
an opportunity to perform better on the in –kind assistance delivery, considering
that a part of the logistics burden has been removed due to C&V programmes.
Consequently, more logistics capacity could be allocated to the in-kind assistance
(Logistics Cluster, 2014).
Nonetheless, if the e-assistance will represent a consequent share of the human-

itarian activities, C&V depends on the markets and the technology, it can’t be
generalised. The traditional logistics supporting the in-kind assistance will always
be required where markets and technology infrastructures are dysfunctioning.
In this perspective, we can imagine that the organisations deep structure would
take an ambidextrous form, with the objective to maintain an equilibrium be-
tween exploitation and exploration in the different facets of the humanitarian
assistance.
5 Conclusion
This research aims to highlight the potential impacts of C&V programmes on

humanitarian logistics. By an analogy with the event logistics and with a reading
through the punctuated equilibrium theory, this communication shows how the
C&V programmes would potentially impact strongly the deep structure of NGOs
and more specifically their logistics strategy.
171
Due to its exploratory nature, this research has important limitations. The first
one is linked to the choice of methodology based on participant observation and
exploitation of secondary data. To increase the validity of our conclusions, it
seems important to develop a new empirical study with multiple cases analysis
and maybe a longitudinal study to observe the real-time evolution of the logistics
strategy of organisations involved in C&V programming. Despite these limits,
this research is, to our knowledge, one of the first on this emerging topic of
humanitarian supply chain digitalization. It opens a door for future researches on
information flow management in crisis context with the e-assistance, the deep
structure of supply chains and the evolution of the humanitarian sector.
Currently C&V programmes can be seen as a shift from the traditional In-kind
assistance toward the e-assistance. In other words, it is a move from a logistics
of physical flow to a logistics of information flow. This may imply three main
modifications in the humanitarian supply chain. First, a focus on new core com-
petencies in information flow management with a delegation to new actors such
as C&V technology providers to pilot virtually the supply chain. Secondly, the
reconfiguration of the traditional logistics roles with the delegation of the down-
stream supply chain management to local traders. Thirdly, the emergence of an
ambidextrous supply chain with a redefinition of the logistics function in order to
manage both in-kind and e-assistance.
But in this evolution it is important to keep in mind that the rush on C&V pro-
grammes and a possible push from donors to deal directly with the private sector,
such as financial institutions and telecom companies, could backfire on aid agen-
cies. They might lose footprint and influence on the ground and their role could
be seriously challenged by new actors.
References
Balcik, B., B. M. Beamon, C. Krejci, K. Muramatsu, and M. Ramire (2010). “Coordination in hu-
manitarian relief chains: Practices, challenges and opportunities”. In: International Journal of
Production Economics 126.1, pp. 22–34.
Beamon, B. M. (2004). “Humanitarian relief chains: Issues and challenges”. In: San-Francisco.
Besson, P. and F. Rowe (2011). “Perspectives Sur le Phénomène de la Transformation Organisa-
tionnelle”. In: Systèmes d’Information et Management 16.1, pp. 3–34.
Besson, P. and F. Rowe (2012). “Strategizing Information Systems-enabled Organizational Trans-
formation: A Transdisciplinary Review and New Directions”. In: Journal of Strategic Information
Systems 21.2, pp. 103–124.
172
References
Cluster, L. (2014). Logistics in Cash & Markets Survey. Tech. rep.

Cohen, L. (2016). “The outsourcing decision process in humanitarian supply chain management
evaluated through the TCE and RBV principles”. In: Lausanne.
Corps, M. (2014). Cash Transfer Programming : E-transfer Implementation Guide. Tech. rep.
Day, J. M. (2014). “Fostering emergent resilience: the complex adaptive supply network of disaster
relief”. In: International Journal of Production Research 52.7, pp. 1970–1988.
Day, J. M., S. Melnyk, P. Larson, E. Davis, and D. Whybark (2012). “Humanitarian and disaster
relief supply chains: a matter of life and death”. In: Journal of Supply Chain Management 48.2,
pp. 21–36.
Devereux, S. (2006). “Cash transfers and social protection.” In: Johannesburg, pp. 9–10.
Doocy, S. and H. Tappis (2016). Cash-based approaches in humanitarian emergencies : a systematic
review. Tech. rep. London: International Initiative for Impact Evaluation.
Eldredge, N. and S. Gould (1972). “unctuated Equilibria: An Alternative to Phyletic Gradualism”. In:
Models in Paleobiology. San-Francisco: Freeman Cooper, pp. 82–115.
Gersick, C. J. G. (1991). “Revolutionary Change Theories: A Multilevel Exploration of the Punctuated
Equilibrium Paradigm”. In: The Academy of Management Review 16.1, pp. 10–36.
Heaslip, G. (2013). “Services operations management and humanitarian logistics”. In: Journal of
Humanitarian Logistics and Supply Chain Management 3.1, pp. 37–51.
Humanitarian Aid & Civil Protection (ECHO), E. C. D.-G. for (2013). The Use of Cash and Vouchers in
Humanitarian Crises. DG ECHO Funding Guidelines. Tech. rep.
Jackson, C. (2014). “The lived esperience o the popular music festival-goer”. PhD in management.
Bournemouth, UK: Bournemouth University.
Kerzner, H. R. (2013). Project Management: A Systems Approach to Planning, Scheduling, and
Controlling. en. 11e édition. Hoboken, N.J: John Wiley & Sons.
Kovacs, G. (2014). “Where next? The future of humanitarian logistics”. In: Humanitarian Logistics:
Meeting the Challenge of Peparing for and Responding to Disasters. 2 ed. London: Kogan Page,
pp. 275–286.
Kovacs, G. and K. Spens (2007). “Identifying challenges in humanitarian logistics”. In: International
Journal of Physical Distribution & Logistics Management 37.2, pp. 99–114.
Kovács, G. and K. M. Spens (2007). “Humanitarian logistics in disaster relief operations”. In: Inter-
national Journal of Physical Distribution & Logistics Management 37.2, pp. 99–114.
Lexhagen, M., H. Nysveen, and L. E. Hem (2005). “Festival coordination: an exploratory study
on intention to use mobile devices for coordination of a festival”. In: Event Management 9.3,
pp. 133–146.
Locatelli, G. and M. Mancini (2014). “Controlling the Delivering of Projects in Mega-Events: An
Application on Expo 2015”. In: Event Management 18.3, pp. 285–301.
Mentzer, J. T., W. DeWitt, J. S. Keebler, S. Min, N. W. Nix, C. D. Smith, and Z. G. Zacharia (2001).
“Defining Supply Chain Management”. en. In: Journal of Business Logistics 22.2, pp. 1–25.
Miles, M. B. and A. M. Huberman (1984). Qualitative data analysis: a sourcebook of new methods.
en. Sage Publications.
O’Toole, W. (2000). “Towards the integration of event management best practice by the project
management process”. In: Event Beyond 2000. Sydney.
Overseas Development Institute (2016). Counting cash: tracking humanitarian expenditure on
cash-based programming. Tech. rep.
173
Pettit, S. and A. Beresford (2009). “Critical success factors in the context of humanitarian aid
supply chains”. In: International Journal of Physical Distribution & Logistics Management 39.6,
pp. 450–468.
Pichault, F. and J. Nizet (2000). Les pratiques de gestion des ressources humaines. Paris: Seuil.
Salaun, V. (2016). “La pérennité des Organisations Temporaires (OT) : compréhension du rôle
conjoint de la pulsation organisationnelle et de la logistique - L’apport de l’étude des festivals
musicaux -”. PhD in Management. Aix-en-Provence: Aix-Marseille Université.
Salaun, V. (2017). “Which evolutions of event logistics? The impact of cashless in music festivals”.
In: Logistique & Management 24.3-4, pp. 207–214.
Silva, L. and R. Hirschheim (2007). “Fighting against Windmills: Strategic Information Systems and
Organizational Deep Structures”. In: MIS Quarterly 31.2, pp. 327–354.
Stevens, G. C. and M. Johnson (2016). “Integrating the Supply Chain … 25 years on”. In: International
Journal of Physical Distribution & Logistics Management 46.1, pp. 19–42.
The Cash Learning Partnership CALP (2016). Is Cash transfer programming fit for the future? Tech.
rep.
Thomas, A. and L. Kopczak (2005). From Logistics to Supply Chain Management: The Path Forward
in the Humanitarian Sector. Tech. rep. San-Francisco: Fritz Institute.
Thomas, A. and M. Mizushima (2005). “Logistics training: neccessity or luxury?” In: Forced MIgration
Review 22, pp. 60–21.
Trushman, M. and C. O’Reilly (1996). “Ambidextrous organizations : managing evolutionnary and
revolutionnary change”. In: California Management Review 38.4, pp. 8–30.
Van Wassenhove, L. (2006). “Blackett Memorial Lecture Humanitarian aid logistics: supply chain
management in high gear”. In: Journal of Operational Research Society 57, pp. 475–489.
Venton, C., S. Bailey, and S. Pongracz (2015). Value for money in cash transfers in emergencies. Final
Report. Tech. rep. UK Department for International Development.
World Food Program (WFP) (n.d.). Cash based transfers. Tech. rep. http://www1.wfp.org/cash-
based-transfers.
174
Business Model Innovation: A Case of
the Offshore Lifting Equipment Supplier
Mikhail Shlopak1 , Richard Myhre2 , Gabriele Hofinger Jünge3
1 – Møreforsking Molde AS
2 – AXTech AS
3 – NTNU
The purpose of this paper is to present the background for and the process of
development of an Internet of Things (IoT) business model (BM) for a Norwegian
offshore lifting equipment supplier. The paper presents both challenges and new
opportunities connected to the case company’s transition towards IoT, service-
based BM. The research methodology is a single case study research. The research
approach involves four steps: theoretical discussion; a case study; analysis of the
case study, and conclusions. The results show that development of an IoT BM by
the case company led to new possibilities for monetization of data and offering
new services. The authors suggest that manufacturing companies considering
transition towards IoT BM focus on the overall IoT business case rather than on
Local IoT solution return on investment (ROI). Studies analyzing the development
process and results of implementation of IoT BMs by manufacturing companies
are scarce. This paper aims to partially fill this gap by analyzing the experience
of a real-world company that has developed and implemented the IoT BM. The
research is limited to a single case company. Although the case company has
developed and implemented its IoT BM, it is still in the transition process. For now,
the company has not yet managed to get its new product rented out, which also
creates limitations for drawing conclusions. This research contributes to the un-
derstanding of IoT BMs and assists managers who are responsible for developing
and implementing IoT BMs.
Keywords: Internet of Things; Business models; Servitization; Engineer-to-Order
175
Business Model Innovation: A Case of the Offshore Lifting Equipment Supplier
1 Introduction
Developing a sound BM is one of the key factors for achieving success in any kind of
business. Many companies tend to focus on technology innovation while paying
less attention to BM innovation. Chesbrough (2010, p.354) argues that “mediocre
technology pursued within a great business model may be more valuable than
a great technology exploited via a mediocre business model”. This idea is also
shared by Amit and Zott (2012) who claim that managers should consider the
opportunities offered by BM innovation to complement, if not substitute for,
innovation in products or processes.
AXTech, a Norwegian company based in Molde, has been delivering customized
heavy-duty lifting and material handling equipment since 2004. The equipment
is produced in low volumes, is capital-intensive and is designed for specific needs
of customers in the oil and gas industry. This form of production, where the
customer order decoupling point is located at the product design stage, is called
engineer-to-order (ETO) manufacturing (Olhager, 2003).
A sharp decline in oil prices (from $115 per barrel in June 2014 to under $35
at the end of February 2016 (Rogoff, 2016)) has led to a significant reduction
of demand for advanced equipment from the companies operating in the oil
and gas sector. Under these circumstances, in the beginning of 2015 the focal
company has started to develop a new BM based on renting out of advanced lifting
equipment, as an alternative to a traditional model of just selling it. By offering
such a solution, the focal company tried to address the customers’ reduced ability
to invest money in the pricey equipment. The first piece of equipment produced
by the focal company under this new BM was a lifting tower Litjkaren that was
ready for market in November 2016. The tower has a lifting capacity of 25 tons,
is very fast to mobilize and can be steered remotely from the focal company’s
headquarters in Molde. The biggest challenge the focal company faces now is,
however, to reach the “right” customers and get Litjkaren to market.
The focal company’s management is therefore currently paying much attention to
further development and refinement of their new BM. In this respect, an Internet
of Things (IoT) BM development process suggested by Slama, Puhlmann et al.
(2015) as part of their “Ignite | IoT Strategy Execution” methodology was applied
to re-consider the focal company’s new “rental” BM.
The purpose of this paper is to describe the core elements of the IoT solution
the Litjkaren’s BM is based on. However, we have paid most of our attention to
176
2 Theoretical background
such elements of the IoT BM as marketing channels, customer relationships, cost

structure and revenue streams, local and overall ROI of an IoT solution, as well as
to nonmonetary effects of the new IoT BM.
The remainder of this paper is structured as following. In section 2, we provide a

definition of the main terms and concepts used in this research. In section 3, we
discuss our method. In Section 4, we present our findings and present how an IoT
BM development process was adopted by the case company. Finally, in Section 5,
we suggest implications for our research, outline the limitations of our study and
suggest further research.
In this chapter, we will briefly discuss the state-of-the-art literature on BM and

business model innovation (BMI). Then we will take a closer look at service BMs.
Further, we will discuss IoT BM development process as a part of the “Ignite | IoT
Strategy Execution” methodology (Slama, Puhlmann et al., 2015). Finally, we
will look at the literature discussing the challenges of getting new products and
services to market, specifically as a result of BMI or development of new BM.
2.1 Business models and business model innovation
BM and BMI are studied widely and we have by no means explored all of it in this
research. In the next sections, we will outline the areas we have focused on.
A recent study by Foss and Saebi (2017) shows that concepts of BM and BMI
over the last 15 years have gained a lot of attention both among researchers and
practitioners. Despite that attention, there is still much ambiguity with respect to
what BM and BMI are.
In different sources BM is defined as either a statement, a description, a repre-

sentation, an architecture, a conceptual tool or model, a structural template, a
method, a framework, a pattern and a set (Zott, Amit et al., 2011). Amit and Zott
(2012, p.42), for instance, define BM as a “system of interconnected and interde-
pendent activities that determines the way the company “does business” with its
customers, partners and vendors”.
177
When it comes to BMI, Foss and Saebi (2017) identified two research streams:
one research stream views BMI as a process, and another views it as an outcome.
For instance, Bucherer, Eisert et al. (2012, p.184) define BMI as “a process that
deliberately changes the core elements of a firm and its business logic”, while
Gambardella and McGahan (2010, p.263) state that BMI “occurs when a firm adopts
a novel approach to commercializing its underlying assets”.
Despite the increasing focus from researchers and practitioners on BM and BMI,
many BMI attempts fail. One of the greatest challenges is that “business models
by their very nature are designed not to change, and they become less flexible
and more resistant to change as they develop over time” (Christensen, Bartman
et al., 2016). They suggest that a BM consists of the following elements: value
proposition, resources, processes and the profit formula. They also claim that a
BM travels a journey through three stages: 1) creation; 2) sustaining innovation;
and 3) efficiency. They conclude that the only innovations that can be performed
in the existing BM naturally are “those that build on and improve the existing
BM and accelerate its progress along the journey”, and thus, in order to achieve
successful BMI, the firms have to “focus on creating new BMs, rather than changing
existing ones” (Christensen, Bartman et al., 2016).
Further, according to (Amit and Zott, 2012, p.44) BMI can occur in several ways:
1) “by adding novel activities” (content); 2) “by linking activities in novel ways”
(structure); 3) “by changing one or more parties that perform any of the activities”
(governance). They have also identified four major value drivers of BMs: 1) novelty
(“the degree of [BMI]”); 2) lock-in (“[BM] activities that create switching costs or
incentives for [BM] participants to stay”); 3) complementarities (“value-enhancing
effect of interdependencies between [BM] activities”); and 4) “efficiency” (cost
savings through the interconnections of the activity system) (Amit and Zott, 2012,
p.45).
Despite the aforementioned ambiguity regarding what BMI is, the majority of
researchers agree on that BMI is essential for any company striving for growth and
better financial performance. Some even argue that BMI can be more efficient
than product, process or technology innovation (Amit and Zott, 2012, Chesbrough,
2010).
178
2.2 Service business models/servitization
Nowadays, pushed by market conditions, competition and new customer de-

mands, many manufacturing companies are moving towards BMs based on ser-
vices (Kindström, 2010). This process is often referred to as “servitization” (Kastalli
and Van Looy, 2013). Examples of services that can be offered by “product-based”
companies include support and service contracts, monitoring and control ser-
vices, process consulting, maintenance contracts, equipment rental etc.
Kindström (2010) analyzed aspects and challenges of companies moving towards

service-based BMs. He argues that for established “product-based” companies,
servitization can be considered as an evolutionary change. The challenge the
companies moving towards service-based BMs often face is the need to develop
both products and services simultaneously. Kindström analyzed such elements
of service-based BMs as ‘value proposition’, ‘revenue mechanisms’, ‘value chain’,
‘value network’, ‘competitive strategy’ and ‘target market’, and came to conclusion
that in order to shift to service-based BM, companies must change all elements of
their BMs. In particular, he suggests companies moving towards service-based BM:
to focus on developing relationship-building competences (with regards to both
customers and suppliers); be more aware of the customer’s processes (because of
the need to interact with the customer in sales, delivery and post-delivery stages);
to design a dynamic portfolio adaptable to needs of different customer segments;
to focus on creating a service delivery infrastructure; and to focus on “developing
new revenue mechanisms based on customer operations and profitability” (which
may lead to cultural change in the organization) (Kindström, 2010, p.489).
Despite the growing popularity of servitization, its impact on the manufacturing

firms’ performance still remains an open question. Contrary to the expected
economic benefits of servitization, some studies show implementation problems
that can lead to the manufacturing businesses’ performance decline, so-called
“servitization paradox” (Kastalli and Van Looy, 2013). In order to overcome this
“servitization paradox”, Kastalli and Van Looy (2013) recommend service-oriented
manufacturing firms the following: adopt an integrated product-service BM; im-
plement practices that generate customer proximity; and consider necessary
investments in services in order to achieve long-term profitability.
179
2.3 IoT business model development
Recently, proliferation of such concepts as Internet of Things (IoT), Industrial

Internet, Industry 4.0 has gone viral. These concepts are interchangeably used
in the context of and in connection to the new wave of disruptive changes. For
manufacturing companies, the spread of IoT means first and foremost the acceler-
ation of the shift towards integrated product-service offerings (Slama, Puhlmann
et al., 2015). This transition is in line with what we have discussed in the previous
section about servitization.
To help companies define their IoT strategies and prepare for IoT adoption, as well
as to create and manage a portfolio of IoT projects, Slama, Puhlmann et al. (2015)
developed a methodology called Ignite | IoT Strategy Execution. This methodology
includes such stages as IoT opportunity identification, IoT opportunity manage-
ment and Initiation. The IoT BM development is considered as a part of the IoT
opportunity management stage. Here, Slama, Puhlmann et al. (2015) refer to the
IoT BM builder developed by Bosch Software Innovations, as the best-practice.
The IoT BM builder is based on the widely used Osterwalder’s Business Model
Canvas (Osterwalder and Pigneur, 2010), and addresses such IoT-specific aspects
as need for clear partner value proposition (since IoT solutions often depend
on partner ecosystem) and the use of data derived from connected things and
services based on top of this information (Slama, Puhlmann et al., 2015). In
particular, the IoT BM builder suggests calculating the total cost of ownership
(TCO) for the solution across all partners involved and “define the return model by
allocating the returns among the stakeholders in a fair manner”, which requires
cost transparency and trust in the IoT ecosystem (Slama, Puhlmann et al., 2015,
p.191). In addition, the IoT BM builder emphasizes the importance of documenting
nonmonetary effects of a BM, such as new market entry, accessing new technology,
coming up with new ideas and new BMs (Slama, Puhlmann et al., 2015).
This said, from practitioner’s perspective, successful transformation to new, IoT

BMs strongly depends on the company’s ability to effectively adapt its marketing
and sales strategies to their new products and services. However, the literature
addressing challenges manufacturing companies face in their sales and marketing
operations and customer relationship management as a result of BMI or deploy-
ment of new BMs, is scarce. This is especially noticeable with regards to service
and IoT BMs.
180
3 Research objectives and methodological approach
It is evident that servitization leads to considerable transformation of how man-

ufacturing companies sell their products and services. This involves the need
for the sales teams to adjust their sales strategy. Slama, Puhlmann et al. (2015)
suggest that incentive models based on upfront revenues need to be substituted
by the models that support recurring revenues. In addition, marketing teams will
need to utilize product usage data to carry out effective marketing campaigns
for different market segments. Another driver for adjusting sales and marketing
strategies of manufacturers is increasing demand for customized products, which
implies that products need to be sold before they have been produced.
Baines and W. Lightfoot (2013) completed a study exploring practices and tech-
nologies successfully servitized manufacturers use in the delivery of advanced
services. Among other practices, efficient customer relationships were identified
as one of the factors for successful delivery of advanced services. They further
point out that moving away from a “transactional approach to doing business, to
one where there are strong relationships in place throughout the life-cycle of the
service offering” can be seen as a “necessity for the service delivery rather than a
feature of the offering” (Baines and W. Lightfoot, 2013, p.21).
3 Research objectives and methodological approach
Based on the theoretical foundations built in the previous chapter, our research
aims to identify the core elements of developing an IoT BM. We have paid most of
our attention to elements such as marketing channels, customer relationships,
cost structure and revenue streams, local and overall ROI of an IoT solution, as
well as to nonmonetary effects of the new IoT BM. Since this is an explorative form
of research, a qualitative research method was chosen. Qualitative research can
be done in several ways, which include ethnography, grounded theory, narra-
tive analysis, case study analysis etc. (Guest, Namey et al., 2013). This research
conducts early theory building through empirical case study. Yin (2013) states
that a case study investigates a contemporary phenomenon in its natural setting
and the outcome is on relevant theories generated from understanding gained
through observing actual practice. We selected our case based on the opportu-
nity to study the development process of an IoT BM and its opportunities and
challenges. Data was mainly qualitative and collected through semi-structured
interviews, observations and discussions. In particular, in the period between
February and May of 2017 there were conducted four interviews with the CEO of
181
IoT BM in Offshore li�ing equipment

BM&BMI IoT BM
ETO se�ng
Original principles derived from Early theory building Problems observed in industry
theory through empirical case
study
Figure 1: Methodological approach
the case company, each of which had a duration of approximately two hours. The
first interview was dedicated to discussion about the strategic situation and main
objectives of the focal company. In the second and the third interviews the desired
properties and the technical solution of the focal company’s new product (lifting
tower “Litjkaren”), as well as the process of development and implementation of
the new BM, were discussed. The results of these interviews were used as input
in sections 4.1. and 4.2. In the last interview, the short-term costs and revenue
streams, as well as the new product’s long-term effects for the focal company
and its customers, were discussed. The data gathered during this last interview
was used in section 4.3. Secondary data such as case company’s steering board
documents, technical product documentation etc. were also gathered and used
in sections 4.1, 4.2 and 4.3. In addition, the second author is the CEO of the case
company and was part of the entire IoT BM development process. Our method-
ological approach as shown in Figure 1, is based on the theory of BM and BMI,
and in more recent years IoT BM and follows Dubois and Gadde (2002) suggestion
of systematic combining logic, where concepts and frameworks evolve during
confrontation with case context and relevant literature throughout the research
process. The arena for the empirical work in this paper is the ETO industry, more
specifically the design and production of offshore lifting equipment.
The research question for this paper is therefore as follows: How to empirically
investigate the development of an IoT BM within the specific setting of offshore
lifting equipment?
3.1 Case company
In order to allow a better understanding of the case company we will give a

detailed description in the following section. AXTech provides engineered and
182
4 Results and discussion
specialized heavy lifting equipment for the marine and offshore industry. The
company provides the equipment on an ETO basis meaning that the equipment
delivered involves project specific engineering to incorporate client, and to some
extent supplier, specific functionality and performance. Fabrication, assembly
and final product testing are managed through selected subcontractors world-
wide.
Traditionally, since start-up in 2004, the focal company has focused on a BM that
allows the company to be competitive by the four strategic standard pillars for
development: Technology, Products, Organization (structure) and Market. The
market (client base) has traditionally sent in Requests for Quote (RFQs) or Invita-
tions to Tender (ITTs) that are a variety of detailed (or not so detailed) functional
specifications. The bidding process can be quite comprehensive and involves
substantial conceptual Front End Engineering and Design (FEED) at own risk, i.e.
“no cure no pay”. The actual value proposition is historically focused on a par-
ticular portfolio of products. Services offered are also targeted towards the very
same installed base of products.
Typical products include winch systems, A-Frame/LARS, module handling. Core
in-house technologies incorporate mechanical and structural design, advanced
analysis, electric, hydraulic and automation skills in addition to particular skills
within contract management, finance and fabrication. The focal company’s port-
folio of products has developed over years to suit specific market needs related
to advanced material handling between offshore construction vessels, the sea
and the seabed. The base of knowledge (within the company) is also used to ex-
plore new market opportunities. It is worth mentioning that some key personnel
(owners, seniors) have a long personal track record for working within this type
of industry. The history of the company must be understood with this particular
background and understanding of the industry particulars.
At peak, the focal company counted some 90 employees located in Norway and
Poland. As of today, the company counts less than 60 people whereas 38 people
are located within Norway.
In this chapter, to understand the background for development of a new BM so-

lution and implementation of a new rental BM (Litjkaren) we will give a detailed
183
description of the strategic situation and main objectives of the focal company.
Further, we will present the process and results of development and implementa-
tion of the IoT, service-based BM by the case company. Finally, we will discuss
the challenges as well as new opportunities connected to the focal company’s
transition towards the new BM.
4.1 Strategic situation and main objectives of the company
During 2014, the market changed dramatically for AXTech. From constant overload
in demand, where the company struggled to get hold of sufficient engineering
capacity, the amount of realistic new project potentials was drastically reduced. In
addition, most of the typical clients of the company were suddenly struggling with
financial liquidity. The dominant North Sea energy company launched a new cost-
cutting regime to be able to cope with the changed market conditions. This regime
implied 20% cost reduction by innovation, 20% by industrialization and further
20% by efficiency. The simple outcome would be to get the same services for half
the cost (0.8x0.8x0.8=0.5). The focal company was confident that the company’s
competence and know-how was still attractive but had to be offered in a different
way. By re-identifying its core technology vital for future success, the company
launched a development program called Litjkaren, or by some called The Little
Swinger. Litjkaren will allow the users to rent the full function of subsea module
handling instead of specifying, buying and installing the equipment onto a vessel.
A prime idea was to utilize whatever knowledge gained over years related both
to the build and to the operational aspects of such equipment. The equipment
is, due to its accurate heave compensation functionality, heavily instrumented
and this allows the potential services to be further optimized. Since most of the
focal company’s clients faced cash-flow challenges, it was important that the
equipment should be ready for use within extremely short time.
Based on a completely new BM (service proposition), the following 12 key proper-

ties were identified as shown in figure 2.
To be built by the focal company’s internal, high-end standards Litjkaren should,

among other things, be cost-effective, be able to operate in harsh weather con-
ditions, have a modular structure, be transport-friendly, be quick to mobilize
(installed on a vessel), be maintenance-friendly, be rugged (robust) and have as
minimal environmental footprint as possible.
184
PERFORMANCE/ GUIDING SPLASH

COST EFFECTIVE MODULAR
WEATHER ZONE & IN AIR
FUNDAMENT TRANSPORT MOBILIZATION MAINTENANCE

ADAPTOR FRIENDLY TIME FRIENDLY
MINIMAL
RUGGED CONTROL SYSTEM SCALABLE
FOOTPRINT
Figure 2: Litjkaren’s desired properties
185
The desired properties became sub-targets for further enhancement of offered ser-
vices and to better utilize the information provided through the applied controls
system.
Services offered to include Front-End Engineering Design (FEED), vessel integra-
tion, optional support structure, installation/mobilization, operation, mainte-
nance, de-mobilization and storage. The FEED would then incorporate the focal
company’s understanding of how to optimize the equipment to any type of vessel
as a suitable working platform.
Due to the nature of offering this concept as a service, the focal company had to
encounter a variety of vessels. The focal company also had to consider vessels
without a “moon-pool” (a shaft through the bottom of a ship for lowering and
raising the equipment into or from the water), which is traditionally used for
subsea module handling and to incorporate features that allow for proper guiding
at the vessel’s side.
Remote diagnostics and operation were also to be offered as various apps or
add-ons to the controls system.
Another feature that was discussed was the focal company’s ability to enhance
the complete operation by tapping into the vessel’s existing Dynamic Position-
ing (DP) system and by that further enhance operational properties. The focal
company has previously developed advanced in-house software that provides
detailed understanding of a vessel’s property (behavior) in combination with such
specialized equipment installed. This allows the focal company to optimize the
operation towards specific needs. For instance, the vessel owner may ask “Can I
recover a 20t module in such defined sea condition?” The focal company’s system
will then optimize not only the equipment, but also the vessel (heading, draft
etc.) for the conditions given and to provide a clear answer. Also by accumulative
knowledge the focal company may also suggest for the vessel owner to optimize
the vessel for further enhancement (like roll/pitch dampening system etc.)
The actual equipment in discussion is a lifting tower designed for safe handling of
subsea modules and tools between a vessel operation in open sea (harsh weather
conditions) and the seabed (figure 3).
This incorporates heave compensated winch systems in addition to various means
of guides to secure the object from any kind of operational damage. The tower size
and capacity are defined by a careful evaluation of available common modules
used for such application.
186
Power and
controls fully
Connected via integrated
Ethernet/4G (within tower)
Cursor Guide
Winch, 12t ART/
Cameras (mul�) Sync
Main Li� Winch,
25t AHC
Li� wire guide
Guide Winch, 5t
AHC/ART
Cursor/prong
guide Energy Storage
(peak power)
Telescopic
Subsea tool/
Cursor Rail
module <25t
Guide
Figure 3: Litjkaren: technical solution
Layered architecture of digital technology consists of four layers: devices, net-

works, services and contents (Yoo, Henfridsson et al., 2010). Below we briefly
characterize each of these layers in Litjkaren’s architecture.
Device Layer: Sensors/Equipment. The equipment is fully instrumented to the

extent that all relevant parameters are fully monitored and controlled. Access to
relevant data is available both through local application servers and by remote
servers via satellite network or 4G communication links.
Network Layer: 4G/Ethernet
Service Layer: Knowledge base (people), in-house developed tools for optimized
operational properties.
Content Layer: Analysis, optimized operations data, remote operation concept

(app), preventive and corrective maintenance system.
187
4.2 Developing and Implementing the IoT BM
For the development of the product, it was necessary to get a full overview of
the BM and look for both internal improvements and for potential new business
based on the fact that the equipment now was fully connected.
Internal Improvements. A core element for the design was to implement and im-
prove on existing solutions for improved maintenance. As the plan was to rent out
the equipment, the focal company needed to secure an improved Overall Equip-
ment Effectiveness (OEE). Examples of such improvements are remote monitoring
and efficient maintenance programs.
In addition, by avoiding expensive and, to some extent, excessive client spec-
ifications the focal company was able to enhance overall quality by focusing
purely on function, reliability and connectivity. Overall target was to gain quality
improvements.
New business. Through IoT opportunity analysis there were also evaluated new
business opportunities that could be developed within this project. The project
provided new business opportunities both as a product-centric BM but also as
a new service-centric BM. The idea was to focus on the product itself (fully op-
timized) and at the same time allow this particular product to form a central
part of the services offered. Revenue would then be generated from not just
the actual rental but also the substantial amount of added services needed. Ex-
amples of such additional services are wire spooling and condition monitoring
non-destructive testing (NDT) services.
The revised BM can be described as outlined in the three phases of the Innovation
Project Canvas with Asset Integration Architecture (AIA) developed by Five I’s
Innovation Management GmbH (Slama, Puhlmann et al., 2015, p.188). The first
phase is to develop the actual value proposition with a repeating review of the
client, client needs, market trends and competition. For the focal company, this
incorporated an evaluation of competitive designs and how to secure that the
final concept could provide a competitive edge not only by functionality but also
by net investment (cost).
In the second phase, there is an evaluation of the actual solutions offered. For
the focal company, it was important to offer the full comprehensive service, like a
payment per lift/operation as this would allow also new clients with less capacity
to incorporate this function as a part of their own ambition. By doing this, the
focal company could enter a position in which the more senior clients would
188
regard the concept also as a threat to their own business, because operators of
smaller and less expensive vessels would now be able to offer advanced subsea
lifting without the need to either invest in or operate the actual equipment.
This allows the services offered to be differentiated and adjusted towards each
specific client. The concept of remote operation is something that will enhance
the value proposition significantly and this subject is currently jointly discussed
with relevant Remotely Operated Vehicle (ROV) services providers. The third step
is to move forward on the development of the services to be offered.
Marketing and sales of such a comprehensive and highly technical service pack-
age is something that requires careful planning and continued efforts over time.
The focal company’s method is built on the well-used word of trust. As most of
this type of equipment is normally presented as ideas on drawings and story-
boards, the focal company was eager to present the very real thing. Key clients
were invited either individually or in groups so that all operational properties
could be demonstrated. Another important method for telling the story is to
attend technological conferences and have papers presented. The focal company
strongly believes in building trust by personal attendance but in this case, a digital
marketing strategy was also formed to make sure the concept made known to
the public. LinkedIn, Facebook and electronic white-papers were all part of this
strategy.
4.3 Challenges and Opportunities
In Figures 4 and 5 below, we briefly analyze Litjkaren’s short-term costs and rev-
enue streams (“local ROI”), as well as the solution’s long-term effects for the focal
company and its customers (“overall IoT business case”) as suggested by Slama,
Puhlmann et al. (2015).
When launching Litjkaren, the focal company’s management consciously ac-
cepted high risk of having negative “local ROI” of the new BM in the short run.
This was due to high upfront capital expenditure (CAPEX) in both hardware and
software and operating expenditure (OPEX) connected to maintenance, repair
and operations of Litjkaren – in the situation of absence of specific customers.
Once rented out, it was planned that Litjkaren would generate both upfront rev-
enues such as payment per mobilization/integration/demobilization, as well as
recurring revenues such as daily rent, service and remote operations payments
(Figure 4).
189
CAPEX OPEX
Asset
~1 mill NOK/device
Enhancements
COST
10-20% of CAPEX/year
Service SW implementation +
Processes, Resources, SW integration (~200k
digital value NOK)
Local ROI
Upfront revenues Recurring revenues
Preparation of the vessel for
tower installation (up to 1 mill
Day rate: 30-40k NOK/day
Asset NOK)
Operations & service: 40-
Mobilization/ integration/
Enhancements demobilizaiton + insurance +
50k NOK/day (up to 3
lifting + wait time etc. (up to 1 persons, 24 h service)
REVENUE
mill NOK)
Service None
Data monetization: “sales”
of proposed improvements
Processes, Resources, to customer based on
digital value collected data
Figure 4: Local IoT solution ROI,Litjkaren (based on (Slama, Puhlmann et al.,

2015))
190
Cost savings & Differen�a�on &

Opera�onal efficiency Strategic Benefits
- Lower TCO and financial risks for the
customer increase product attractiveness
Asset and potential for future sales
Enhancements - Improved service quality
- Reduced complaints
- OEE improvement - Early problem detection
Local ROI - Reduced down�me - Reduced environmental impact because
of improved OEE and energy storage
Service integrated into the equipment
Processes, Resources, - New possibilities with use of AR in
digital value
remote operations /”intelligent
maintenance ”
Contribu�on to overall
business case
Figure 5: Overall IoT business case, Litjkaren (based on (Slama, Puhlmann et al.,
2015))
From the overall business case perspective, in the long run, the new solution was
supposed to lead to substantial strategic benefits for the focal company (Figure 5).
New rental BM would mean lower total customer’s costs of ownership (TCO) of the
solution and would decrease customer’s financial risks connected to ownership of
the pricey equipment. The new solution would also provide for early problem de-
tection and thus help increase overall equipment effectiveness (OEE). In addition,
development of Litjkaren has led to a new idea for enhancing maintainability by
implementing Augmented Reality (AR) technology. Since the tower is designed to
be maintained with a minimum amount of personnel, the focal company needs
to ensure specialist assistance and guidance on rather complicated machinery
and controls. An internal AR project is now formally initiated to explore these
opportunities as something that can add up to the value proposed. At start, the
focal company could not see the ROI for this investment, but as the technology
becomes more practical in use, the company now assumes that also clients are
more willing to pay for such services.
191
5 Closing remarks
Despite the ambiguity regarding what BMI is, the majority of researchers agree
on that BMI is essential for any company striving for growth and better financial
performance. In recent years, BMs based on services have gained a big popularity
(and many have proven successful) among many manufacturing companies. A
powerful push to the manufacturing companies’ transition towards service-based
BMs, or so-called “servitization”, was given by the rapid development of the IoT.
In response to new market conditions and in anticipation of new customer de-
mands, the focal company started developing a product based on a completely
new, service-based, IoT BM.
At the point as this research was ended, Litjkaren was not assigned for any partic-
ular work. The market situation seems to be still at a stage where the existing fleet
of equipment/vessels has covered the needs of such subsea projects. It is a fact
that most of this work has been performed with significantly larger equipment
than strictly needed. Over time, the focal company is confident that there will be
a market for smaller light-weight systems that can be rented for a fraction of the
investment price needed when a full-size tower is integrated on a vessel.
For the focal company this project has been a good opportunity to further explore
new business potentials and to enhance the applied technology. Exploring the
business of servitization has introduced new thinking on how to enhance the
value proposed.
Finally, how to promote BMI would appear to be critical business management
skills. The authors regard the measurement of effectiveness of BMI to be de-
pendent on empirical observation and more similar research is needed. As this
reseach is limited to a single case, the research team is currently exploring the
possibilities to re-apply the suggested concept of developing and implementing
the IoT BM to several companies supplying advanced offshore equipment.
Acknowledgements
The authors would like to express their appreciation to the Norwegian Research
Council as well as the case company participated in this research cooperation.
Without their contribution, this research would have never materialized.
192
References
References
Amit, R. and C. Zott (2012). “Creating value through business model innovation”. In: MIT Sloan
Management Review 53.3, p. 41.
Baines, T. and H. W. Lightfoot (2013). “Servitization of the manufacturing firm: Exploring the
operations practices and technologies that deliver advanced services”. In: International Journal
of Operations & Production Management 34.1, pp. 2–35.
Bucherer, E., U. Eisert, and O. Gassmann (2012). “Towards systematic business model innovation:
lessons from product innovation management”. In: Creativity and Innovation Management 21.2,
pp. 183–198.
Chesbrough, H. (2010). “Business model innovation: opportunities and barriers”. In: Long range
planning 43.2, pp. 354–363.
Christensen, C. M., T. Bartman, and D. Van Bever (2016). “The hard truth about business model
innovation”. In: MIT Sloan Management Review 58.1, p. 31.
Dubois, A. and L.-E. Gadde (2002). “Systematic combining: an abductive approach to case research”.
In: Journal of business research 55.7, pp. 553–560.
Foss, N. J. and T. Saebi (2017). “Fifteen Years of Research on Business Model Innovation: How Far
Have We Come, and Where Should We Go?” In: Journal of Management 43.1, pp. 200–227.
Gambardella, A. and A. M. McGahan (2010). “Business-model innovation: General purpose tech-
nologies and their implications for industry structure”. In: Long range planning 43.2, pp. 262–
271.
Guest, G., E. E. Namey, and M. L. Mitchell (2013). “Qualitative research: Defining and designing”. In:
Collecting Qualitative Data, pp. 1–40.
Kastalli, I. V. and B. Van Looy (2013). “Servitization: Disentangling the impact of service business
model innovation on manufacturing firm performance”. In: Journal of Operations Management
31.4, pp. 169–180.
Kindström, D. (2010). “Towards a service-based business model–Key aspects for future competitive
advantage”. In: European Management Journal 28.6, pp. 479–490.
Olhager, J. (2003). “Strategic positioning of the order penetration point”. In: International journal
of production economics 85.3, pp. 319–329.
Osterwalder, A. and Y. Pigneur (2010). Business model generation: a handbook for visionaries, game
changers, and challengers. John Wiley & Sons.
Rogoff, K. (2016). Web Page. https://www.weforum.org/agenda/2016/03/what-s-behind-the-drop-
in-oil-prices/.
Slama, D., F. Puhlmann, J. Morrish, and R. M. Bhatnagar (2015). Enterprise IoT: Strategies and Best
Practices for Connected Products and Services. ” O’Reilly Media, Inc.”
Yin, R. K. (2013). Case study research: Design and methods. Sage publications.
Yoo, Y., O. Henfridsson, and K. Lyytinen (2010). “Research commentary—the new organizing logic
of digital innovation: an agenda for information systems research”. In: Information systems
research 21.4, pp. 724–735.
Zott, C., R. Amit, and L. Massa (2011). “The business model: recent developments and future
research”. In: Journal of management 37.4, pp. 1019–1042.
193
Supply Chain Flexibility and SMEs
Internationalization. A Conceptual
Framework
Sylvia Mercedes Novillo Villegas1 , Hans–Dietrich Haasis1
1 – University of Bremen
This paper has the aim of analyzing the relationship of logistics capabilities and
supply chain flexibility (SCF) as part of small and medium-sized enterprises (SMEs)
internationalization strategy through a review of relevant research in these areas.
This review will constitute the foundation for developing an integrative conceptual
framework to understand the relationship. An integrative literature review in inter-
nationalization approaches, SCF and logistics capabilities from 2006 through 2016
was conducted. The analysis of an extract from this literature review is presented
to identify the key elements that contribute to the SMEs’ flexibility as a strategy
to achieve their internationalization goals. The role that logistics capabilities
play in achieving SCF as part of SMEs internationalization strategies has not been
analyzed in literature from an integrative conceptual perspective. Further, the
areas of research were conducted mostly in large firms from developed markets.
A multi-disciplinary approach is used in this research to address this gap. As such,
this paper is the first on analyzing the relationship between the areas of interest.
This constitutes the initial phase of building a theory on the relationship between
SMEs internationalization, SCF, and logistics capabilities. Further research will be
necessary to empirically test the conceptualized relationships in both developed
and emerging markets. This paper also presents managerial implications. The
relationships between the presented areas contribute to a better understanding
of the influence of logistic capabilities regarding SMEs flexibility in a supply chain
context. This will lead to improving the SMEs’ innovation and logistics manage-
ment while designing internationalization strategies.
Keywords: maritime security; illicit trade; detection architecture; discrete event

simulation
195
Supply Chain Flexibility and SMEs Internationalization.
1 Introduction
Internationalization is a high-risk strategy that any firm can undertake due to

the complexity and uncertainty of international environments. Nevertheless not
embracing this strategy connotes to the firms a major disadvantage in relation to
their competitors who have an international orientation (Leonidou, 2004). Fur-
thermore, internationalized companies have shown a better capacity to innovate
and adapt to the fluctuations of the demand (Johanson and Vahlne, 2009). In
addition, the business scenario has switched from firms competing as indepen-
dent entities to a supply chain competing scenario (Christopher, 2000). Small
and medium-sized enterprises (SMEs) cope with many limitations due to poor
capabilities (e.g. innovativeness) and lack of resources (e.g. production capacity)
though they are recognized for being flexible, which can lead to a faster response
and to have closer relationships with customers and suppliers (Singh, Garg and
Deshmukh, 2008; Ismail, Poolton and Sharifi, 2011). Singh, Garg and Deshmukh
(2008) argued that it is necessary to assume a holistic approach in order to sustain
SMEs competitiveness, as the firms must assess their resources, procedures, and
performance regarding the referents in the industry.
In a dynamic business environment, with the aim of developing a competitive ad-

vantage, companies have to synchronize their strategies, resources, capabilities,
and operations, with suppliers to respond positively to customers’ demand and
consequently together achieve higher levels of flexibility beyond individual firms
(Gligor, 2014). Hence, firms adopting supply chain flexibility (SCF) strategies will
have an edge over their competitors (Singh and Acharya, 2013). Moreover, in a
supply chain scenario, companies have to unify, integrate and synchronize their
own logistics capabilities within their network partners in order to compete with
firms outside their supply chain (Gligor, 2014). Gligor and Holcomb (2012) argued
that supply chain agility is prompted by joining together logistics capabilities at a
network level. Mentzer, Soonhong and Bobbitt (2004) grounded the statement
that an integral element of the management of the supply chain is logistics; there-
fore logistics capabilities constitute a key element for developing supply chain
broad capabilities. Further, from an empirical study, Mandal (2016) found that
logistics capabilities impact positively on SFC which in return affects positively
the performance of the supply chain.
There is an extensive body of literature regarding the influence of logistics capa-

bilities in achieving sustainable competitive advantage regarding a supply chain
context (Mentzer, Soonhong and Bobbitt, 2004; Stank, Davis and Fugate, 2005;
196
2 Methodology
Esper, Fugate and Davis-Sramek, 2007; Gligor and Holcomb, 2012; Gligor, 2014;
Mandal, 2016). With respect to the research conducted about SMEs on these
fields, Gelinas and Bigras (2004) examined the features and characteristics of
SMEs in order to recognize their impact on the integration of logistics. They found
that in some aspects, SMEs appeared ”dynamically suited to integration”. The
flexibility of the SME, the growth of the entrepreneur and its sustainability goals,
the simplified decision-making processes, the closeness of operational and or-
ganizational levels were categorized as well-suited with integrated logistics. By
contrast, the authors categorized as unfavorable the “firms’ focus on effectiveness
rather than efficiency, their tendency to underutilize information technologies,
and their short-term strategic planning” (p.276). In addition, previous research
on these three areas of interest was conducted mostly in large firms from de-
veloped markets (Gelinas and Bigras, 2004; Verdú-Jover, Lloréns-Montes and
García-Morales, 2006; Mellat-Parast and Spillan, 2014; Felzensztein et al., 2015;
Zhang et al., 2014).
In the regard of SMEs internationalization, the role of logistics capabilities in
achieving SCF as a competitive advantage has not been analyzed in the research.
This gap highlights the necessity of an integrative conceptual frame with the
aim to understand the relationship. Thus, it is necessary the development of
an integrative framework to unify the areas under study. This paper presents
the initial stage of developing an integrative conceptual framework through an
integrative literature review.
2 Methodology
An integrative literature review was conducted to establish a comprehensive

outlook to define the relationship of logistics capabilities in gaining SCF when
designing internationalization strategies for SMEs.
The first section introduced the motivation for this work, the areas of interest
(i.e. SMEs’ internationalization approaches, SCF and logistics capabilities) which
have led to the problem formulation. The review of literature related to the areas
of interest from 2006 through 2016 was conducted. Section 3 presents a quick
summary of the literature reviewed in internationalization theories (to better
understand the drivers and factors that influence SMEs internationalization); SCF
(to establish the characteristics and elements of SCF as a competitive strategy);
and logistics capabilities (to provide the elements that influence SCF and SMEs
197
internationalization). In section 4, the literature review is analyzed and interpreted

to integrate a conceptual framework. The last section includes the theoretical
conclusions and managerial implications.
3 Literature Review
3.1 SMEs Internationalization
Leonidou (2004, p. 281) defined firms’ expansion as “the firms’ ability to initi-
ate, to develop, or to sustain business operations in overseas markets”. Three
approaches of internationalization (i.e. stage approach, network approach and
entrepreneur approach) are presented to formulate an integrative framework
regarding the phenomenon of SMEs internationalization.
3.1.1 Stage Approach
To the authors of this approach, firms’ internationalization is considered as an

“evolutionary process” where firms gradually increment their involvement in
international markets going from one stage of internationalization to the next
(Cavusgil, 1984, p.196). Firms should select the optimal entry mode to foreign
markets by evaluating their risks, market constraints, and their own resources.
This approach has two main subdivisions: the innovation-related international-
ization model (I-model) (Bilkey and Tesar, 1977; Cavusgil, 1980) and the Uppsala
model (Johanson and Vahlne, 1977).
The innovation-related internationalization model (I-model) compares the pro-
cess of firms’ internationalization with the stages of adoption for a new product.
Under this approach, each internationalization stage is studied as an innovation
stage for the firm. On the other hand, Johanson and Vahlne (1977) developed an
internationalization model, known as the Uppsala model, regarding the learning
curve of a firm. This model is grounded on two main components: the market
knowledge (including operations and markets overseas), and the market commit-
ment, i.e. the number of committed resources to foreign markets.
Firm’s current activities and experience of operations in foreign markets con-
tribute to structure its market knowledge. The gained knowledge influences the
level of commitment with the foreign market according to the decisions made
198
3 Literature Review
and the subsequent activities that result from those decisions as well as the num-
ber of committed resources. Johanson and Vahlne (1977, p.23) characterized
the approach as dynamic since internationalization “is the product of a series
of incremental decisions” which lead from one level of commitment to the next
stimulating more learning.
This model also introduced the notion of ”psychic distance” defined as “the sum
of factors preventing the flow of information from and to the market” (Johan-
son and Vahlne, 1977, p.24). Culture and language dissimilarities, the degree
of development in the industry, among others factors affect the way to receive
and transfer information in foreign markets. Thus the diminution of the psychic
distance through an integrated flow of information might enhance the firm’ access
to more geographically distant targeted markets.
3.1.2 Network Approach
Johanson and Vahlne (1990) introduced a business network approach. The au-
thors stated that firm’s internationalization is affected by the business relation-
ships in the networks of foreign markets. The argument is that the existing net-
work relationships affect firm’s decision to access a specific market and the choice
of the entry mode. Further, Johanson and Vahlne (2009) analyzed the drivers and
the modes of internationalization by mobilizing internal and inter- organizational
relations. They included the “recognition of opportunities” as part of the market
knowledge, due to the importance of this capability as it prompts the internation-
alization process. This capability is developed by creating financial, market and
technological links with other network partners and increases progressively its
operations from local markets to foreign markets. The authors set the second
state variable, “network position”, due to their assumption that firm’s internation-
alization process is determined by the position and partnership of the firm within
a network. Relationships depend on particular levels of knowledge, trust, and
commitment which vary from one partner to another.
For future references in this work, the supply chain will be the network scenario
of analysis (Stevenson and Spring, 2007).
199
3.1.3 International Entrepreneurship
The international entrepreneurship approach emerged from entrepreneurship

literature when it comes to understanding entrepreneurial processes to explain
the internationalization of a firm (Freiling and Schelhowe, 2014). An interna-
tional entrepreneur is the result of the combination of a proactive, innovative,
and risk-taking behavior which leads to cross national borders by creating value
in the firm to establish business operations in foreign markets (McDougall and
Oviatt, 2000). Therefore, three key elements are identified in this approach: in-
novativeness, proactiveness and risk-taking (Oviatt and McDougall, 2005). The
proactive behavior of firms leads them to take risky measures to overcome psy-
chic distance through innovating their products, production processes, marketing
processes, sales and service support with the aim of satisfying the requirements
of a multicultural customer base. Innovativeness is firm’s capability to “promote
new and creative ideas, products, and processes designed to service the market”
(Felzensztein et al., 2015, p.149). This implies that the more diverse are the tar-
get markets, the more innovative the firm might need to be in order to develop
successful strategies for those markets. Proactiveness refers to the firm’s ability
to acquire, use, and exchange market knowledge in a way that firm is able to
commit resource in a marketplace. Internationalization demands a higher level
of resources (i.e. production, financial, human) and capabilities that might be
acquired through the market knowledge and proactive behavior. Thus, the firm is
stimulated to undertake risky decisions. They also need the capability to learn
from their competitors and their international network in order to identify oppor-
tunities, make more risky decisions and work better with customers (Freiling and
Schelhowe, 2014).
3.2 Supply Chain Flexibility (SFC)
3.2.1 Defining the Concept
The actual business environment is highly dynamic and competitive as a result

of the introduction of new technologies and the sophistication of customers’ de-
mand, who require customized product in shorter lead times. Therefore, firms
have to deal with more complex scenarios filled with uncertainties and turbu-
lences (Stevenson and Spring, 2007). Considering these circumstances, flexibility
emerges as a strategic capability to effectively adapt to dynamic environments.
200
3 Literature Review
Flexibility is the capacity to shift or adapt with slight punishment in terms of

performance, cost, time or effort (Upton, 1994). Furthermore, for achieving the
flexibility required to add value to the customers, suppliers and distribution chan-
nels as well as to the organization, the firm has to act beyond its internal domain
(Martínez Sánchez and Pérez Pérez, 2005; Kumar et al., 2006).
The study of SFC has its roots in the literature in manufacturing flexibility and arose
in the decades of 1980 and 1990 (Slack, 1983; Sethi and Sethi, 1990; Koste and
Mal-hotra, 1999). Early studies demonstrated that flexibility impacts positively on
the performance of the firm. The literature on manufacturing flexibility focuses
mainly on the physical resources and the internal performance of a firm thus is
not sufficient to the complexity and dynamics of supply chains in which firms are
embedded (Lummus, Duclos and Vokurka, 2003). Moreover, supply chain operates
as a complex system, where the performance of every part involved affects the
entire system performance. Consequently, the flexibility of each network partner
and their interrelationship result in the flexibility of the entire supply chain (Duclos,
Vokurka and Lummus, 2003).
Authors and practitioners have different perceptions about the concept of flexibil-
ity. Lee (2004) explained the flexible capability of a firm through three different
elements, i.e. adaptability, alignment, and agility. Supply chain adaptability is
the capability to set the network design to meet structural changes upstream or
downstream, adjust the strategies of the supply network, and design products
and services. Alignment is the capability to generate incentives between the
participants within the network for a better overall performance. Agility is the
capacity to act quickly to short-term fluctuations upstream or downstream and
manage external disturbances smoothly.
Stevenson and Spring (2007) stated that SCF is the ability to generate a quick
response to different disruptions in demand and supply along with modifications
in other environmental parameters e.g. capacity limits, lead-time, and exchange
rate. Kumar et al. (2006, p.305) defined SCF as “the ability of supply chain partners
to restructure their operations, align their strategies, and share the responsibility
to respond rapidly to customers’ demand at each link of the chain, to produce a
variety of products in the quantities, costs, and qualities that customers expect,
while still maintaining high performance”.
Lummus, Duclos and Vokurka (2003) identified the components of SCF (i.e. oper-
ations systems, supply network, logistics processes, information systems, and
organizational design) and also the potential features of these components that
produce flexibility in the supply chain. Further, the authors introduced a supply
201
chain model regarding the identified components which might lead to customer
satisfaction when working effectively in a coordinated and collaborative manner.
In addition, Kumar et al. (2006) introduced an integrative framework for the im-
plementation and management of SCF. The framework includes three stages, i.e.
required flexibility identification procedure; implementing and sharing responsi-
bilities; and feedback and monitoring.
Stevenson and Spring (2007) highlighted some characteristics of flexibility. They

argued that this capability is always present in some degree. Moreover, it is
also a multi-dimensional and complex concept, which is difficult to measure,
challenging to gain and hard to imitate. The authors developed a hierarchical
taxonomy of SCF based on the different components of manufacturing flexibility.
Moreover, some dimensions of flexibility influence others, e.g. supply chain design
flexibility is influenced by sourcing flexibility (Gosling, Purvis and Naim, 2010).
3.2.2 Drivers and Analysis
It is important to understand the drivers of SCF to establish the role of logistics ca-
pabilities in accomplishing SCF strategies as a sustainable competitive advantage
for SMEs internationalization. Flexibility drivers are situation or factors that fash-
ion the need for flexibility. At the same time, each driver is associated with some
kind of uncertainty, e.g. external downstream driver for SFC is the customer who
is responsible for demand uncertainty and likewise external upstream driver is the
supplier, who is responsible for sourcing uncertainty that may drive the focal firm
to maintain a pool of suppliers (Tiwari, Tiwari and Samuel, 2015). Kumar, Shankar
and Yadav (2008) established mutual relationships and interactions among the
flexibility enablers through the development of an interpretive structure model.
A supply chain, characterized by important levels of process and information
integration, collaborative relationships, and responsiveness flexibility, will be
able to respond to the different sources of uncertainty (Gligor and Holcomb, 2012).
The impact that integration generates in achieving SCF is a fundamental theme
in the research. The integrated flexible supply chain constitutes a competitive
strategy to develop a domestic and overseas leadership in a dynamic environment
of fluctuating customer demands (Kumar, Shankar and Yadav, 2008).
202
3 Literature Review
3.3 Logistics Capabilities
Logistics capabilities are used to align, combine, adapt, and reorganize functional
competences, resources and structural skills to improve the overall performance
(Gligor and Holcomb, 2012). Capabilities are the combination of dynamic rou-
tines and procedures which show how the resources are structured, applied and
synchronized with the environment (Stank, Davis and Fugate, 2005) in order ”to
achieve superior performance and sustained competitive advantage over com-
petitors” (Morash, Droge and Vickery; 1996, p.1). They also constitute complex
sets of knowledge and skills, which determine the firm’s capacity of general ability
and efficiency. Moreover, studies have demonstrated that logistics capabilities
constitute a competitive advantage to the firm. Though the concepts of supply
chain management and logistics are related, there are individual differences.
Logistics is in charge of planning, executing, and monitoring the effective and effi-
cient flow and storing of resources, information, and services associated with the
processes of the supply chain from the sourcing place to the consumption place
with the aim of fulfilling customer demands (Mentzer et al., 2001). In contrast,
supply chain management is the set of related activities and resources of each
supply chain partner which might be considered as logistics systems integrated
into a network. The success of a supply chain depends on a high degree of the ca-
pabilities of individual logistics systems, particularly for quality- and time-based
competition (Duclos, Vokurka and Lummus, 2003).
Mentzer, Soonhong and Bobbitt (2004) studied the strategic relationship of logis-
tics and its capabilities within the frame of firm theories. The authors presented
a ”unified theory of logistics”, that includes the conceptualization of logistics
capabilities which result in a competitive advantage. These logistics capabilities
are categorized into four broad groups, i.e. logistics quality and customer service
(capabilities of demand-management interface); low supply and distribution cost
(capabilities of supply-management interface); information technology and infor-
mation sharing (information-management capabilities); and interior and exterior
coordination capabilities.
Esper, Fugate and Davis-Sramek (2007) studied how firms develop and acquire
their logistics capabilities and the way they are used to gain a sustainable competi-
tive advantage. The authors highlighted that the literature referred to the logistics
capabilities mainly included integration capabilities, demand management capa-
bilities, supply management capabilities, measurement capabilities (the degree
to which firms monitor internal and external processes and the achievement of
203
strategies), and information exchange capabilities. Three fundamental aspects

are included into the classification of logistics capabilities, i.e. effectiveness, ef-
ficiency, and differentiation. Effectiveness is the result of collaborative efforts
through integration capabilities. Supply-management capabilities stimulate the
firm to undertake efficiency while demand-management capabilities enable dif-
ferentiation strategies. Information exchange is a requirement to accomplish
the expected internal and external results that have to be measured (Gligor and
Holcomb, 2012).
4 Analysis and Integration of the Concepts
It is necessary the understanding of the relationship between the drivers of flexi-

bility, logistics capabilities as a source of flexibility and SCF for developing the
present conceptual framework. It will provide the indispensable elements to
analyze the role of logistics capabilities in accomplishing SCF strategies for SMEs’
internationalization. A synthesis of the literature regarding the relationship be-
tween these three areas is presented in figure 1.
Logistics creates advantages such as the customer value enhancement, the pro-
ductivity assessment, and functional effectiveness. One of the unique attributes
of logistics is the active coordination with inside- and outside- functions of the
firm. Therefore logistics capabilities have two dimensions, i.e. internal and exter-
nal (Gligor and Holcomb, 2012). The internal dimension works directly with other
functions and competences to organize, integrate and design cross-functional
processes inside the firm (Morash, Droge and Vickery, 1996). The external di-
mension expands logistics outside the company boundaries to link suppliers
and customers. As strategy, logistics generates the capability to synchronize
and incorporate interdependent processes regarding the flow of goods, services
and related information across major functional areas both materials manage-
ment and physical distribution. Through this unique attribute from logistics, the
firm can generate a supply chain capability by integrating processes, resources,
operational and systems interfaces to maintain operational coordination while
decreasing redundancy (Mentzer, Soonhong and Bobbitt, 2004). Moreover, logis-
tics capabilities enable the firm to generate flexibility in the supply chain to cope
up with the uncertainties of nowadays business environment.
Five relational functions to manage SMEs internationalization (i.e. market man-
agement, knowledge management, network management, innovation manage-
204
Figure 1: Relationship between drivers of flexibility, logistics capabilities and SCF
ment, and resource management) are identified from the literature review on
internationalization. For the purpose of this work, a relational function for interna-
tionalization refers to the management of activities to set the internationalization
strategies of a firm and coordinate its internal functions as well as the relationships
and efforts within its business network. The firm will be able to accomplish in-
ternationalization objectives through the effective use of the available resources,
functional competences, and organizational abilities within the supply chain.
The functions of market management and knowledge management set the main
internationalization strategy that the firm will adopt through the analysis of the
body of knowledge gained from the experiences overseas and the information
flow within its supply chain. The recognition of opportunities is one of the main
capabilities related to these functions as it prompts the design, adoption or modi-
fication of internationalization strategies through the valuation of future oppor-
tunities in foreign markets (e.g. selecting the entry mode). The stage approach
identified the market knowledge as a key factor to succeed overseas (Johanson
and Vahlne, 1977). Moreover, it highlights the influence of internal and external
drivers of internationalization and stresses the need for a different exporting
strategy according to each degree of internationalization and the commitment
205
with foreign markets (Johanson and Vahlne, 2009). Thus, the development of
information exchange is essential for managing, analyzing, acquiring, storing, and
distributing strategic and tactical information in the interior as well as with the
exterior of the firm (Mentzer, Soonhong and Bobbitt, 2004). Further, a sustainable
competitive advantage is the result of organizational learning (Mentzer, Soonhong
and Bobbitt, 2004) and SCF (Singh and Acharya, 2013) among other elements.
Information exchange prompts collective learning processes about organizing
manufacturing abilities and the incorporation of technology streams (Mentzer,
Soonhong and Bobbitt, 2004). SCF, as strategy improves the ability to respond to
fluctuations in the environment. Aligning information systems enables firms to
satisfy changing information requirements within the supply chain in order to ac-
complish strategic goals and identify business opportunities (Singh and Acharya,
2013).
Through the network management, the firm sets its internationalization strategies
regarding its relationships within the supply chain. Further, this function depends
on the firm’ networking capabilities as the internationalization process depends
on the network relationships of the firm (Johanson and Vahlne, 2009). In addition,
highly specialized and transferable resources (except the most strategic ones) are
appreciated in a networking firm, but the full use of these resources will depend
on the firm’ networking capabilities. Pihkala, Varamaki and Vesalainen (1999,
p.340) defined networking capabilities as the “abilities such as communicating
skills, cooperativeness, ability to share a vision, trust, ability to act as a network
broker, customer orientation, ability to use market information, knowledge of
co-operative arrangements and market orientation”. In contrast to the case of
born global, if the firm has poor of these capabilities, the internationalization
process will be in an incremental and progressive way (Felzensztein et al., 2015).
In addition, this function is responsible for the process of constructing resources
and competences as it represents the main source of exchange, and opportuni-
ties as well as prompts innovation and creativity (Johanson and Vahlne, 2009).
The internationalization- and competitiveness’ degree of the network influences
the internationalization- and competitiveness’ degree of the firm. Consequently,
the relationships within the network stimulate the process of firm’ international-
ization (Johanson and Vahlne, 2009). Gligor (2014, p.586) suggested that firms
aspiring to integrate supply chain strategies must have supply chain orientation
and “must develop firm-level strategies consistent with their supply chain orienta-
tion and their objective of competing through agile response”. Firms that develop
flexible strategies will recognize the importance of integrating demand and supply,
along with process flexibility. Thus, flexibility in the supply chain needs alignment
206
between each supply chain partners and it is accomplished through information

sharing (Singh and Acharya, 2013). Furthermore, logistics capabilities enable
the development of differentiation strategies of products and service through
value-added activities. Mandal (2016) found that trust, commitment, power, and
reciprocity the socio-exchange attributes have a positive and direct impact on in-
tegrated logistics capabilities that are also positively related to the enhancement
of the supply chain performance. Trust and commitment are explicit elements
of the network approach (Johanson and Vahlne, 2009) which implies a learning
process embedded between internationalization, the development of logistics
capabilities to achieve SCF.
Innovation management refers to firm’s usage of resources and competences,
knowledge, and risk-taking and proactive behavior to create value and differ-
entiation through the transnational coordination of the firm business network.
The international entrepreneurship approach explains firms’ internationalization
through innovativeness, and risk-taking and proactiveness (Felzensztein et al.,
2015). Additionally, superior levels of trust, knowledge, and commitment lead to
efficient and innovative processes (Johanson and Vahlne, 2009). Creating supply
chain flexibility is one method to face the uncertainty of demand, especially in
innovative categories of products (e.g. electronic devices) or mass customized
products (Stevenson and Spring, 2007). Moreover, firms are forced to plan the
strategic use of their resources and manage their innovation regarding the cus-
tomer demand and competitors behavior (Lummus, Duclos and Vokurka, 2003).
Further, “value chain flexibility reflects the current state of embedding process
innovation into the supply chain operations and being proactive in managing
supply-demand fulfillment” (Hock Soon and Mohamed Udin, 2011, p.507). For
developing innovation, it is necessary to determine the functional responsibili-
ties across the supply chain in a rapid and effective way. High levels of logistics
capabilities within the supply chain partners might increase the development of
new products, services and processes to reach competitive differentiation (Esper,
Fugate and Davis-Sramek, 2007).
Last, resource management sets appropriate strategies to acquire, develop, adjust
and coordinate the sourcing and use of resources along firm’s business network.
Johanson & Vahlne (2009) stated that the adjustment of resources and the coordi-
nation of activities within the “outsidership” are necessary. Thus, the networks are
fundamental to enable SMEs the development of their limited resources (Pihkala,
Varamaki and Vesalainen, 1999). The acquisition and development of logistics
capabilities might influence the generation of sustainable internationalization
strategies, since they facilitate the total cost reduction in supply chain operations,
207
and enable postponement, modularization, and standardization strategies (Esper,

Fugate and Davis-Sramek, 2007). Moreover, through the development of logistics
capabilities, implementing SCF strategies will mean a competitive advantage as
the firm enhances its ability for shipping and receiving goods rapidly and effi-
ciently as sources of supply and customers fluctuate (Stevenson and Spring, 2007;
Singh and Acharya, 2013).
5 Conclusions and Future Research
From the multidisciplinary literature review, it is concluded that internationaliza-

tion is a complex process. For achieving a competitive and sustainable advantage
in foreign markets, the role of decision makers is to select and develop effec-
tive resources and competences. The market knowledge, network relationships,
value system, and an entrepreneur behavior are indispensable elements in SMEs’
internationalization. Stank, Davis and Fugate (2005, p.29) stated “creating and
sustaining competitive advantage is an important part of the strategic planning
process”. A firm’s strategy built up regarding external factor from the environment
will drive the process to develop a structural and operational organization. “Firms
that have properly aligned strategy with structure are expected to perform better
than competitors that lack the same degree of strategic fit”. As the competition
scenario shifted from single firms’ competition to supply chains’ competition,
SMEs have to implement a supply chain orientation, particularly to face inter-
national dynamic markets. This might lead to accomplish a competitive and
sustainable advantage. SMEs in a supply chain have to coordinate and incorpo-
rate logistics capabilities as individual firms with their supply chain partners to
compete with enterprises outside their supply chain. It is also necessary to men-
tion that the main research of the areas of interest is conducted in developed and
industrialized markets. There are only a few studies analyzing the impact of inter-
nationalization of SMEs from emerging markets, particularly in Latin American
countries.
This research presents the first phase in building an integrative conceptual frame-
work to integrate and describe the relationship between the three areas of in-
terest (i.e. internationalization approaches, SCF and logistics capabilities) (see
figure 2).
We identified five relational functions to manage SMEs internationalization pro-
cesses (i.e. market management, knowledge management, network manage-
208
5 Conclusions and Future Research
Figure 2: Conceptualizing SMEs internationalization, logistics capabilities and SCF

relationship
209
ment, innovation management and resource management). These functional

areas will enable the integration of information, resources, processes, operations,
and capabilities with the SMEs-supply-chain partners. In the other hand, the
development of logistics capabilities will enable the alignment of the firm with
the customer demand as well as the supply side. The result of this iteration will
contribute to the achievement of SCF strategies and as a consequence to the
development of a sustainable competitive advantage. Further hypothesis and
empirical research are needed to support the conceptualized relationships.
Additionally to the contribution to the theory regarding the role of the logistics
capabilities in achieving SCF strategies for SMEs internationalization, there are
also managerial implications. Managers and decision makers need to recognize
the necessity of developing logistics capabilities in order to create sustainable
internationalization strategies. SMEs might shift and respond more effectively to
the dynamics of international markets by increasing their interior and exterior
flexibility, alignment, and integration. Moreover, this will lead to improving SMEs
innovation and logistics management due to the positive influence of logistics
capabilities and SCF on creating value and better managing limited resources.
References
Bilkey, W. J. and G. Tesar (1977). “The Export Behavior of Smaller-Sized Wisconsin Manufacturing
Firms”. In: Journal of International Business Studies 8.1, pp. 93–98.
Cavusgil, S. T. (1980). “On the internationalization process of firms.” In: European research 8.6,
pp. 273–281.
Cavusgil, S. T. (1984). “Differences Among Exporting Firms Based on Their Degree of International-
ization”. In: Journal of Business Research 12.2, pp. 195–208.
Christopher, M. (2000). “The agile supply chain: competing in volatile markets”. In: Industrial
marketing management 29.1, pp. 37–44.
Duclos, L. K., R. J. Vokurka, and R. R. Lummus (2003). “A conceptual model of supply chain flexibil-
ity”. In: Industrial Management & Data Systems 103.6, pp. 446–456.
Esper, T. L., B. S. Fugate, and B. Davis-Sramek (2007). “Logistics Learning Capability: Sustaining the
Competitive Advantage Gained Through Logistics Leverage”. In: Journal of Business Logistics
28.2, pp. 57–82.
Felzensztein, C., L. Ciravegna, P. Robson, and J. E. Amorós (2015). “Networks, Entrepreneurial
Orientation, and Internationalization Scope: Evidence from Chilean Small and Medium Enter-
prises”. In: Journal of Small Business Management 53, pp. 145–160.
Freiling, J. and C. L. Schelhowe (2014). “The Impact of Entrepreneurial Orientation on the Perfor-
mance of Internationalization”. In: Journal of Entrepreneurship, Management and Innovation
10.4, pp. 169–199.
210
References
Gelinas, R. and Y. Bigras (2004). “The Characteristics and Features of SMEs: Favorable or Unfa-
vorable to Logistics Integration?” In: Journal of Small Business Management 42.3, pp. 263–
278.
Gligor, D. M. (2014). “The role of demand management in achieving supply chain ag ility”. In: Supply
Chain Management: An International Journal 19.5/6, pp. 577–591.
Gligor, D. M. and M. C. Holcomb (2012). “Understanding the role of logistics capabilities in achiev-
ing supply chain agility: a systematic literature review”. In: Supply Chain Management: An
International Journal 17.4, pp. 438–453.
Gosling, J., L. Purvis, and M. M. Naim (2010). “Supply chain flexibility as a determinant of supplier
selection”. In: International Journal of Production Economics 128.1, pp. 11–21.
Hock Soon, Q. and Z. Mohamed Udin (2011). “Supply chain management from the perspective of
value chain flexibility: an exploratory study”. In: Journal of Manufacturing Technology Manage-
ment 22.4, pp. 506–526.
Ismail, H. S., J. Poolton, and H. Sharifi (2011). “The role of agile strategic capabilities in achieving
resilience in manufacturing-based small companies”. In: International Journal of Production
Research 49.18, pp. 5469–5487.
Johanson, J. and J.-E. Vahlne (1977). “The Internationalization Process of a Firm - a Model of
Knowledge Foreign and Increasing Market Commitments”. In: Journal of International Business
Studies 8.23, pp. 23–32.
Johanson, J. and J.-E. Vahlne (2009). “The Uppsala internationalization process model revisited:
From liability of foreignness to liability of outsidership”. In: Journal of International Business
Studies 40.9, pp. 1411–1431.
Johanson, J. and J.-E. Vahlne (1990). “The Mechanisms of Internationalization”. In: International
Marketing Review 7.4, pp. 11–24.
Koste, L. L. and M. K. Malhotra (1999). “A theoretical framework for analyzing the dimensions of
manufacturing flexibility”. In: Journal of operations management 18, pp. 75–93.
Kumar, P., R. Shankar, and S. S. Yadav (2008). “Flexibility in global supply chain: modeling the
enablers”. In: Journal of Modelling in Management 3.3, pp. 277–297.
Kumar, V., K. A. Fantazy, U. Kumar, and T. A. Boyle (2006). “Implementation and management
framework for supply chain flexibility”. In: Journal of Enterprise Information Management 19.3,
pp. 303–319.
Lee, H. L. (2004). “The triple-A supply chain”. In: Harvard Business Review 82.10, pp. 102–120.
Leonidou, L. C. (2004). “An Analysis of the Barriers Hindering Small Business Export Development”.
In: Journal of Small Business Management 42.3, pp. 279–303.
Lummus, R. R., L. K. Duclos, and R. J. Vokurka (2003). “Supply chain flexibility: building a new
model”. In: Global Journal of Flexible Systems Management 4.4, pp. 1–13.
Mandal, S. (2016). “Towards an Integrated Logistics Capabilities Model of Supply Chain Flexibility:
a Social Exchange Perspective”. In: Romanian Economic and Business Review 11.3, pp. 44–67.
Martínez Sánchez, A. and M. Pérez Pérez (2005). “Supply chain flexibility and firm performance: A
conceptual model and empirical study in the automotive industry”. In: International Journal of
Operations & Production Management 25.7, pp. 681–700.
McDougall, P. P. and B. M. Oviatt (2000). “International entrepreneurship: The intersection of two
research paths”. In: Academy of management Journal 43.5, pp. 902–906.
Mellat-Parast, M. and J. E. Spillan (2014). “Logistics and supply chain process integration as a
source of competitive advantage”. In: The International Journal of Logistics Management 25.2,
pp. 289–314.
211
Mentzer, J. T., J. S. Keebler, N. W. Nix, C. D. Smith, and Z. G. Zacharia (2001). “Defining Supply Chain
Management”. In: Journal of Business Logistics 22.2, pp. 1–25.
Mentzer, J. T., M. Soonhong, and L. M. Bobbitt (2004). “Toward a unified theory of logistics”. In:
International Journal of Physical Distribution & Logistics Management 34.8, pp. 606–627.
Morash, E. A., C. L. M. Droge, and S. K. Vickery (1996). “Strategic logistics capabilities for competitive
advantage and firm success”. In: Journal of Business Logistics 17.1, pp. 1–22.
Oviatt, B. M. and P. P. McDougall (2005). “Defining international entrepreneurship and modeling
the speed of internationalization”. In: Entrepreneurship: Theory and Practice 29.5, pp. 537–553.
Pihkala, T., E. Varamaki, and J. Vesalainen (1999). “Virtual organization and the SMEs: a review
and model development”. In: Entrepreneurship & Regional Development 11.4, pp. 335–349.
Sethi, A. K. and S. P. Sethi (1990). “Flexibility in manufacturing: A survey”. In: International Journal
of Flexible Manufacturing Systems 2.4, pp. 289–328.
Singh, R. K., S. K. Garg, and S. Deshmukh (2008). “Strategy development by SMEs for competitive-
ness: a review”. In: Benchmarking: An International Journal 15.5, pp. 525–547.
Singh, R. K. and P. Acharya (2013). “Supply chain flexibility: A frame work of research dimensions”.
In: Global Journal of Flexible Systems Management 14.3, pp. 157–166.
Slack, N. (1983). “Flexibility as a manufacturing objective”. In: International Journal of Operations
& Production Management 3.3, pp. 4–13.
Stank, T. P., B. R. Davis, and B. S. Fugate (2005). “A Strategic Framework for Supply Chain Oriented
Logistics”. In: Journal of Business Logistics 26.2, pp. 27–46.
Stevenson, M. and M. Spring (2007). “Flexibility from a supply chain perspective: definition and
review”. In: International Journal of Operations & Production Management 27.7, pp. 685–713.
Tiwari, A. K., A. Tiwari, and C. Samuel (2015). “Supply chain flexibility: a comprehensive review”.
In: International Journal of Retail & Distribution Mangement 38.7, pp. 767–792.
Upton, D. M. (1994). “The Management of Manufacturing Flexibility”. In: California Management
Review 36.2, pp. 72–89.
Verdú-Jover, A. J., F. J. Lloréns-Montes, and V. J. García-Morales (2006). “Environment – Flexibility
Coalignment and Performance : An Analysis in Large versus Small Firms”. In: Journal of Small
Business Management 44.3, pp. 334–349.
Zhang, X., X. Ma, Y. Wang, and Y. Wang (2014). “How can emerging market small and medium-sized
enterprises maximise internationalisation benefits? The moderating effect of organisational
flexibility”. In: International Small Business Journal 32.6, pp. 667–692.
212
Part II
Advanced Manufacturing and

Industry 4.0
Data Mining and Fault Tolerance in
Warehousing
Christopher Reining1 , Omar Bousbiba1 , Svenja Jungen1 , Michael ten Hompel1
1 – TU Dortmund University
This paper surveys the significance of data mining techniques and fault tolerance
in future materials flow systems with a focus on planning and decision-making.
The fundamental connection between data mining, fault tolerance, and materials
flow is illustrated. Contemporary developments in warehousing are assessed to
formulate upcoming challenges. In particular, the transition towards distributed
systems and the increasing data volume is examined. The significance of taking
fault tolerance into account is emphasized. Ultimately, research issues are derived
by conflating the previous findings. They comprise a holistic approach towards the
integration of data science and fault tolerance techniques into future materials
flow systems. Tackling these research issues will help to proactively harmonize the
data representation to specific data mining techniques and increase the reliability
of such systems.
Keywords: Materials Flow System; Data Mining; Fault Tolerance; Survey
215
Data Mining and Fault Tolerance in Warehousing
1 Introduction
1.1 Increasing Data Volume and Data-Driven Operations
Nowadays the amount of available data in materials flow systems grows faster
than the performance of computers to process them. Due to the implementa-
tion of superior sensors, more data is available for analysis. The application of
distributed systems and current trends such as Internet of Things (IoT), Cyber-
physical systems (CPS) and Industry 4.0 reinforce this trend (Xu, He and Li, 2014;
International Controller Association, 2014; Hofmann and Rüsch, 2017). Smart
devices are expected to decide and act autonomously or in collaboration with
each other (Schuh and Stich, 2014, pp.203–213). Decentralization is seen as neces-
sary in order to increase flexibility, reduce reaction time, and adapt to unplanned
scenarios (Wilke, 2008). Accordingly, the area of decision-making broadens, in-
put factors diversify, and standardized processes are fragmented or eliminated.
Then again, there is a desire for predetermination when it comes to planning and
controlling materials flow systems. Paradoxically, further information gathering
and processing do not necessarily facilitate decision-making (Günthner and Ten
Hompel, 2010, pp.2–5). Emerging data analysis methods are expected to create
valuable information for both humans and machines to enable enhanced cooper-
ation (Klötzer and Pflaum, 2015). Superior data processing shall relieve human
workers by taking on recurring decisions autonomously.
In contrast to other scientific fields, there is little knowledge about data analyt-
ics in logistics so that further research is crucial (Rahman, Desa and Wibowo,
2011). Firms consider data and its analysis as a relevant resource to ensure their
future competitiveness (Mazzei and Noble, 2017). At the current stage, data in
warehousing systems is mainly used for anomaly detection and process control.
To achieve additional benefit from the potential offered by the available data,
processing methods for proactive optimization and prediction have to be enabled.
Data-driven operations require relevant information to be obtained from raw data
(Manyika et al., 2015). Their implementation connects human employees and
their creativity in solving problems with state-of-the-art technology so that both
can act ideally in real-time. This results in an environment which actively sup-
ports the cooperation of man and technology. The transition from warehousing
as an isolated task into digital social networks is proposed by “Social Networked
Industry” (ten Hompel, Putz and Nettsträter, 2017).
216
1 Introduction
1.2 Rising communication complexity
From the findings of the previous subsection it can be concluded that distributed
systems will be more widely used in future logistics facilities. In such systems,
tasks are spread and solved in smaller groups of autonomous cooperating com-
puting systems (Becker, Weimer and Pannek, 2015), that are also referred to as
nodes. A distributed system can be described as a collective set of nodes, which
interact with each other through message exchange. In contrast to centralized
systems, there is no so-called “master-node”, which holds the full control over all
components in the system (Tanenbaum and Steen, 2008). Moreover, the applica-
tion of a distributed system is often predetermined by the physical distribution of
its components.
Each node has to be able to cooperate with each other node, e.g. by exchanging
messages. All tasks have to be distributed and processed by each participating
node in the distributed system. If several individual tasks exhibit time causal
dependency or are limited due to physical restrictions, nodes have to be able to
coordinate their tasks with each other. Two autonomous intelligent forklift trucks
that try to take goods from a narrow alleyway are an exemplary case. If they are
unable to coordinate, a deadlock might occur. A deadlock describes a blocking
state of a system in which each participating node Ni blocks each other node Nj
(i 6= j) by holding a lock on a particular object Oi (object Oi is locked by node
Ni ) of interest. Suppose that a group of cooperating autonomous forklift trucks
tries to gather goods from a narrow alleyway. Further suppose that one of the
forklift trucks crashes. As long as the fault is not detected, all forklift trucks may
remain in a blocking state.
The coordination problem can be simplified when all fault-free nodes of a dis-
tributed system share the same global view of open tasks as well as a global view
of the progress of accomplished tasks. Depending on their malfunction, faulty
components can hinder the fault-free nodes from reaching a global view.
Due to the complexity of future distributed systems and the increasing reliability
requirement (e.g., autonomous cooperating forklift trucks working in the same
area as a human worker), it is important to make these systems resilient against
faults. Otherwise, corruption (e.g., flash corruption due to supply voltage faults)
of subcomponents will result in high costs and/or safety risks. This may cause the
distributed system to enter an unsafe state. The problem of ensuring that fault-
free nodes of a distributed system always share the same global view (reaching
an agreement or consensus) independent of any presence of faulty nodes is also
217
known as Byzantine agreement problem. The Byzantine generals’ problem was

first introduced by Lamport, Shostak and Pease (1982). With the help of agree-
ment protocols, reaching a global view among fault-free nodes becomes possible.
Byzantine agreement protocols have a long tradition. Thus, the state-of-the-art
in the design of agreement protocols compromised a huge number of solutions
(Lamport, Shostak and Pease, 1982; Jochim and Forest, 2010). Therefore, the fun-
damental problem of reaching agreement in the presence of faulty components
can be seen as solved. Its unsolvability for different fault assumptions and/or
system models has been proven by both Fischer, Lynch and Paterson (1985) and
Santoro and Widmayer (2007).
Now, the challenge has become an optimization problem of the communication
complexity (e.g., reducing the total number of communication rounds, number
of redundant nodes and message transmission per communication round) which
depends on the number of tolerated faults (Jochim and Forest, 2010; Khosravi
and Kavian, 2012; Bousbiba, 2015).
1.3 Contribution and Outline
This paper contributes scientific hypotheses and questions to be evaluated in fur-

ther projects in the field of materials flow systems, data mining and fault tolerance.
It constitutes the necessary knowledge basis and sensitizes the ensuing issues
from both an academic and an industrial point of view (BVL - Bundesvereinigung
Logistik e. V., 2017).
The rest of the paper is organized as follows: The second section contemplates
warehousing from a researcher’s perspective. The link to knowledge discovery
and the transition towards distributed systems are discussed. The third section
further illustrates the significance of fault tolerance in this context. Finally, re-
search issues are derived based on the previous findings.
218
2 Challenges in Materials Flow Systems
2.1 Warehousing Operations and Tasks
As shown in figure 1, warehousing employs three elements, namely systems,

processes and management. Thus, a feasible way to formalize operations is to
consider them as a combination of these elements. This approach implies a wide
variety of configurations. In basic terms, warehousing processes are conducted
within a system in accordance with an underlying management strategy. Typical
planning and decision tasks interact with these operations. They comprise tech-
nology selection, capacity planning of personnel and equipment and packaging
planning.
Figure 1: Elements of warehousing (Rohrhofer and Graf, 2013, pp.8–9)
219
As illustrated in figure 1, the storage space allocation of each article is defined

by the warehouse management. It influences the effort of each order picking
process, for example the travel time to corresponding storage locations. Likewise,
the layout of the order picking system has a potentially major impact as it deter-
mines each shelf’s position. More sophisticated warehousing principles cause
more complex combinations of the elements. In a two-stage order picking system
the interaction with the conveyor system would have to be taken into account.
Then again, a conveyor itself is a electromechanical component. The aforemen-
tioned context emphasizes the interdependency of planning and decision tasks
with warehousing. Other points of view, e.g. from maintenance or performance
availability, reveal further ties.
On one hand, formalization of warehousing operations is desired in order to

apply methods from other fields of science and represents a current research
issue. On the other hand, task solving requires a high degree of domain knowl-
edge. Running a warehouse is often characterized not only by interdependencies
but uncertainties and rising task-complexity (Faber, de Koster and Smidts, 2013;
Schieweck, Kern-Isberner and ten Hompel, 2016). These aspects hinder the appli-
cation of analytical approaches. The required data, target figures and restrictions
are depending on the operation’s configuration and decisions already taken in
related tasks.
Analytical approaches can be found in a wide range of technical standards and

scientific works. They are limited to specific subsystems and use cases because
warehousing is characterized by parallel and sequential operations that rigid
formulas can hardly deal with. Materials flow simulation offers an alternative to
deal with this issue. The commonly used discrete event simulation yet requires
formalized operations and input parameters to obtain useful information (Timm
and Lorig, 2015). Once the observed (sub-)system exceeds a certain level of
complexity both analytical and simulative approaches tend to be infeasible to
use in practical application (Roessler, Riemer and Mueller, 2015). They are either
based on heavily simplified assumptions or they require a disproportional effort,
e.g. specialized personnel or long preparation and running time of the simulation
models (Frank, Laroque and Uhlig, 2013).
Applicable methods of data collection rely on observation and posterior analysis.

In most cases, the conventional method of turning raw data into useful knowl-
edge is done by manual analysis and interpretation (Bohács, Gáspár and Rinkács,
2012). Due to the necessary domain knowledge and insufficient formalization,
220
the analysis and information extraction is conducted by specialists. While plan-

ning tasks usually offer a relatively long time frame, many decisions concerning
warehousing are of short-term nature. For example, newly incoming goods in a
distribution centre have to be assigned to one of several possible storage loca-
tions without a time lag. The assignment to a location is then performed based
on a multitude of input factors of which some, e.g. the daily retrieval volume
for the upcoming weeks, may not be available yet. Nevertheless, these figures
may be predictable from past data. In such cases, warehousing operations can be
expected to improve once they are data-driven.
2.2 Application of Knowledge Discovery and Data Mining
It is estimated that about 60% of data mining projects fail (Goasduff, 2015). Con-
cerning warehousing, potential issues leading to this unsatisfactory situation are
outlined.
The figure 2 below illustrates a widely recognized approach towards Knowledge
Discovery in Databases (KDD) as proposed by Fayyad, Piatetsky-Shapir and Smyth
(1996). Its nine steps are considered more closely to assess the applicability of
the KDD process in warehousing. An overview of alternative process models is
provided by Mariscal, Marbán and Fernández (2010). In literature, there is a wide
variety of coexisting terminologies for data science related terms. Still, the drawn
conclusions are valid regardless of the used terminology. A potential distinction
of related terms is provided by Mitchell-Guthrie (2014).
The first step in the process is to develop an understanding of the observed system,
gather necessary domain knowledge and identify the goal of the KDD process.
While the warehouse operators may provide the necessary domain knowledge,
goals can be derived from different sources. In warehousing there is a wide variety
of key figures to examine. The goals may be set by technical staff or stipulated by
the senior management.
Next, a target data set has to be created. For example, data about the past through-
put volume or technical properties of the stored goods and their assignment to a
storage location may be of interest. This data can be taken from the Warehouse
Management System (WMS), Enterprise Resource Planning System (ERP) or other
sources of information. Most IT-systems in use have evolved historically with-
out regards to Knowledge Discovery. Human practitioners often generate article
master data with emphasis on easy comprehensibility. Redundancy or noise
221
Figure 2: Overview of the KDD Process according to Fayyad, Pi-atetsky-Shapir and

Smyth (1996)
avoidance has usually not been the priority when implementing the information
systems.
The third step comprises data cleansing and pre-processing. This is necessary
because the gathered data will very likely have inconsistencies, errors, out of
range values, missing values and so forth. The goal of this step is to bring the
data into a state which enables data mining methods to work as intended. While
being time-consuming and labour-intensive, pre-processing facilitates the pro-
ceeding steps (Witten et al., 2017, pp.56–65). The same is true for the fourth step,
data transformation. It comprises the search for useful features to represent
the distinctive characteristics of the data. This procedure often goes along with
dimensionality reduction. Data cleansing, pre-processing and transformation
have to be conducted in accordance with the requirements of each data mining
method. Accordingly, previous steps have to be repeated iteratively once an oper-
ation or an outcome turns out to be infeasible later on. The often given short-term
nature of warehousing operations suggests that the utilization of data mining is
vitally dependent on an efficient data preparation. An overview of pre-processing
methods is provided by García, Luengo and Herrera (2016).
Data mining itself appears as the fifth step of the KDD process. A particular data
mining method has to be matched with the predefined goal of the KDD process.
The sixth step comprises the selection of a hypothesis and a data mining algorithm.
The search for useful patterns and interesting information is conducted in the
222
seventh step. It is crucial to recognize helpful data mining methods and algorithms
for each warehousing task to exploit their potential. The chosen method and
algorithm codetermine the optimal configuration of the overall KDD process.
There are two thinkable scenarios but not limited to the following. Either the
sheer amount of data is too big to be manually analysed, impeding data-driven
decisions or the necessary information for making a valid decision is unknown
beforehand.
While the first scenario is a topic of descriptive data mining, the second is a con-
ceivable scenario of predictive data mining. Both scenarios may coincide in prac-
tice.
A common way to differentiate data mining methods is given by Fayyad, Piatetsky-
Shapir and Smyth (1996):
— Regression
— Anomaly detection
— Association rule learning
— Clustering
— Classification
— Summarization
The data mining algorithm defines the model representation, the evaluation and
the search method. For example, concrete algorithms of the regression method
are, among others, the linear, polynomial and logistical regression. It is obvious
that practitioners that choose to implement data mining methods are required to
observe a multitude of possibilities. The question arises, which of the available
methods and algorithms correspond to one or more warehousing tasks.
In the eight step the mined patterns are interpreted, e.g. by visualization. Previous
steps may be repeated in further iterations of the KDD process. This is the case
when the knowledge gain and its usability turn out to be insufficient. Often occur-
ring short-term nature of warehousing operations suggests that time-consuming
iterations are disadvantageous. They are ought to be avoided by a reasonable
configuration of the KDD process in the first place. In the final step, the extracted
knowledge is validated and made available for further processes.
To conclude this section, existing approaches to integrate KDD and data mining in
warehousing are referenced. Ming-Huang Chiang, Lin and Chen (2014) presented
223
an approach to improve the efficiency of order picking systems by finding associa-

tions between orders. The associations are the basis for heuristics to minimize the
travel distance. Pang and Chan (2017) utilized data mining to find relationships
among costumer orders. The identification of relationships helps to determine
rational storage locations to minimize the picking effort. Gils et al. (2016) in-
troduced an approach to predict the workload of order picking zones based on
past data. The forecast can be used to determine the required number of em-
ployees and their assignment within the warehouse in a rational manner. Knoll,
Prüglmeier and Reinhart (2016) presented an approach to use machine learning
for predictive inbound logistics planning. It facilitates the system’s adaptation
to future scenarios. Their approach is based on the development of a logistics
information structure (ontology).
The literature review shows that so far there is little effort spent on embedding
data mining methods and algorithms into a holistic KDD process. They rather
focus on finding suitable methods and algorithms for specific problems.
2.3 Distributed Systems in Warehousing
A central unit (master-node) controls the majority of current materials flow com-
ponents. For example, a high-bay storage, a sorting system or a pick-by-light
system are self-contained units in the sense that all information gathered and all
decisions made within this system refer to a central unit. Therefore, warehousing
operations are not triggered and conducted along several units. They are coordi-
nated with help of the ERP, WMS or other central IT systems instead. To face the
increasing flexibility demand, the distribution of warehousing operations among
autonomous units is a current topic in both academic and industrial research.
The underlying idea of realisation approaches is a plug-and-play behaviour of au-
tonomous agents also referred to as ad-hoc-networking. Each agent is a modular
unit (Lieberoth-Leden, Regulin and Günthner, 2016; Seibold and Furmans, 2017).
In a distributed system, the warehousing system components are fluctuating. For
example, the layout of the order picking zones cannot be assumed static once the
facility is built. The system is repeatedly shifting its form in short terms. Agents
may enter or leave the system for a multitude of reasons such as the required
performance capacity, maintenance intervals and so forth. Furthermore, a single
benign malfunctioning agent (e.g. fail-silent or crash) does not necessarily lead
to a system crash, as it would be the case with a malfunctioning conveyor system
in a high bay storage. Exemplary realisation approaches are referenced below.
224
Karis Pro (Colling et al., 2016) is an automated guided vehicle. Decentralized

job creation without a prior teaching of transport connections between transfer
points offers flexibility towards layout changes. Due to its modular construction
and decentralized control, additional hardware like a central master computer
are not necessary. Toru Robot (Kremen, 2016) is a warehouse robot that is able to
pick items off shelves autonomously. In order to identify objects the robots create
a map using laser sensors that scan the environment. This data is shared with
other robots working in the same warehouse. The Grid Sorter (Colling, Seibold
and Furmans, 2016) is a conveyor system which provides efficient and space-
saving sorting of goods. The system’s decentralized controllers and its structure
consisting of identical rectangular conveyor modules allow flexible adaption to
changing requirements. A reservation algorithm facilitates the transportation
of goods, which are larger than a single conveyor module. The use of Cellular
Transport Vehicles (Kirks et al., 2012) offers possibilities for areas where flexibility
and changeability is required, planning reliability is not guaranteed or automation
is not desired due to the lack of flexibility. The communication and control among
the entities is realized by a decentral architecture. They are flexible in their topol-
ogy meaning the formation of the entities is changeable at any time. Research
and applications of distributed materials flow systems are neither geographically
restrained to Europe (GreyOrange, 2017) nor solely focused on packaged goods
(Serva Transport Systems, 2017). A further overview is given by Karabegović
(2015).
The literature review shows that current research on distributed materials flow
systems mainly focuses on technical issues regarding control and organization
of the agents such as deadlock prevention or safety aspects (see e.g. (Kopecki,
2015)). So far, the data collection and processing, e.g. by the agent’s environmen-
tal sensors, are subordinate to this purpose. Therefore, the data representation
among agents from different suppliers is very likely not harmonized. As each
agent is a source of data, the number and heterogeneity of data sources increases.
Consequently, manual data analysis becomes increasingly impractical. The find-
ings in this subsection further emphasize the urgency of proactively integrating
knowledge discovery methods in future materials flow systems.
225
3 Fault Tolerance
The information in future distributed materials flow systems will be gathered

from different locations and mobile systems. Using a centralized approach of
data collection will be slow and expensive because all transferred information
must be gathered by the master node. In other words, the entities and the master
node need to be fully connected. Due to the communication bottleneck, costs and
high communication overhead, it can be concluded that a distributed solution
(i.e. a distributed knowledge discovery system) will be applied.
Data mining techniques in a distributed knowledge discovery system may help
to improve the overall system by providing useful information to the group. This
information (e.g., on several blocking areas) can be used by the autonomous
robots to adapt their path planning in real-time in order to meet their constraints.
Thus, the group will not reach a temporary deadlock.
In a distributed system, a global view has to persevered even if some of the com-
ponents behave faulty. Otherwise, the extracted knowledge from the gathered
datasets may result in faulty decisions. Another important aspect is the time in
which a decision has to be made. Cooperating agents may rely on the knowledge
extracted by the distributed knowledge discovery system in real time. This implies
that the decision made by a distributed knowledge discovery system has a big
impact on the effectiveness and efficiency of the overall logistics system (Dou,
Chen and Yang, 2015).
Although the information provided by data mining can help to improve the effec-
tiveness and efficiency of the overall logistics system, fault tolerance techniques
must be applied nevertheless. The efficiency of large and/or complex distributed
system heavily depends on the correct behaviour of its components and the
decision made with the help of a distributed knowledge discovery system. For
instance, Lee and Lin (2006) have shown that data mining techniques combined
with fault tolerance techniques are able to cope with a certain degree of polluted
data (containing a certain degree of noise, which – without using fault tolerance
techniques – would lead to ambiguous conclusions).
226
4 Identification of Research Issues
4 Identification of Research Issues
In section 2.2 it has been stated that an efficient design and quick application of
the KDD process require the data mining method and algorithm to be determined
beforehand. Consequently, it is necessary to define a framework which allows the
assignment of data mining methods and algorithms on one side to warehousing
tasks and the given time horizon on the other side. An approach to this issue
is given in figure 3. While data science provides methods and algorithms, ware-
housing tasks need to be formalized and abstracted to be assigned to them. The
time horizon possibly restricts feasible methods. As illustrated in figure 3, the
long-term packaging planning (e.g., the necessary amount of pallets or packages
in a specific time frame) is a possible use case of regression. In contrast, storage
space allocation as a short-term decision could be a use case for classification. In
such a case, formalization becomes an issue. The classification of articles is not
static over time. The retrospectively identified best classification is not necessar-
ily the actually applied classification at the time the goods had been examined
for the first time. This discrepancy is usually not shown in the provided data. A
high degree of domain knowledge is necessary to retrospectively identify the best
classification of an article.
The overview is far from being complete and serves as a starting point for further
research. An identified match of a warehousing task and a data mining method
does not necessarily mean that data science provides the best solution. Analytical
or simulative approaches may provide better solutions in regards to accuracy
and/or expenditure of time. Determining the limits of these approaches may
reveal a variety of feasible use cases of data mining in warehousing for further
examination.
In distributed systems, the harmonization of data representation among the

agent becomes an issue. A standardized framework can take the requirements
of efficient data mining into account. An industrial standard to be used by all
suppliers of materials flow components may be the final goal of this approach.
The question arises in what way the necessary computing capacity for short-term
data mining applications is provided. The data could be gathered and processed
by a central unit, by each agent on its own (with sufficient computing capacity) or
cooperative by a multitude of agents (distributed). As stated in Section 3 any form
of cooperation requires a certain degree of fault tolerance, such as in a distributed
knowledge discovery system and/or cooperative multi-agent system.
227
Figure 3: Research Gap between data mining methods and warehousing tasks
With respect to fault tolerant solutions in future distributed materials flow systems,
the following research issues were identified. Designing an efficient Byzantine
agreement protocol for different network topologies (Tanenbaum and Steen,
2008) is required as the complexity of materials flow systems will grow. In general,
agreement protocols require a quadratic communication overhead to solve the
Byzantine agreement problem. By relaxing the fault assumption (Jochim and
Forest, 2010; Khosravi and Kavian, 2012) to simple faults (e.g., which cause fail
silent behaviour) it is possible to solve the problem with linear communication
overhead. However, such solutions cannot be applied in the presence of malicious
faults (e.g., benign or malicious Byzantine faults).
One way to reduce the communication overhead is to design or deploy new

efficient fault tolerant signature algorithms (Echtle, 1999). Many agreement pro-
tocols suffer from redundant message transmissions (Jochim and Forest, 2010).
To provide fault tolerance, messages need to be retransmitted in recurrent com-
munication rounds. It raises the question whether the design of a new signature
algorithm may help to reduce the number of unnecessary retransmissions while
simultaneously providing sufficient fault protection.
228
5 Conclusion
With the help of fault diagnosis algorithms and gossip-based membership pro-
tocols (Aljeri, Almulla and Boukerche, 2013) faults can be detected. Appropriate
measures will be taken faster. As the detection of faulty behaviours without hu-
man interventions will play an important role in future logistics processes, such
algorithms will help groups of cooperating nodes to remain functional in the pres-
ence of faults or breakdowns. For instance, if an intelligent forklift truck, which
has reserved an area in order to complete its task, breaks down unexpectedly,
it will block an area for a non-specific time. However, if the fault is detected by
the other intelligent forklift trucks in the group, the area can be freed and the
information about the faulty intelligent forklift truck can be further distributed to
a sink, e.g. to a service engineer as well as to the knowledge discovery system.
5 Conclusion
The datasets being processed and analysed in future industry 4.0 warehousing
applications will be tremendous. In order to extract useful knowledge from the
gathered raw data, fault tolerant distributed knowledge discovery systems need
to be applied. While there is a consensus that data science methods offer a
great potential for planning and decision-making in warehousing, there is little
knowledge about feasible use cases for data mining methods among typical
warehousing tasks. This paper reviewed the status quo of data mining and fault
tolerance. Crucial research issues have been derived and concrete approaches to
address them have been outlined. In the Innovationslabor Hybride Dienstleistung,
the Chair of Materials Handling and Warehousing at the TU Dortmund University
plans to further work on this topic.
Acknowledgements
Parts of the work on this publication have been supported by Deutsche Forschungs-
gemeinschaft (DFG) in the context of the research project ”Human Activity Recog-
nition in the Commissioning Process”.
229
References
Aljeri, N., M. Almulla, and A. Boukerche (2013). “An efficient fault detection and diagnosis protocol
for vehicular networks”. en. In: ACM Press, pp. 23–30.
Becker, T., D. Weimer, and J. Pannek (2015). “Network structures and decentralized control in
logistics: topology, interfaces, and dynamics”. In: International Journal of Advanced Logistics
4.1, pp. 1–8.
Bohács, Gáspár, and Rinkács (2012). “Adaptive simulation techniques for modeling material
flow systems”. In: 2012 IEEE 10th Jubilee International Symposium on Intelligent Systems and
Informatics, pp. 559–564.
Bousbiba, O. (2015). “Reducing the Communication Complexity of Agreement Protocols By Ap-
plying A New Signature Scheme called SIGSEAM”. In: DEPEND 2015 : The Eigth International
Conference on Dependability, pp. 29–32.
Bundesvereinigung Logistik e. V. (BVL) (2017). Digitalisierung in der Logistik - Antworten auf Fragen
aus der Unternehmenspraxis.
Colling, D., S. Ibrahimpasic, A. Trenkle, and K. Furmans (2016). “Dezentrale Auftragserzeugung
und -vergabe für FTF”. In: Logistics Journal : Proceedings 2016.10.
Colling, D., Z. Seibold, and K. Furmans (2016). “GridSorter – Dezentral gesteuertes Materialflusssys-
tem zum Transport unterschiedlich großer Ladungsträger”. In: Logistics Journal : Proceedings
2016.05.
Dou, J., C. Chen, and P. Yang (2015). “Genetic Scheduling and Reinforcement Learning in Multi-
robot Systems for Intelligent Warehouses”. en. In: Mathematical Problems in Engineering 2015,
e597956.
Echtle, K. (1999). “Avoiding Malicious Byzantine Faults by a New Signature Generation Technique”.
In: Dependable Computing ? EDCC-3. Ed. by G. Goos, J. Hartmanis, J. van Leeuwen, J. Hlavi?ka, E.
Maehle, and A. Pataricza. Vol. 1667. DOI: 10.1007/3-540-48254-7_9. Berlin, Heidelberg: Springer
Berlin Heidelberg, pp. 106–123.
Faber, de Koster, and Smidts (2013). “Organizing warehouse management”. In: International
Journal of Operations & Production Management 33.9, pp. 1230–1256.
Fayyad, U., G. Piatetsky-Shapir, and P. Smyth (1996). “From Data Mining to Knowledge Discovery
in Databases”. In: AI Magazine, pp. 37–54.
Fischer, M. J., N. A. Lynch, and M. S. Paterson (1985). “Impossibility of Distributed Consensus with
One Faulty Process”. In: J. ACM 32.2, pp. 374–382.
Frank, M., C. Laroque, and T. Uhlig (2013). “Reducing computation time in simulation-based
optimization of manufacturing systems”. In: IEEE, pp. 2710–2721.
García, S., J. Luengo, and F. Herrera (2016). “Tutorial on practical tips of the most influential data
preprocessing algorithms in data mining”. In: Knowledge-Based Systems 98, pp. 1–29.
Gils, T. v., K. Ramaekers, A. Caris, and M. Cools (2016). “The use of time series forecasting in zone
order picking systems to predict order pickers’ workload”. In: International Journal of Production
Research 0.0, pp. 1–14.
Goasduff, L. (2015). Gartner Says Business Intelligence and Analytics Leaders Must Focus on Mindsets
and Culture to Kick Start Advanced Analytics.
GreyOrange (2017). GreyOrange Butler.
Günthner, W. A. and M. ten Hompel (2010). Internet der Dinge in der Intralogistik. ger. VDI-Buch.
OCLC: 845595346. Berlin: Springer.
230
References
Hofmann, E. and M. Rüsch (2017). “Industry 4.0 and the current status as well as future prospects
on logistics”. In: Computers in Industry 89, pp. 23–34.
Hompel, M. ten, M. Putz, and A. Nettsträter (2017). Social Networked Industry. Whitepaper. Fraun-
hofer IML, p. 10.
International Controller Association (2014). “Dream Car of the ”Ideenwerkstatt” at the ICV 2014”.
In: pp. 1–59.
Jochim and Forest (2010). “An Efficient Implementation of the SM Agreement Protocol for a Time
Triggered Communication System”. In: SAE Convergence 2010.
Karabegović, I., E. Karabegović, M. Mahmić, and E. Husak (2015). “The application of service robots
for logistics in manufacturing processes”. In: Advances in Production Engineering & Management
10.4, pp. 185–194.
Khosravi, A. and Y. S. Kavian (2012). “Reaching an agreement in a distributed environment in
absence of omission faults”. In: 2012 8th International Symposium on Communication Systems,
Networks Digital Signal Processing (CSNDSP), pp. 1–4.
Kirks, T., J. Stenzel, A. Kamagaew, and M. ten Hompel (2012). “Zellulare Transportfahrzeuge für
flexible und wandelbare Intralogistiksysteme”. In: Logistics Journal 2012.01.
Klötzer, C. and A. Pflaum (2015). “Cyber-physical systems as the technical foundation for problem
solutions in manufacturing, logistics and supply chain management”. In: 2015 5th International
Conference on the Internet of Things (IOT), pp. 12–19.
Knoll, D., M. Prüglmeier, and G. Reinhart (2016). “Predicting Future Inbound Logistics Processes
Using Machine Learning”. In: Procedia CIRP. The Sixth International Conference on Changeable,
Agile, Reconfigurable and Virtual Production (CARV2016) 52, pp. 145–150.
Kopecki, L. (2015). “An Approach for Deadlock Prevention in Decentralized Controlled Pre-storage
Zones with Dynamic Behavior and Time Windows”. In: International Journal of Materials, Me-
chanics and Manufacturing 4.3, pp. 208–212.
Kremen, R. (2016). Watch this new warehouse robot pick individual items from shelves.
Lamport, L., R. Shostak, and M. Pease (1982). “The Byzantine Generals Problem”. In: ACM Trans.
Program. Lang. Syst. 4.3, pp. 382–401.
Lee, G. and Y. t. Lin (2006). “A study on Proportional Fault-tolerant Data Mining”. In: 2006 Innovations
in Information Technology, pp. 1–5.
Lieberoth-Leden, C., D. Regulin, and W. A. Günthner (2016). “Efficient Messaging through Clus-
ter Coordinators in Decentralized Controlled Material Flow Systems”. en. In: MATEC Web of
Conferences 81, p. 06005.
Manyika, J., M. Chui, P. Bisson, J. Woetzel, R. Dobbs, J. Bughin, and D. Aharon (2015). Unlocking
the potential of the Internet of Things | McKinsey & Company. Tech. rep.
Mariscal, G., Ó. Marbán, and C. Fernández (2010). “A survey of data mining and knowledge discov-
ery process models and methodologies”. In: The Knowledge Engineering Review 25.2, pp. 137–
166.
Mazzei, M. J. and D. Noble (2017). “Big data dreams: A framework for corporate strategy”. In:
Business Horizons.
Ming-Huang Chiang, D., C.-P. Lin, and M.-C. Chen (2014). “Data mining based storage assignment
heuristics for travel distance reduction”. en. In: Expert Systems 31.1, pp. 81–90.
Mitchell-Guthrie, P. (2014). Looking backwards, looking forwards: SAS, data mining, and machine
learning.
231
Pang, K.-W. and H.-L. Chan (2017). “Data mining-based algorithm for storage location assignment in
a randomised warehouse”. en. In: International Journal of Production Research 55.14, pp. 4035–
4052.
Rahman, F. A., M. I. Desa, and A. Wibowo (2011). “A review of KDD-data mining framework and its
application in logistics and transportation”. In: The 7th International Conference on Networked
Computing and Advanced Information Management, pp. 175–180.
Roessler, M. P., J. Riemer, and M. Mueller (2015). “Decision Support for Choosing an Appropriate
Simulation Method for Dynamic Material Flow Analysis”. In: Journal of Industrial and Intelligent
Information 3.4.
Rohrhofer, C. and H.-C. Graf (2013). Weissbuch der Intralogistik und Logistiktechnologie: Sys-
temüberblick, Implementierung, Optimierung. ger. OCLC: 854726145. Aachen: Shaker.
Santoro, N. and P. Widmayer (2007). “Agreement in synchronous networks with ubiquitous faults”.
In: Theoretical Computer Science. Structural Information and Communication Complexity
(SIROCCO 2005) 384.2, pp. 232–249.
Schieweck, S., G. Kern-Isberner, and M. ten Hompel (2016). “Intralogistik als Anwendungsgebiet
der Antwortmengenprogrammierung – Potenzialanalyse”. In: Logistics Journal : Proceedings
2016.05.
Schuh, G. and V. Stich (2014). Enterprise-Integration: auf dem Weg zum kollaborativen Unternehmen.
ger. VDI-Buch. OCLC: 879580815. Berlin: Springer Vieweg.
Seibold, Z. and K. Furmans (2017). “Plug&Play-Fördertechnik in der Industrie 4.0”. de. In: Handbuch
Industrie 4.0 Bd.3. Ed. by B. Vogel-Heuser, T. Bauernhansl, and M. t. Hompel. Springer Reference
Technik. DOI: 10.1007/978-3-662-53251-5_2. Springer Berlin Heidelberg, pp. 3–20.
Serva Transport Systems (2017). EVE - Effiecient Variable Ergonomic: The driverless transport system
for intralogistics.
Tanenbaum, A. S. and M. v. Steen (2008). Verteilte Systeme: Prinzipien und Paradigmen. ger. 2.,
aktualisierte Aufl. it-informatik. OCLC: 237201545. München: Pearson Studium.
Timm, I. J. and F. Lorig (2015). “Logistics 4.0 - A challenge for simulation”. In: 2015 Winter Simulation
Conference (WSC), pp. 3118–3119.
Wilke, M. (2008). “Dezentral steuern, zentral kommunizieren - ein Steuerungskonzept für wandel-
bare Materialflussysteme”. In: Logistics Journal : referierte Veröffentlichungen 2008.April.
Witten, I. H., F. Eibe, M. A. Hall, and C. J. Pal (2017). Data mining: practical machine learning tools
and techniques. Fourth Edition. Amsterdam: Elsevier.
Xu, L. D., W. He, and S. Li (2014). “Internet of Things in Industries: A Survey”. In: IEEE Transactions
on Industrial Informatics 10.4, pp. 2233–2243.
232
Influence of Patterns and
Data-Analytics on Production logistics
Benjamin Kormann1 , Susanne Altendorfer-Kaiser2
1 – Robert Bosch GmbH
2 – Montanuniversität Leoben
The flow of information is an essential part of Industry 4.0, as more reliable process
data is available and subsequent hardware changes provide for processing power
to enable large-scale data analysis. Due to the fact that most data analytics and
big data frameworks presume that Data-Mining- and Data-Analytics-Activities are
conducted in form of projects, this paper focuses on the integration of data analyt-
ics and data mining into operational processes and the resulting consequences of
the organization. Therefore a framework to implement data analytics workflows
in production logistics to improve decision-making and processes is presented.
By integrating data analytics workflows in production logistics applying the pre-
sented framework, more resources can be devoted to proactively discover and
counteract possible bottlenecks or constrictions instead of resorting to firefighting
and taskforce-activities. The methodology to derive such framework consists of
developing and implementing data analytics use-cases along the supply chain
in production logistics according to current big-data and data-analytics frame-
works in cooperation with a large automotive supplier and modifying current
frameworks and approaches to fit the company’s requirements.
Keywords: Knowledge Discovery in Databases; Data Analytics; Process model;

Supply Chain Analytics
233
Influence of Patterns and Data-Analytics on Production logistics
1 Introduction
Current initiatives and research approaches investigate in intelligent, autonomous

and decentralized subsystems that should lead to more competitive production
and logistics processes - often revered to as ”Industry 4.0” or ”Internet of Things”.
These initiatives are not only driven by new technologies and methods of tech-
nical integration, in this context the information element is an essential asset
for successful business: every business process - either service or production
processes - requires a fundamental perception of information management to
implement scientific logistics approaches to achieve an optimized supply chain
(Jamil, Soares and Pessoa, 2017). Thus, Altendorfer-Kaiser (2015) states ”There-
fore the effective and economical integration of information and decision-making
bodies is relevant”.
Data and information are currently omnipresent and this oversupply of data
and information for the business environment obtains more disadvantages than
potentials for a company. In this context Information logistics becomes a cor-
nerstone for companies: The goal of information logistics is to deliver the right
information, in the right format, at the right place at the right time for the right
people at the right price. (Uckelmann, 2012) Therefore an efficient information
supply management is essential. However, information does not exist without
data. And this is even important for the production industry as here data, infor-
mation and generated insight based on the generated information are a major
cornerstone for successful business in order to manage production processes
and the resulting supply chains. (Fosso Wamba S., Akter S., 2015; Gimenez Isasi,
Morosini Frazzon and Uriona, 2015; Militaru, Pollifroni and Ioanid, 2015; Meudt,
et al., 2016; Corte-Real, Oliveira and Ruivo, 2017; Zhang, et al., 2017b; Zhang, et
al., 2017a) Therefore this paper deals with the importance of big data in terms of
production logistics and how big-data and analytics principles can be integrated
in business processes extending existing knowledge discovery process models
and which potential benefits can be derived by applying said model to business
situations in production logistics.
The remainder is structured as follows: Section 2 provides background material

on production logistics, big data analytics, smart data analytics, supply chain
analytics, information and knowledge discovery process models. Section 3 dis-
cusses a possible extension of existing Knowledge discovery process models to
ensure a lasting integration of big data analytics into business processes in pro-
duction logistics. Section 4 discusses current application of the proposed model
234
2 Background
in use-cases in the automotive supplier industry. Section 5 provides a summary

and concluding remarks.
2 Background
2.1 Production logistics
Production logistics will be defined in this article as all operational purchasing,

in-house material flow, in-house material handling, operational distribution and
information flow processes which do need to be organized, controlled, executed
and optimized in order to supply necessary raw-materials, perform manufacturing
operations and physically distribute finished goods to customers. (Arnold, 2006;
Günther and Tempelmeier, 2014; Wang, et al., 2016; Krieger W., 2017). Therefore,
in order to facilitate further references to the mentioned process the SCOR-Model
reference framework will be used to cluster the activities into Plan-, Source-, Make-
and Deliver-Processes in Production Logistics (Bolstorff, Peter A., Rosenbaum,
Robert G., Pohula, Rolf G., 2007; Supply-Chain Council, 2008) effectively defining
production logistics as the management of in-house supply chains supplying man-
ufacturing activities. In the context of Industry 4.0 many manufacturing related
activities are being digitalized and products or manufacturing equipment are
transformed in to cyber-physical systems which constantly gather data. (Schöning
H., 2017) The resulting data can be considered Big Data, as the gathered data is
being generated in near real-time. (Schöning H., 2017) In order to cope with such
systems, pattern identification and other analytics tools related to Big Data serve
as a viable way to handle complexity (Wehberg G., 2016) and in turn by actively
handling complexity creating a basis which ensures future competitiveness and
adaptability. (Schuh, Krumm and Amann, 2013; Leveling, Edelbrock and Otto,
2014)
2.2 Data and Information
In order to talk about information in general and for logistics in particular, it

is important at this point to define the terms data and information to connect
information and logistics in a more appropriate way. One approach can be seen
in Figure 1.
235
Data: The noun data is defined as facts and statistics collected together for
reference or analysis. The term itself comes from the Latin plural of “datum”.
Today organizations generate large amounts of multi-spectral data. In view of

its discrete form, data in itself may not be very useful, so it is often referred to as
the original knowledge asset. When data is processed into a context, it becomes
information. (Bali, Wickramasinghe and Lehaney, 2009)
Information: For this paper the relevant definition of information is defined

as something that is conveyed or represented by a particular arrangement or se-
quence. The term information origins in the Latin verb “informare” (in English” to
inform”), which means ‘to give from’ or ‘to form an idea of’. Furthermore the Latin
noun “informatio” had already had the meaning of concept and idea”. (Bali, Wick-
ramasinghe and Lehaney, 2009) mentioned that information is data that has been
arranged into a meaningful pattern and thus has an identifiable shape. An exam-
ple is a report created from intelligent database queries. (Bali, Wickramasinghe
and Lehaney, 2009) also found that information and communication technology
not only increase the communication abilities with data but also accelerate the
transferring and processing of this data into information.
When talking about data and information it is also interesting to have look on
IBMs “Business Information Maturity Model”, which defines five levels of data
management and is shown in figure 2.
The focus of data at the lowest level is from an operational perspective. At the next
level, the Information is used to manage the company. The Information becomes
a strategic asset at the next level. At the next level the Information becomes a
form of special expertise. Finally, at the top level, the Information is what give the
company a competitive advantage and therefore often needed to be protected
against external actors. (Arlbjørn and Haug, 2010)
Figure 1: Interconnection between data, information and knowledge (Auer, 2008)
236
2 Background
Figure 2: Business Information Maturity Model (Arlbjørn and Haug, 2010)
2.3 Big and Smart Data Analytics
(Provost and Fawcett, 2013) define “Big Data as ‘datasets that are too large for
traditional data-processing systems and that therefore require new technologies”.
Therefore big data is mainly produced by machines and thus often represented
as machine data, too. According to (Cooper M., 2012) ‘Big data is where the
data volume, acquisition velocity, or data representation limits the ability to
perform effective analysis using traditional relational approaches or requires the
use of significant horizontal scaling for efficient processing.” Such views focus
on the data domain and are therefore insufficient to cover all topics faced in
managing Production logistics processes, as for example complexity or analyzing
the data gathered is not covered in the definitions above. In order to connect
the different characteristics of Big Data to Production logistics the mentioned
view needs to be expanded further creating an integrated view of the topic, which
includes dependencies of characteristics, business intelligence, statistics, data
and characteristics clustering (Wu, Buyya and Ramamohanarao, 2016; Lee, 2017).
In doing so the characteristics of Big Data can be clustered in three domains. (Wu,
Buyya and Ramamohanarao, 2016)
237
Data Domain
The Data Domain covers the Variety, Velocity and Volume of Big Data. (Wu, Buyya
and Ramamohanarao, 2016) Volume measures the amount of data an organiza-
tion generates and has already generated, Velocity measures the speed of data
generation and Variety measures the different types of data accessible to poten-
tial analysis. (Kaisler, et al., 2013) These characteristics cover the descriptions of
Provost and Fawcett as well as Cooper’s and Mel’s view of the subject.
Statistics Domain
The Statistics Domain covers Veracity, Validity and Variability. The characteristics
cover issues in data quality, reliability, complexity, variation, decay and data
uncertainty. (Wu, Buyya and Ramamohanarao, 2016)
Business Intelligence Domain
The Business Intelligence Domain includes the topics of Visibility, Verdict and
Value. (Wu, Buyya and Ramamohanarao, 2016) By adding such characteristics
views on decision making, value of data for the business namely extracting valu-
able information and data hindsight, insight and foresight are also covered. (Wu,
Buyya and Ramamohanarao, 2016)
Integrating all characteristics implies that they cannot be viewed independently
from each other. (Lee, 2017) Increases in volume, variety and velocity increase
complexity, variability and value but decrease veracity. (Lee, 2017) In order to
develop a process model, all aspects need to be considered accordingly as the goal
is to ensure a lasting integration of Big Data into business processes in production
logistics.
Additionally, the topic of value must be further investigated. Big Data itself does
not necessarily generate value for the company. The generated information must
be presented, aligned with current processes, kept safe, enhanced with previous
experiences and aggregated accordingly in order to enable human interactors to
decide based on results efficiently and effectively. (Reich R., Mohanty S., Litzel N.;
Kaisler, et al., 2013; Coffey L., 2014; Heuring W., 2015; Jähnichen S., 2015; Diesner
M., 2016) Big Data Analytics becomes Smart Data Analytics. Smart Data Analytics
thus requires also to consider information management aspects which deepen
238
2 Background
the information process view perspective to be able to manage the information

and the consequent distribution of such generated insights by data analysis pro-
cesses. Additionally, the connection with smart data analytics and supply chain
management has to be clarified in order to elaborate potential benefits of data
mining and data analytics in supply chains.
2.4 Supply Chain Analytics
Applying smart data in supply chains or production logistics domain results in

supply chain analytics which are defined by applying big data and data analytics
in the context of logistics and supply chains. (Wang, et al., 2016) Additionally
Supply Chain Analytics can be seen as dynamic capabilities of a company. (Chae
and Olson, 2013) These capabilities are analytical, IT-enabled, improve supply
chain performance. (Chae and Olson, 2013) They consist of a data management,
an analytical supply chain process and a supply performance management ca-
pability. (Chae and Olson, 2013) These capabilities are necessary to gain useful
information and knowledge from supply chain data. (Chae and Olson, 2013) The
use of the gathered information and knowledge results in optimized operational,
tactical and strategic decisions in the plan-, source-, make-, deliver- and return-
phases of supply chains. (Souza, 2014) The necessary analytics-capabilities can
be divided into four sections: generating information, insights, decisions and
actions (Sivarajah, et al., 2017) These sections are (Sivarajah, et al., 2017):
Descriptive analytics Descriptive analytics provide a basis to understanding what

has happened in supply chains.
Inquisitive/diagnostic analytics Diagnostic analytics generate insight into why

the results derived in the descriptive phase have occurred.
Predictive analytics Predictive analytics is based on the complete understanding

of past events generated in the descriptive and diagnostic phase and
based on these findings anticipates future events within certain statistical
boundaries.
Prescriptive analytics Prescriptive analytics provides a basis to decision-making

processes helping decision-makers gain objective and transparent views
on historical events in combination with results from the predictive phase.
Questions like ””What now ?” can be answered
239
Pre-emptive analytics This step recommends based on the decisions made pos-
sible courses of action to prevent certain events derived in the predictive
phase.
The application and implementation of supply chain analytics must be measured

to derive measures for each stage to take in order to apply supply chain analytics
most efficiently and effectively to improve business processes. (Arunachalam,
Kumar and Kawalek, 2017) The proposed methodology by Arunachalam, Kumar
and Kawalek (2017) for measuring divides the capabilities of supply chain ana-
lytics in analytics, visualization, data-driven culture, data generation and data
integration. Measuring all these capabilities the following stages are identified in
Figure 3. (Arunachalam, Kumar and Kawalek, 2017)
Companies must therefore initiate supply chain analytics with data gathering
and being able to understand past events of internal supply chains with data.
(Initiation stage) This must then be expanded to predictions inside internal supply
chains and the inclusion of external data. (Adoption stage) The routinization stage
consists of generating knowledge and value of the derived information effectively
turning data into a competitive advantage. The stages in figure 3 show that
data and analytics capabilities go hand-in-hand in order to derive information,
knowledge and value from data.
2.5 Information Management Models
Information Management Models aim to provide reference activities to define

one possible efficient and effective way of managing the resource information
in organizations. (Krcmar, 2015) There exist some well described Information
Management Models, whereas everyone has its individual focus. Three different
approaches are (Stahlknecht and Hasenkamp, 2002), Heinrich L. J., Stelzer D.
(2011) and Mertens, et al. (2012). Stahlknecht and Hasenkamp (2002) focus on
Information as the third production factor and concentrate on the technical infor-
mation acquisition and storage. Mertens, et al. (2012) also address the technical
focus. Heinrich L. J., Stelzer D. (2011), however, focus on the management aspects
of information management and aims to establish an information management
infrastructure that supports decision-making. In the context of this paper only
the model of Krcmar (2015) is analysed in detail as it covers all characteristics
mentioned with smart data. Krcmar’s model divides information management in
four sections. These are executive activities, information economics (demand,
240
2 Background
supply and consumption), software management and hardware management.

(Krcmar, 2015) These sections respectively include: (Krcmar, 2015)
Executive activities Such activities include strategic aspects of information man-

agement, IT-Governance, information processes, human resources, fi-
nancial controlling and IT-Security.
Information economics Information economics activities result in exact knowl-

edge of information demand, information supply and by consequently
combining the two deducting the necessary information use of the orga-
nization.
Software and hardware management These activities ensure that life-cycles of

applications and hardware are generated, maintained and observed. Ad-
ditionally these sections ensure that the data necessary to satisfy demand
is supplied in the most efficient and effective manner possible.
Figure 3: Supply Chain Analytics capability framework (Arunachalam, Kumar and

Kawalek, 2017)
241
The presented information model does not include a section relevant to the cre-
ation of value for the organization based on the information provided. Therefore
the current landscape on big data reference process models must also be analyzed
in depth.
2.6 Big Data and Knowledge discovery process reference models
In order to work with Big Data and master its characteristics certain process steps
need to be considered in order to create knowledge about one specific domain.
Furthermore the analytics perspective needs to be incorporated into Big Data
in order to consider that identified patterns are transformed in usable process
knowledge. Such processes can be summarized by the concept of Knowledge
Discovery in Databases. (Fayyad, U., Piatetsky-Shapiro G., Smyth G. P., 1996) The
academic world and industries alike have developed different kinds of approaches
on how and what to face when trying to gain knowledge out of data. The concepts
analyzed in the context of this paper are:
— Knowledge Discovery in Databases (KDD) Process

(Fayyad, U., Piatetsk-Shapiro G., Smyth G. P., 1996).
— Information Flow in a Data Mining Life Cycle
(Ganesh, et al., 1996; Kopanakis I., 1999).
— SEMMA
(SAS Institute Inc, 1997)
— Refined KDD paradigm
(Collier, K., Carey, B., Grusy, E., Marjaniemi, C., Sautter, D., 1998).
— Knowledge Discovery Life Cycle (KDLC) Model
(Lee and Kerschberg, 1998).
— Cross-Industry-Standard Process for Data Mining (CRISP-DM)
(CRISP-DM Consortium, 2000).
— RAMSYS
(Moyle; Blockeel and Moyle, 2002).
— Generic Data Mining Life Cycle (DMLC)
(Hofmann, 2003; Hofmann and Tierney, 2009).
242
2 Background
— Ontology Driven Knowledge Discovery Process (ODKD)

(Gottgtroy, 2007).
— Adaptive Software Development-Data Mining (ASD-DM) Process Model

(Alnoukari, Alzoabi and Hanna, 2008).
— A Data Mining & Knowledge Discovery Process Model

(Marbán Ó, Mariscal G., Segovia J., 2009).
— AgileKDD
(Nascimento G., 2012).
— Big Data Analytics Activity Reference Framework

(Yew, 2015).
— Knowledge Discovery in Data Science

(Grady, 2016).
All mentioned methodologies focus on the data mining tasks and only partially
consider how to transfer the knowledge gathered into the organization. Further-
more, all methodologies presume that all process steps are performed as projects
and not continuous improvement tasks to optimize business processes constantly.
Additionally, no explicit loops are considered to stop the cycle once it becomes
evident that with the current quality and availability of prerequisites such as data
itself for example the project needs to be modified and transformed into a change
process. This process then aims to transform the involved business processes in
such a manner that they fulfill the prerequisites and provide the basis to improve
business processes with data mining and knowledge discovery methods. Pivk,
A., Vasilecas O., Kalibatiene D., Rupnik R.(2013) propose a methodology to use
data mining to optimize business processes. Adapting processes to incorporate
gathered knowledge is being considered, yet this model does propose to perform
loops when required, certain prerequisites are not met or model deployment
uncovers possible changes necessary to business processes in order to perform
optimizations with data mining and knowledge discovery in databases.
243
3 Extension of existing approaches to data mining and

knowledge discovery in databases
In order to solve the weaknesses mentioned above, a combination of the AgileKDD

model, the Generic Data Mining Life Cycle, the Data Mining & Knowledge Discovery
Process Model by Marban (2009), the proposed implementation of datamining
to perform business process optimization (Pivk, A., Vasilecas O., Kalibatiene D.,
Rupnik R., 2013), the Knowledge Discovery in Data Science approach (Grady, 2016),
the Big Data Analytics Activity Reference Framework (Yew, 2015) and Krcmar’s
Information Management Model (Krcmar, 2015) is proposed. Combining these
models, the following reference activities must be incorporated when trying to
derive a process reference model suitable to integrate knowledge discovery results
into business processes under the aspect of performing loops when prerequisites
of stages are not met:
— Identify potentials in business processes where data mining activities

can improve performance. (Identify potentials)
— Provide a transparent and decisive process to evaluate benefits and nec-

essary resources for chosen activities ideally expressed in monetary units.
(Benefits)
— Apply traditional data mining activities which potentially can generate

necessary knowledge to realize identified performance improvements
based on a predefined selection matrix customized to production logistics
needs. (Data Mining)
— Apply chosen techniques within a standardized environment mainly con-

sisting of a big-data-platform which provides all necessary hardware, soft-
ware, a knowledge repository documenting previous activities, a model
repository documenting previous modelling approaches and access to
various data sources in a timely manner. (Deployment)
— Continuously evaluate the progress of activities related to all necessary

steps to allow modifications to previous steps once disrupting events are
uncovered during the execution of tasks. (Evaluation)
— Provide suitable organizational methods and means to preserve the gen-

erated knowledge, share it across the organization and ensure experi-
244
3 Extension of existing approaches to data mining and knowledge discovery in databases
ences made in the application of the results are considered in future

iterations of the process. (Integration)
These steps are only to be executed with an implemented and fully operational
big-data platform allowing the use of standardized hardware, software, scaling,
interfaces, data sources, knowledge repositories, model repositories, data gover-
nance protocols, privacy protection and security measures.
The ”identify potentials”-phase consists of a thorough analysis of the business

processes to be improved with a focus on the information flows and the conse-
quent tasks mainly triggered by generated information during the execution of
the processes. Such focus is necessary as the model aims to optimize information
flows and due to the focus on information all data understanding topics are also
covered. This phase can be compared to Hofmann’s phase of constantly loop-
ing through the business understanding and data preparation phase until the
understanding of both domains is sufficient to set the objectives and hypothesis
clearly to proceed with data processing. (Hofmann, 2003) Furthermore as the
objective is to optimize the process a change process must be conceived once the
desired results are generated. This aspect also incorporates the Assessment- and
Business-Process-Renovation-phase in the proposed approach of (Pivk, A., Vasile-
cas O., Kalibatiene D., Rupnik R., 2013) excluding application analysis, design,
development and design. Additionally, the first six steps of the Big Data Analytics
Activity Reference Framework by Yew (2017) are considered.
The benefits-phase determines the necessary resources to optimize the chosen

business processes with supply chain data analytics and weigh them against
potential benefits gained from the generated insight of the knowledge discovery
process. This step mainly incorporates Marban’s (2009) approach to data mining
where software engineering and project management practices are incorporated.
(Marbán Ó, Mariscal G., Segovia J., 2009) If the benefits-process is triggered by the
processes downstream the necessary changes to be made must be assessed and
weighted against the resources necessary to implement them.
The Data Mining phase consists of the core Knowledge Discovery process covering
data transformation, pre-processing, choosing suitable data mining techniques,
developing a model, apply the model and evaluate results. (CRISP-DM Consor-
tium, 2000; (Marbán, et al., 2007)
The deployment-phase consists of adequately scaling the results and the model
to suit the business process to be optimized within the limits of the platform.
245
Furthermore, the generated model and the gathered knowledge are to be trans-
ferred to the respective repositories on the platform resembling the steps in the
Hofmann (2003) approach.
The evaluation- phase covers the maintenance and monitoring of the deployed so-
lution and should verify the benefits derived in the identify-potentials-phase. The
focus should be put on the improvement of decision-making based on the gener-
ated knowledge. It can also be considered as a pilot-implementation-phase.
The last step consists of integrating the improved process into the organization
by developing a roll-out-plan which consists of activities in the Integral process
domain cited in Marbán’s model. (Marbán Ó, Mariscal G., Segovia J., 2009) These
are, for example, training of employees or create a documentation.
The loops in shown in Figure 4 allow for the mentioned weakness of transform-
ing the process into a change process to ensure the readiness of the process to
incorporate data mining as an improvement tool. This change process is then
started during the benefits phase if the potential benefit is sufficient to justify the
necessary changes in business processes. The change process can be triggered in
various other phases of the model.
The model therefore enables companies by repeatedly executing the steps to
build up their supply chain analytics capabilities and steadily implement truly
data-driven supply chains as it can be applied in the four stages of Supply Chain
analytics and because of the loops dynamically adapted to suit the needs in every
phase.
246
3 Extension of existing approaches to data mining and knowledge discovery in databases
Figure 4: Knowledge Discovery Model for Production Logistics
247
4 Application in the automotive supplier industry
The developed model was applied in cooperation with a large international auto-
motive supplier with the goal to integrate data mining and knowledge discovery
in databases within its production logistics of the electronics manufacturing do-
main.
The framework was developed while trying to integrate data analytics in business
process in the plan-, soure-, make- and delivery-domains with current methods
mentioned in Section 3. During the application of the methods the mentioned
weaknesses became evident resulting in the development of the new reference
model. The application of the model to the use-cases is still ongoing.
Currently the focus is to thoroughly define, implement and the ”Identify poten-
tials”- and ”Benefits”-phases of the model. The ”Identify potentials”-phase in
the application domain consists of a combination of three approaches to sys-
tematically determine possible applications of Supply Chain Analytics. These
are:
1. Interviews
The first and currently most used approach is to interview managers

about current issues regarding information deficits and labor-intensive
processes necessary to provide the organization with decision-relevant
information. The results are then evaluated and processed in order to de-
termine if theoretical benefits can be generated by applying descriptive
analytics, diagnostic-, predictive-, prescriptive- or pre-emptive analytics.
The general rule currently being applied is that possible use-cases only
generating descriptive insight are not being pursued as this kind of activ-
ity is mostly connected to reporting-topics which require a completely
different kind of methodological approach. If theoretical benefits are
found, the next process step can be started.
2. Process analysis
As a prerequisite possible processes suitable for improvement must be

chosen. This requires a rather superficial and fast way to analyze pro-
cesses. In the automotive context the SIPOC-approach was chosen. The
approach consists of describing the supplier, input, process, output and
customer in a very short and precise way. (Toutenburg and Knöfel, 2009).
248
4 Application in the automotive supplier industry
Based on the rough process analysis inefficiencies and optimization po-

tentials are then identified. These inefficiencies and potentials are then
evaluated by the possible benefits of the application of data analytics.
The focus is to be laid on the input, output, suppliers and customers
as the knowledge generated by the application of data analytics can ei-
ther help to improve the information content of input/output or help to
better understand the behavior of the customer/supplier. If the SIPOC-
analysis has generated promising potentials a more detailed analysis of
the process to be improved can be started.
This step requires a precise documentation of processes ideally using
the BPMN-notation, as this notation also includes data flows. These
flows provide basis for an analysis based on information management
principles and by determining deviations of the current state to the target
picture. The deviations are then further investigated in order to determine
if the application of analytics-tools provides an improvement leading
to a decrease of the determined deviation. If the reduction is deemed
sufficient enough the next process step of the model can be started.
3. Benchmarking
Benchmarking activities in this context include extensive literature re-
search and the compilation of locally implemented data-analytics solu-
tions within the manufacturing network of the company. These two views
must then be combined, generating a view on the subject containing in-
ternal and external activities. This view must then be used to determine
possible use-cases using data analytics in a workshop with operational
management.
Once possible use-cases are determined a benefit analysis is conducted in
order to determine the economic benefits. This benefit analysis contains
qualitative and quantitative aspects.
The qualitative aspects include compliance improvements, conformance
improvements, decision-making basis improvements, increase in trans-
parency, analytics maturity and risks regarding goal-achievements. The
analytics maturity is evaluated using the desired data analytics level
(descriptive-, diagnostic-, predictive-, prescriptive-, pre-emptive analyt-
ics). The further the use-case advances the higher the evaluation.
The quantitative aspects include rough estimates regarding costs of devel-
opment, implementation and maintenance. The costs are then compared
249
against possible cost savings and performance improvements. Further-

more the qualitative aspects are included into the cost-saving-calculation
with a factor.
No conclusive results can be presented at this moment in time, yet the
approach of aligning data analytics activities with the available resources
of the platform and implemented business processes has rendered great
acceptance of the proceeding within the affected parts of the organization
and first pilot-phases have shown great potential to transform production
logistics into a data-driven endeavor seamlessly integrateable in other
Industry-4.0-activities within the company. Furthermore, the operational,
tactical and strategic decisions are already partially based on the results
of the pilots which are now based on knowledge previously unused re-
sulting in focusing available resources more efficiently and effectively
on current business challenges. These changes in resource allocation
are now used to proactively uncover potential disruptions to processes
avoiding fire-fighting and task-forces activities.
5 Conclusion
The presented supply chain analytics and knowledge discovery model does by
design align with current business processes and therefore can seamlessly be
integrated into current continuous improvement activities which facilitates the
acceptance of data analytics solutions among all parties involved. No definitive
results can be presented at this time to prove the effectiveness and efficiency
of the model in the context of production logistics. Current results of the run-
ning pilots have the desired managerial implications of being able to allocate
resources to proactively discover and counteract potential process disruptions
which can be discovered by data-based decisions enabled by the application of
the presented knowledge discovery model. Information can then be used as a
competitive differentiator. Furthermore the first two steps of the model have
been systematically applied with success in order to determine possible pilot
use-cases to test and develop the model further.
The model does have its limitations as the application has been limited to in-house
supply-chain-processes in the automotive supplier industry. The implementation
of the model does proceed accordingly, yet only the first two steps in the model
have been thoroughly defined and applied to real-world applications. The other
250
References
steps are currently highly theoretical and in pilot phases. Further research has to
be performed to specify the remaining processes in the model more thoroughly
in order to make them applicable to other domains in supply chain management.
Additionally, the platform needs to be specified further.
References
Alnoukari, M., Z. Alzoabi, and S. Hanna (2008). “Applying adaptive software development (ASD)
agile modeling on predictive data mining applications: ASD-DM methodology”. In: 2008 Inter-
national Symposium on Information Technology. IEEE, pp. 1–6.
Altendorfer-Kaiser, S. (2015). “Information Logistics Means to Support a Flexible Production?”
In: Advances in Production Management Systems: Innovative Production Management Towards
Sustainable Growth. Ed. by S. Umeda, M. Nakano, H. Mizuyama, N. Hibino, D. Kiritsis, and G.
von Cieminski. Vol. 459. IFIP Advances in Information and Communication Technology. Cham:
Springer International Publishing, pp. 414–421.
Arlbjørn, J. S. and A. Haug (2010). Business process optimization. 1. edition. Århus: Academica.
Arnold, D. (2006). Intralogistik. Berlin, Heidelberg: Springer Berlin Heidelberg.
Arunachalam, D., N. Kumar, and J. P. Kawalek (2017). “Understanding big data analytics capabilities
in supply chain management: Unravelling the issues, challenges and implications for practice”.
In: Transportation Research Part E: Logistics and Transportation Review.
Auer, T. (2008). “Wissensmanagement: Reizwort oder zeitgemäße Notwendigkeit ?” In: Controller
Leitfaden 12.
Bali, R. K., N. Wickramasinghe, and B. Lehaney (2009). Knowledge management primer. Routledge
series in information systems. New York and London: Routledge.
Blockeel, H. and S. Moyle (2002). “Collaborative data mining needs centralised model evaluation”.
In: Proceedings of the ICML-2002 Workshop on Data Mining Lessons Learned, pp. 21–28.
Bolstorff, Peter A., Rosenbaum, Robert G., Pohula, Rolf G. (2007). Spitzenleistungen im Supply
Chain Management. Berlin, Heidelberg: Springer Berlin Heidelberg.
Chae, B. and D. L. Olson (2013). “Business analytics for supply chain: a dynamic-capabilities
framework”. In: International Journal of Information Technology & Decision Making 12.01, pp. 9–
26.
Coffey L. (2014). Smart Data - What is it & How it is Different From Big Data?
Collier, K., Carey, B., Grusy, E., Marjaniemi, C., Sautter, D. (1998). A perspective on Data Mining.
Ed. by Center for Data Insight. USA.
Cooper M., M. P. (2012). Tackling Big Data. Ed. by Federal Computer Security Managers’ Forum.
USA.
Corte-Real, N., T. Oliveira, and P. Ruivo (2017). “Assessing business value of Big Data Analytics in
European firms”. In: Journal of Business Research 70, pp. 379–390.
CRISP-DM Consortium (2000). CRISP-DM 1.0: Step-by-step data mining guide.
Diesner M. (2016). “Massendaten im Fertigungsumfeld”. In: visIT[Big Data] - Fraunhofer-Institut für
Optronik, Systemtechnik und Bildauswertung IOSB, pp. 6–7.
Fayyad, U., Piatetsky-Shapiro G., Smyth G. P. (1996). “From data mining to knowledge discovery in
databases”. In: AI Magazine, pp. 37–54.
251
Fosso Wamba S., Akter S. (2015). “Big data analytics for supply chain management: A literature
review and re-search agenda”. In: Enterprise and Organizational Modeling and Simulation,
pp. 61–72.
Ganesh, M., E.-H. Han, V. Kumar, S. Shekhar, and J. Srivastava (1996). Visual Data Mining: Framework
and Algorithm Development.
Gimenez Isasi, N. K., E. Morosini Frazzon, and M. Uriona (2015). “Big Data and Business Analytics
in the Supply Chain: A Review of the Literature”. In: IEEE Latin America Transactions 13.10,
pp. 3382–3391.
Gottgtroy, P. C. M. (2007). “Ontology Driven Knowledge Discovery Process: A proposal to integrate
Ontology Engineering and KDD”. In: PACIS.
Grady, N. W. (2016). “KDD meets Big Data”. In: 2016 IEEE International Conference on Big Data (Big
Data). IEEE, pp. 1603–1608.
Günther, H.-O. and H. Tempelmeier (2014). Supply Chain und Operations Management. 11., verbesserte
Auflage. Vol. / Horst Tempelmeier; Hans-Otto Günther ; [Hauptbd.] Produktion und Logistik.
Norderstedt and Norderstedt: Books on Demand.
Heinrich L. J., Stelzer D. (2011). Informationsmanagement. Oldenburg: Oldenbourg Wissenschaftsver-
lag.
Heuring W. (2015). Warum Big Data zu Smart Data werden muss!
Hofmann, M. (2003). “The development of a generic data mining life cycle (DMLC)”. Ph. D. Ireland:
School of Computing.
Hofmann, M. and B. Tierney (2009). “An enhanced data mining life cycle”. In: 2009 IEEE Symposium
on Computational Intelligence and Data Mining. IEEE, pp. 109–117.
IARIA, ed. (2012). The Seventh International Conference on Software Engineering Advances.
Jähnichen S. (2015). “Von Big Data zu Smart Data – Herausforderungen für die Wirtschaft”. In:
Newsletter Smart Data August 2015, pp. 1–2.
Jamil, G. L., A. L. Soares, and C. R. M. Pessoa (2017). Handbook of Research on Information Man-
agement for Effective Logistics and Supply Chains. IGI Global.
Kaisler, S., F. Armour, J. A. Espinosa, and W. Money (2013). “Big Data: Issues and Challenges Moving
Forward”. In: 2013 46th Hawaii International Conference on System Sciences. IEEE, pp. 995–1004.
Kopanakis I., T. B. (1999). “Visual data mining & modeling techniques”. In: University of Manchester
Institute of Science and Technology - Department of Computation.
Krcmar, H. (2015). Informationsmanagement. Berlin, Heidelberg: Springer Berlin Heidelberg.
Krieger W. (2017). Definition Produktionslogistik. Ed. by Gabler Wirtschaftslexikon.
Lee, I. (2017). “Big data: Dimensions, evolution, impacts, and challenges”. In: Business Horizons
60.3, pp. 293–303.
Lee, S. W. and L. Kerschberg (1998). “A methodology and life cycle model for data mining and
knowledge discovery in precision agriculture”. In: SMC’98 Conference Proceedings. 1998 IEEE
International Conference on Systems, Man, and Cybernetics (Cat. No.98CH36218). IEEE, pp. 2882–
2887.
Leveling, J., M. Edelbrock, and B. Otto (2014). “Big data analytics for supply chain management”.
In: 2014 IEEE International Conference on Industrial Engineering and Engineering Management.
IEEE, pp. 918–922.
Marbán Ó, Mariscal G., Segovia J. (2009). “A Data Mining & Knowledge Discovery Process Model”.
In: Data Mining and Knowledge Discovery in Real Life Applications. Ed. by K. A. Ponce J. I-Tech
Education and Publishing, Ch. 1.
252
References
Marbán, Ó., G. Mariscal, E. Menasalvas, and J. Segovia (2007). “An Engineering Approach to Data
Mining Projects”. In: Intelligent Data Engineering and Automated Learning - IDEAL 2007. Ed. by
H. Yin, P. Tino, E. Corchado, W. Byrne, and X. Yao. Vol. 4881. Lecture Notes in Computer Science.
Berlin, Heidelberg: Springer Berlin Heidelberg, pp. 578–588.
Mertens, P., F. Bodendorf, W. König, A. Picot, M. Schumann, and T. Hess (2012). Grundzüge der
Wirtschaftsinformatik. 11. Aufl. 2012. Springer-Lehrbuch. Berlin, Heidelberg: Springer Berlin
Heidelberg and Imprint: Springer.
Meudt, T., M. P. Rößler, J. Böllhoff, and J. Metternich (2016). “Wertstromanalyse 4.0”. In: ZWF
Zeitschrift für wirtschaftlichen Fabrikbetrieb 111.6, pp. 319–323.
Militaru, G., M. Pollifroni, and A. Ioanid (2015). “BIG DATA IN SUPPLY CHAIN MANAGEMENT: AN
EXPLORATORY STUDY”. In: Network Intelligence Studies 6, pp. 103–108.
Moyle, S. (n.d.). “RAMSYS - A methodology for supporting rapid remote collaborative data mining
projects”. In:
Nascimento G., O. A. (2012). “AgileKDD: An Agile Business Intelligence Process Model”. In: The
Seventh International Conference on Software Engineering Advances. Ed. by IARIA, pp. 598–604.
Pivk, A., Vasilecas O., Kalibatiene D., Rupnik R. (2013). On approach for the implementation of data
mining to business process optimisation in commercial companies. The Free Library.
Ponce J., K. A., ed. (2009). Data Mining and Knowledge Discovery in Real Life Applications. I-Tech
Education and Publishing.
Provost, F. and T. Fawcett (2013). Data science for Business: What you need to know about data
mining and data-analytic thinking. First edition. Sebastopol, Calif.: O’Reilly.
Reich R., Mohanty S., Litzel N. (n.d.). Nach Big Data kommt Smart Data.
SAS Institute Inc (n.d.). SEMMA: Data mining and the case for sampling. SAS Institute Inc.
Schöning H., D. M. (2017). “Big Smart Data – Intelligent Operations, Analysis und Process Align-
ment”. In: Handbuch Industrie 4.0 Bd.2. Ed. by B. Vogel-Heuser, T. Bauernhansl, and M. ten
Hompel. Springer Reference Technik. Berlin, Heidelberg: Springer Berlin Heidelberg, pp. 457–
469.
Schuh, G., S. Krumm, and W. Amann (2013). Chefsache Komplexität. Wiesbaden: Springer Fachme-
dien Wiesbaden.
Sivarajah, U., M. M. Kamal, Z. Irani, and V. Weerakkody (2017). “Critical analysis of Big Data chal-
lenges and analytical methods”. In: Journal of Business Research 70, pp. 263–286.
Souza, G. C. (2014). “Supply chain analytics”. In: Business Horizons 57.5, pp. 595–605.
Stahlknecht, P. and U. Hasenkamp (2002). Einführung in die Wirtschaftsinformatik. Berlin, Heidel-
berg: Springer Berlin Heidelberg.
Supply-Chain Council (2008). Supply-Chain Operations Reference-Model Version 8.0. Pittsburg, USA.
Toutenburg, H. and P. Knöfel (2009). Six Sigma: Methoden und Statistik für die Praxis. 2., verbesserte
und erweiterte Aufl. Berlin and Heidelberg: Springer.
Uckelmann, D. (2012). Quantifying the Value of RFID and the EPCglobal Architecture Framework in
Logistics. Berlin, Heidelberg: Springer Berlin Heidelberg.
Umeda, S., M. Nakano, H. Mizuyama, N. Hibino, D. Kiritsis, and G. von Cieminski, eds. (2015).
Advances in Production Management Systems: Innovative Production Management Towards
Sustainable Growth. IFIP Advances in Information and Communication Technology. Cham:
Springer International Publishing.
Vogel-Heuser, B., T. Bauernhansl, and M. ten Hompel, eds. (2016). Handbuch Industrie 4.0 Bd.3:
Logistik. Wiesbaden: Springer Science and Business Media and Springer Vieweg.
253
Automatisierung. Springer Reference Technik. Berlin, Heidelberg: Springer Berlin Heidelberg.
Wang, G., A. Gunasekaran, E. W. Ngai, and T. Papadopoulos (2016). “Big data analytics in logis-
tics and supply chain management: Certain investigations for research and applications”. In:
International Journal of Production Economics 176, pp. 98–110.
Wehberg G. (2016). “Big Data – Mustererkennung als Erfolgsfaktor der Logistik 4.0”. In: Handbuch
Industrie 4.0 Bd.3. Ed. by B. Vogel-Heuser, T. Bauernhansl, and M. ten Hompel. Wiesbaden:
Springer Science and Business Media and Springer Vieweg, pp. 377–392.
Wu, C., R. Buyya, and K. Ramamohanarao (2016). “Big Data Analytics = Machine Learning + Cloud
Computing”. In: CoRR abs/1601.03115.
Yew, S. L. B. (2015). “Building use Cases With Activity Reference Framework for Big Data Analytics”.
In: 2015 9th International Conference on IT in Asia (CITA). IEEE, pp. 1–7.
Zhang, Y., S. Ren, Y. Liu, T. Sakao, and D. Huisingh (2017). “A framework for Big Data driven product
lifecycle management”. In: Journal of Cleaner Production 159, pp. 229–240.
Zhang, Y., S. Ren, Y. Liu, and S. Si (2017). “A big data analytics architecture for cleaner manufactur-
ing and maintenance processes of complex products”. In: Journal of Cleaner Production 142,
pp. 626–641.
254
Use Case of Self-Organizing Adaptive
Supply Chain
Julia Wagner1 , Henning Kontny 1
1 – HAW Hamburg
The main drivers for the development and implementation of self-organizing adap-
tive operation processes are the increase of flexibility while lowering costs in times
of growing consumer demands, shaping the supply chain into ”the source of en-
terprises core competence”. This paper is based on the ”Guidelines for conducting
and reporting case study research in software engineering” and it demonstrates
the design of the self-organizing adaptive supply chain through an integration
of processes and systems. As a result, the automated near real-time data flow
enables shorter order-lead-time with a high accuracy of information. Although
the study includes the architecture for the self-organizing adaptive supply chain,
it cannot be completely standardized due to individual processes and IT-systems
within different companies. Academics and practitioners may find it useful to
identify appropriate scenarios while looking for the ways to digitize their supply
chains.
Keywords: Self-Organizing Systems; Adaptive Supply Chain; Digitalization; Use

Case
255
Use Case of Self-Organizing Adaptive Supply Chain
1 Introduction
Today’s consumers require the products they have ordered to be delivered faster
and faster, and innovative products should be offered within shorter and shorter
time-to-market cycles. What industry is able to achieve today, will soon be in-
sufficient (Bauernhansl, ten Hompel and Vogel-Heuser, 2014). This expands the
role of Supply Chains (SC) to the “source of competitive advantages”(Ponte et
al., 2017), considering that they include all the material and information flows as
well as processes from sourcing (purchasing), internal activities (manufacturing,
assembly, storage etc.) to product shipping. In order to stay competitive and to
be able to satisfy growing customer demands, companies are striving toward the
application of new IT systems. There are some industries which even completely
depend on the IT in supply chains such as the meat industry, where transparency
through the shared data between different suppliers and buyers could help to
prevent scandals involving meat products (Kassahun, Hartog and Tekinerdogan,
2016). In recent studies the supply chain is viewed as one of the most important
criteria for commercial success (Wellbrock, 2015; Yuvaraj and Sangeetha, 2016),
which also should be adaptable to changing demands in the interest of potential
growth (Bogaschewsky et al., 2016; Fuller et al., 2013) and as a consequence is
even called „moving authority“ (Plattform Industrie 4.0, 2016).
Nevertheless, small and medium-sized enterprises (SMEs) have rather immature

IT-systems along supply chains, despite the fact that they generate around 50% of
GDP of developed countries and are responsible for 60% of employment (Kumari
et al., 2015). Although SMEs usually have an ERP and production systems, many
decisions such as production planning, purchasing volumes for raw materials
etc. are usually taken with help of numerous Excel sheets or Access Databases
also known as Shadow IT (Rentrop and Zimmermann, 2012; Jones et al., 2004).
Other scientists (O’Leary, 2008) underline that around 90% of supply chains are
fragmented and not digitalized end-to-end. As a result, the entire process of order
fulfillment remains very slow and is vulnerable to mistakes.
At the same time, the vision of the Internet of Things (IoT) or also known as
Industry 4.0 is arising, targeting the interconnection of physical objects such as
product or a machine with the enterprise systems, allowing companies to collect
real-time data as well as to automate simple decisions based on historical data.
“The IoT allows things’ and ‘objects’, such as RFID, sensors, actuators, mobile
phones, which, through unique addressing schemes, (…) interact with each other
and cooperate with their neighboring ‘smart’ components, to reach common
256
2 Theoretical Background
goals”(Dittrich et al., 2008). This leads to a decentralized execution of business

processes, thereby increasing process efficiency and transparency.
Research studies underline the need to provide a reference architecture integrat-

ing IoT services (Dar et al., 2015) in order to enable the spread of new technologies.
Other authors(Gunasekaran et al., 2016; Duan and Xiong, 2015) add that visibil-
ity as one of the most important organizational capabilities can “revolutionize
existing supply chains”. However, as the McKinsey Survey of over 300 industrial
companies from the USA, Japan and Germany shows, most enterprises have
difficulties in defining a convertible strategy for the implementation of the Indus-
try 4.0 to gain competitive advantages, despite their high interest in doing so.
“Digitalization is important, but we are not prepared enough” (McKinsey Digital,
2016).
For these reasons, the main objective of this paper is to describe a process of the
development of a Self-Organizing Adaptive Supply Chain based on the case study
for a greater understanding of such a complex project.
2.1 Self-Organizing Adaptive Logistics
Although adaptable processes which allow balanced execution of operations

under unstable conditions are in focus of many research articles, the design of
such remains little known (Bogaschewsky and Müller, 2016).
The concepts of Self-Organization/Autonomy are used in this case study in the

context of an automated data analysis, with an independent reaction to changes,
including self-configuration and self-optimization by systems, up to the level of an
entire self-organization. This is an indispensable component of the Industry 4.0 or
Internet of Things. Examples for self-organizing are the generation of automated
procurement orders, control of supplies or consumption of materials etc. (Agiplan
GmbH, Fraunhofer IML and Zenit GmbH, 2015). Self-Organizing Adaptive Logistics:
“focuses on the entire inter- and intra-logistics structure” in order to increase
flexibility and reaction time using efficient data analysis to make better decisions
or counteract possible disturbances of processes. It is closely linked to the logic of
Order-Controlled Production(Plattform Industrie 4.0, 2016) although it encloses
257
the integration of all operational processes and material movements (End-to-End

Supply Chain).
Autonomous logistic components which are capable to communicate among
each other will become “Smart Objects”, which are intelligently connected with
each other and are able to independently optimize themselves as an adaptive,
co-operating group, whose further development will lead to a decentralized self-
organized logistics system with complete fusion between the physical and the
virtual world (Voß, 2015).Although adaptable processes which allow balanced
execution of operations under unstable conditions are in focus of many research
articles, the design of such remains little known(Kersten et al., 2015). On the other
hand, IBM, Siemens, Cisco and other firms (O’Leary, 2008) underline the need for
autonomic systems, arguing that human intervention requires higher costs and
longer periods of time.
2.2 Related Work
The estimation of the efficiency of decision-making by human participants in

comparison to IT systems has occupied scientists for a long time. Measurable
results for such comparison were provided at the XXI International Conference
on Industrial Engineering and Operations Management 2015, where scholars
(Costas et al., 2017) presented results of an IT system (Agent-Based Modelling and
Simulation) playing the well-known Beer Distribution Game. BDG was presented
by Sterman (1989) at the MIT Sloan School of Management and is widely used to
observe the Bullwhip Effect known from Forrester (1968), which represents one of
the basic concepts of Supply Chain Management. Human players were not able
to deal with a high-load of information and uncertainties and as a result, they
were outperformed by the IT system in every single parameter, which managed to
keep a low inventory level ($3,641 IT vs. $21,662 Human decision) in combination
with low lost sales ($4 vs. $652).
258
3 Methods and Approach
3.1 Design of Use Case
The use case was designed according to “Guidelines for conducting and reporting
case study research in software engineering”(Runeson and Höst, 2009), which
underlines the importance of case studies for investigation of complex issues,
especially involving humans in interaction with technology. Case study methodol-
ogy is closely related to action research (Dittrich et al., 2008; Gorschek et al., 2006),
which is focused on change processes such as software process improvement
or technology transfer studies. Whereas the analysis of the effects of a change
is classified as a case study (Runeson and Höst, 2009). The case study allows to
“investigate a contemporary phenomenon within its real-life context”(Shahin,
2015) and represents a comprehensive analysis of several of cases (Waterman,
2014). This paper represents the cumulative results from several projects:
1. Coffee manufacturer, over 2.000 employees, project ”End-to-End digitiza-
tion of Supply Chain”.
2. LED (light-emitting diode) producing company, over 2.000 employees,
focus on digitization of Demand and Supply Chain processes.
3. Manufacturer of air filters and signaling technologies, over 500 employ-
ees, project ”Cyber Assembly” focusing on the digitization of assembly,
production, warehousing and purchasing.
The above-mentioned projects present the following similarities:
— Focus on the digitization of decision processes and automated data flow.
— Aiming at fast and precise data exchange across at least four independent
departments.
— New and/or changes in customer orders trigger the calculation of pro-
duction, warehouse etc. volumes.
The setup of the use case includes the subsequent steps, which according to
Perry et al. (Perry, Sim and Easterbrook, 2004) separate the case study from an
experience report:
1. Clearly defined research question
259
2. Systematic data collection and analysis in order to answer the research

question
3.2 Research Question and Data Collection
Due to the high pressure of consumer expectations, it is crucial to understand

the requirements of the market and to be able to adapt to them instantly, which
leads us to the research question:
What are the main requirements and design principles for the Self-Organizing
Adaptive Supply Chain?
As the proper design of the case study needed systematic data collection, it was
carried out on the basis of face-to-face interviews as well as through workshops
with managers and scholars. The process mapping and development of the tar-
get process acquired in each case over ten single interviews (see Table 1) with
experts from departments of production, logistics, controlling and planning on
national and international levels. The results were documented as event-driven
process chains (EPC) on a very detailed level as well as two flow charts on a higher
level, each of the process ”as is” and target process. Both EPCs and flow charts
were reviewed and if required supplemented during three two-day workshops
with the responsible managers. Subsequently, the developed architecture of a
current status of processes and systems was confirmed as appropriate by respon-
sible managers. At this stage of the projects, the data was analyzed mainly on a
qualitative level in form of process steps, timeline and information flow.
Table 1: Interviews for process mapping.
Company Timeline Interviewer Interviews

Coffee 10/2011 – First Author, external 14
manufacturer 12/2011 Consultant
LED producer 11/2013 – First Author 23
04/2014
Manufacturer of 09/2016 – Both Authors 12
air filters 12/2016
260
In pursuance of a development of a robust target process, the method of the

“operation recording” (Herrmannsdoerfer and Koegel, 2010; Lippe and Van Oost-
erom, 1992; Langer et al., 2013) was adopted – which is a documentation of the
processes within the modelling environment, while the processes are actually
being performed.Subsequently, the developed architecture of a current status of
processes and systems was confirmed as appropriate by company executives.
3.3 Process Mapping
In order to document and present the results of data collection in a structured

and comprehensible way the concept of Enterprise Architecture was chosen,
which is according to Ross (2006), includes business processes and IT infrastruc-
ture in order to fulfill long-term operation needs as well as to plan the projects.
ISO/IEC/IEEE defines an architecture “as composed of: (a)the fundamental orga-
nization of a system embodied in its components; (b) their relationships to each
other, and to the environment; and (c) the principles guiding its design and evolu-
tion.”(ISO/IEC 42010: 2007-Systems and software engineering–Recommended
practice for architectural description of software-intensive systems, 2007). Other
scholars (Tekinerdogan, 2014) add that a software architecture represents the
structure of the systems on a high-level and is relevant for the stakeholder com-
munication, design decisions including the planning of implementation project
and allocation of tasks.
In this way, the study supports the IT experts with the business knowledge that is
necessary for the supply chain digitalization. Although IT units primary knowledge
is technical, it is essential that IT managers understand the potential challenges
for the SC digitalization from a business perspective (Xue, 2014), which leads to
more effective communication between IT and operation departments and thus
better results for the development projects.
261
4 Architecting the Self-Organizing Adaptive Supply Chain
4.1 Case Description
Company X is a typical medium-sized enterprise with a business model of pur-

chasing raw materials from outside and executing operations in-house (Kumari
et al., 2015), then selling products to diverse customers (as shown in Material
Flow in figure 1). For example, a coffee company would buy green coffee beans,
roast them (equal to the “pre-production” process step), blend the beans and
pack them in diverse types of packaging in units of 1000g, 500g or 250g (equal to
“production planning 2” step as shown in figure 1). As with many SMEs, company
X had no in-house consulting or a strategic IT department. Subsequently, the busi-
ness grew over the years, but supporting systems stayed at the same level as of
their implementation over ten years ago. In contemplation of significant increase
of customer demand other production locations with additional warehousing
were acquired. Despite additional production capacity, the order-lead-time (time
between the customer order and delivery of the order) was not shortened. In
times, where Amazon and other “Big Players” offer very short delivery times,
order-lead-time became one of the critical factors of success in the consumer
market. For these reasons, the main aim of the project was to provide the IT
solution for an adaptive supply chain with the following targets:
— End-to-end digitalization of data along the supply chain
— Shorter order-lead-time (from over 2 weeks to 1 week)
— Lower level of inventory (reducing the average 4-6 week safety stock to 2
weeks)
4.2 Adaptive Supply Chain Architecture
The approach of the visualizing of existing business processes and systems was
based on principles presented by Panunzio (Panunzio and Vardanega, 2014), who
underlines that software architecture:
— Should stay at a high-level, “does not pollute”

— And at the same time “meaningfully represents those entities and their
semantics”.
262
Figure 1: Operational Processes ”As Is”
263
For these reasons the Enterprise Architecture was structured into the three most
important levels:
— High-level business processes (which also match the organizational struc-
ture of departments)
— Information flow including systems used for decision-making
— Material flow from suppliers to customers.
Based on numerous face-to-face interviews the first version of the Enterprise
Architecture “As Is” was created (the processes outside the supply chain such
as finance, marketing etc. were left out of scope). It showed that although the
ERP system contains all the information about articles such as master data, bill
of material, the information about customers and suppliers; the relevant deci-
sions on volume of orders, evaluation of supplier performance (at purchasing
departments) and production planning (in production departments) were still
met based on diverse reports exported into Excel. This way of working caused a
very slow reaction by the company to the changes in customer orders (as shown
in figure 1). Additionally, the data in different departments had different levels
of accuracy in comparison to the real-time data. For example, the production
orders for Monday in the production system were based on data from Thursday
of the last week, whereas pre-production operations were lagging even further
behind. The purchasing department, which purchases the raw materials, takes
that decision not based on actual customer orders (they will still be taken into a
consideration), but on the information from production, thus creating additional
waiting time. All these facts lead to the conclusion, that the more participants,
who change the data and need 1-3 days for the alignment with others, a process
has, the longer the order-lead-time will be.
The next step was the planning and execution of workshops with managers of
respective departments who are responsible for business processes. As a result,
all participants agreed that the current IT systems do not provide the required level
of support to the processes and should be improved. The material flow was left
out-of-scope for the target architecture because the increase in its complexity over
recent years due to new production bases and warehouse as well as an additional
customer channel had no significant impact on the structure of the processes
or systems. Considering that the decision-making process for the production,
pre-production and purchasing of goods took over 50% of the order-lead-time, it
was agreed to automate those calculation processes without substantial changes
in material flow.
264
Figure 2: Adaptive Supply Chain
265
Which led to the target process, as shown in figure 2. In target process “Adap-
tive Supply Chain with Autonomous Information Flow” the customer orders are
viewed as a trigger for the new calculation of production volume, which leads to
the calculation of the pre-production volume, which as a result can cause new
purchasing orders. Thus all departments will receive the data at the same time
and the entire Supply Chain will be able to adapt itself to new customer orders
within a short period of time.
The target architecture of systems and processes is not only intended to imple-
ment a new customized IT solution but also to change the information flow and
as a result the entire method of working. It was agreed to define the target pro-
cess based on the concept of Self-Organization/Autonomy with interconnected
“Smart Objects” for autonomous data processing in each process group, such as
Purchasing, Pre-Production, Production, as well as Warehouse & Transport as
defined by Bischoff and ten Hompel (Agiplan GmbH, Fraunhofer IML and Zenit
GmbH, 2015). In this way, they can independently analyze and share the data as
well as take decisions on production and purchasing volumes. As a consequence,
the people who work, for example, in purchasing department will be supported
with data in near real-time mode, and employees in different departments will
have same data at the same time.
4.3 Quality Requirements for Master Data
The self-organizing system can only exist if the master data is correct and the
connection between single products, parts etc. is explicitly defined. For example,
from the perspective of production planning of LED bulbs the end product of
so-called twin-blister (2 products in 1 packaging) is the same product as “normal”
bulb of the same type and can be produced in the same batch. For the packaging,
warehouse and customer/consumer these are unequal products. If the twin-
blisters should be sold from display instead of from regular shelves, they will be
seen as different product type by the marketing department, even if they seem to
be the same product for the customers and consumers. Consequently, the End-to-
End communication strongly depends upon the ability to dissolve and reassemble
the data generated by the needs of each process group or “Smart Object”. Thus for
the accurate information exchange throughout numerous departments master
data should fulfill following requirements:
266
— All product variants such as differences in packaging for the co-equal

product, products in marketing campaigns, packaging with different
languages etc. should be differentiated in a system.
— All product components should be clearly identified in a system, regard-

less of their size.
— All bills of material and packing lists should be kept up-to-date.
— Explicit assignment of data – one product number should not be used for
more than one type of product.
4.4 Implementation
Although the target architecture at a high-level was developed in one piece, the
development and deployment of the software were separated into work packages
according to the principles of the continuous deployment (Shahin, 2015):
— Small and independent deployment units.
— Considering operational aspects.
— Testability inside the architecture.
The operational focus in the architecture allowed the separation of the software
solution into manageable independent units, starting with the operation next to
the customer orders and moving toward the purchasing orders. Such an approach
freed up the focusing on one task at a time, making the whole process transparent
and manageable.
For example, the first task was to automate the decision in the production plan-
ning “should we produce the ordered volume or should we deliver the goods
from the available stock?” At first sight, it is a very complex decision which in-
volves many factors such as information about customers, seasonality for some
products due to Christmas and other typical marketing promotions. There are
many theories, on how to make an accurate decision as well as demonstrations of
wrong decisions i.e. well-known Beer Distribution Game from Sterman (Sterman,
1989), which represents a tendency to accumulate unnecessary inventories called
the Bullwhip effect (Forrester, 1968). Nevertheless, even under such complex
circumstances, it is possible to automate the decision process and even improve
267
the overall performance (Costas et al., 2017) using statistics and algorithms for
the calculation of the safety stock.
Subsequently, the target process for the production planning was tested for one
product group simultaneously to the “as is process”, which led to further improve-
ments on the master data (s. Chapter 4.4). As a result, the production planning
was performed ongoing in the background, starting the new calculation as soon
as a new customer order which was higher than average planned volume entered
the ERP system. In comparison to the old process, where the production planning
was performed once per week one week upfront to the start of production, it
drastically increased the reaction time of the whole supply chain and shortened
the order-lead-time.
Other working packages were designed based on same logic as an example, mak-
ing the whole process robust and transparent, saving time for the data analysis
and decision-making.
In the first project the validation of the software was based on the Requirements
Specifications which represented the target Architecture of IT-Systems on a de-
tailed level. The requirements were documented and divided into mandatory and
optional functions of the software. The validation of the second project solution
was result-based: the automated flow of data improved the forecast accuracy by
13% five months in a row, which led to more precise purchasing volumes and as a
consequence to lower level of stocks of finished goods and spare parts. The third
project will be validated in 2017 based on a simulation of the business processes
in a real-time mode simultaneously to the processes ”As Is”.
4.5 Risks and Benefits
Pressure to deliver high-quality value and deploy faster (Shahin, 2015) may cause
a situation where sub-optimal solutions will be chosen over originally required
systems, which would cause “architectural technical debt”, which is similar to a
financial debt and has to be repaid in the long term (Vogel-Heuser et al., 2015).
In our case, it would mean the risk of introducing a new solution which is incom-
plete, so that both old and new way of working would exist at the same time,
causing high costs, low efficiency of a new system and most importantly, mistakes
in relevant business decisions such as purchasing volume, production and the
delivery of goods.
268
5 Discussion
Despite the risks and high expenses of the implementation, the Adaptive Self-
Organizing Supply Chain can produce multiple benefits, which can be divided
into operational and strategic (Xue, 2014).
— Operational: lowering logistics costs, lowering operational costs, improv-

ing inventory control, increasing inventory turnover.
— Strategic: establishing a competitive advantage, improving customer
relationships, improving supplier relationships.
In our use case, all the companies benefited from a shortening of order-lead-
time by at least two weeks, which lead to lower level of inventory, not only for
finished goods but also for spare parts and consequently, the costs for storage
and handling. Thus, the benefits unquestionably outweighed the risks, which
reduced the resistance towards changes within the company and accelerate the
process of gaining acceptance.
5 Discussion
5.1 Contribution to Research and Practice
This study contributes to the topics of Self-Organizing Systems and Adaptive

Supply Chains by providing a clear structured use case. The combination of real-
time online data and increasing computer processing power has an immense
impact on architecting new systems since it allows autonomous data collection
and processing (Mortenson, Doherty and Robinson, 2015). In our use case, the
automated data flow had a massive impact on the order-lead-time and accuracy
of information.
However, as exemplified in the analysis of 52 research papers undertaken by Dikert
(Dikert, Paasivaara and Lassenius, 2016) and a survey of almost 4,000 respondents
executed by VersionOne Inc.(VersionOne Inc, 2016), the following barriers must
be taken into consideration on the way to the implementation of projects:
— General organizational resistance. As stated in the ”Manifesto for Agile

Software Development”, which came out in 2001 in USA (Fowler, Martin,
Highsmith, Jim and Cockburn, Alistair, 2017), it is important that business
people work together with the software development throughout the
269
whole project in order to enforce the communication while lowering the

resistance.
— Pre-existing rigid framework. Since the new software solutions will be

implemented on a basis or with the co-existence with older software
models, it is crucial to understand the evolution of predecessor models
(Langer et al., 2013). Clearly defined requirements for the software allow
for the technical implementation as well as the contractual relationship
between the software providers and the process owners, which allow
mitigating of the relevant project risks (Panunzio and Vardanega, 2014).
5.2 Limitations
The impact of the new system could be quantified only in terms of order-lead-time
and accuracy of information. The impact on logistics and operational costs could
not be precisely measured, due to complexity in the real-world environment:
— Change of sales volume at the end of various periods of time. Sales

volume has a massive impact on costs per product since logistics costs
are highly impacted by rent costs of the warehouse, costs of warehouse
operator etc. Therefore it is hard to evaluate the exact impact of the new
software implementation if the sales and product range were changed;
even if the costs at the end of the reporting period were lower than before
the rollout of the software.
— Other projects and marketing campaigns. Since the software project was
not the only project in companies, it is impossible to separate the impact
of one project on a company’s performance (in all three cases the compa-
nies operate globally). A good example is the almost 40% reduction of
so-called penalty costs (costs to B2B customers for not achieving service
level agreements) after the roll-out of software. Although the software
doubtless supported this achievement, it is still difficult to measure the
exact level of influence, due to simultaneous application of a new system
of financial reward for the employees by the management.
In order to investigate the potential for costs reduction due to the implementa-
tion of Self-Organizing Adaptive Supply Chain, the Business Innovation Lab at
University of Applied Sciences in Hamburg will launch a research project, in which
270
6 Summary
the influence of single parameters can be changed and evaluated on demand in

controlled environments, based on the simulation of events.
A further limitation results in the difficulty of standardization of an IT-Solution
for the development and implementation of a Self-Organizing Adaptive Supply
Chain since every company has a specific business model and already has one
or more customized IT-System with which new software will have interfaces. For
this reason, it is important to have a methodological approach for evaluating the
potentials and problems in each single case.
Due to the high complexity of software projects, the study is limited to the area
of supply chain processes. However, the proposed approach can still be used
for the development of an architecture for Self-Organizing Systems within other
business areas such as marketing, R&D and sales.
6 Summary
The main task of the Supply Chain is to steer the processes throughout the differ-
ent departments both inside and outside the company. This can only increase
the performance of the entire company if the data collected by different partic-
ipants and systems are interlinked with each other instead of the digitalization
of the activities of only one department (Yuvaraj and Sangeetha, 2016; Coltman,
Devinney and Midgley, 2011; Xue, Ray and Sambamurthy, 2013). It is also im-
portant to keep in mind, that an autonomous information flow does not mean
100% computer-based decisions. There is still some low percentage of errors or
situations, where human interaction is required. However, even the partial imple-
mentation of the automated information flow makes the whole supply chain more
robust and flexible to the changes in customer demand. The implementation of
the self-organizing adaptive supply chain as a new and innovative business strat-
egy supports companies by the extension of transparency of organization-wide
processes. This in turn, enables data-based decision-making processes, which
impact the market share and even help us to discover new strategic capabilities
(Hazen et al., 2016). When supply chain activities achieve better results through
the automation of decisions, it generates cost savings and improves the economic
benefit for the company, shaping the supply chain into “the source of enterprises
core competence” (Vogel-Heuser et al., 2015).
271
References
Agiplan GmbH and Fraunhofer IML and Zenit GmbH (2015). Erschliessen der Potentiale der Anwen-
dung von ”Industrie 4.0” im Mittelstand.
Bauernhansl, T., M. ten Hompel, and B. Vogel-Heuser, eds. (2014). Industrie 4.0 in Produktion, Au-
tomatisierung und Logistik. de. DOI: 10.1007/978-3-658-04682-8. Wiesbaden: Springer Fachme-
dien Wiesbaden.
Bogaschewsky, R. and H. Müller (2016). Industrie 4.0: Wie verändern sich die IT-Systeme im Einkauf
und SCM?
Coltman, T., T. M. Devinney, and D. F. Midgley (2011). “Customer relationship management and
firm performance”. In: Journal of Information Technology 26.3, pp. 205–219.
Costas, J., B. Ponte, D. de la Fuente, J. Lozano, and J. Parreño (2017). “Agents Playing the Beer
Distribution Game: Solving the Dilemma Through the Drum-Buffer-Rope Methodology”. In:
Engineering Systems and Networks. Springer, pp. 337–345.
Dar, K., A. Taherkordi, H. Baraki, F. Eliassen, and K. Geihs (2015). “A resource oriented integration
architecture for the Internet of Things: A business process perspective”. en. In: Pervasive and
Mobile Computing 20, pp. 145–159.
Dikert, K., M. Paasivaara, and C. Lassenius (2016). “Challenges and success factors for large-scale
agile transformations: A systematic literature review”. en. In: Journal of Systems and Software
119, pp. 87–108.
Dittrich, Y., K. Rönkkö, J. Eriksson, C. Hansson, and O. Lindeberg (2008). “Cooperative method
development: Combining qualitative empirical research with method, technique and process
improvement”. en. In: Empirical Software Engineering 13.3, pp. 231–260.
Forrester, J. W. (1968). “Industrial dynamics—after the first decade”. In: Management Science 14.7,
pp. 398–415.
Gorschek, T., P. Garre, S. Larsson, and C. Wohlin (2006). “A model for technology transfer in practice”.
In: IEEE software 23.6, pp. 88–95.
Gunasekaran, A., M. Kumar Tiwari, R. Dubey, and S. Fosso Wamba (2016). “Big data and predictive
analytics applications in supply chain management”. en. In: Computers & Industrial Engineering
101, pp. 525–527.
Hazen, B. T., J. B. Skipper, J. D. Ezell, and C. A. Boone (2016). “Big data and predictive analytics for
supply chain sustainability: A theory-driven research agenda”. en. In: Computers & Industrial
Engineering 101, pp. 592–598.
Herrmannsdoerfer, M. and M. Koegel (2010). “Towards a generic operation recorder for model
evolution”. In: Proceedings of the 1st International Workshop on Model Comparison in Practice.
ACM, pp. 76–81.
Jones, D., S. Behrens, K. Jamieson, and E. Tansley (2004). “The rise and fall of a shadow system:
Lessons for enterprise system implementation”. In: ACIS 2004 Proceedings, p. 96.
Kassahun, A., R. Hartog, and B. Tekinerdogan (2016). “Realizing chain-wide transparency in meat
supply chains based on global standards and a reference architecture”. en. In: Computers and
Electronics in Agriculture 123, pp. 275–291.
Kersten, W., T. Blecker, C. M. Ringle, and epubli GmbH, eds. (2015). Sustainability in logistics and
supply chain management: new designs and strategies. eng. 1st ed. Proceedings of the Hamburg
International Conference of Logistics (HICL) 21. OCLC: 933219309. Berlin: epubli GmbH.
272
References
Kumari, S., A. Singh, N. Mishra, and J. A. Garza-Reyes (2015). “A multi-agent architecture for
outsourcing SMEs manufacturing supply chain”. en. In: Robotics and Computer-Integrated
Manufacturing 36, pp. 36–44.
Langer, P., M. Wimmer, P. Brosch, M. Herrmannsdörfer, M. Seidl, K. Wieland, and G. Kappel (2013).
“A posteriori operation detection in evolving software models”. en. In: Journal of Systems and
Software 86.2, pp. 551–566.
Lippe, E. and N. Van Oosterom (1992). “Operation-based merging”. In: ACM SIGSOFT Software
Engineering Notes. Vol. 17. ACM, pp. 78–87.
McKinsey Digital (2016). Industry 4.0 after the initial hype. Where manufacturers are finding value
and how they can best capture it.
Mortenson, M. J., N. F. Doherty, and S. Robinson (2015). “Operational research from Taylorism to
Terabytes: A research agenda for the analytics age”. en. In: European Journal of Operational
Research 241.3, pp. 583–595.
O’Leary, D. E. (2008). “Supporting decisions in real-time enterprises: autonomic supply chain
systems”. en. In: Information Systems and e-Business Management 6.3, pp. 239–255.
Panunzio, M. and T. Vardanega (2014). “A component-based process with separation of concerns
for the development of embedded real-time software systems”. en. In: Journal of Systems and
Software 96, pp. 105–121.
Perry, D. E., S. E. Sim, and S. M. Easterbrook (2004). “Case studies for software engineers”. In:
Software Engineering, 2004. ICSE 2004. Proceedings. 26th International Conference on. IEEE,
pp. 736–738.
Plattform Industrie 4.0 (2016). Aspects of the research roadmap.
Ponte, B., E. Sierra, D. de la Fuente, and J. Lozano (2017). “Exploring the interaction of inventory
policies across the supply chain: An agent-based approach”. en. In: Computers & Operations
Research 78, pp. 335–348.
Rentrop, C. and S. Zimmermann (2012). “Shadow IT”. In: Management and Control of Unofficial IT.
ICDS, pp. 98–102.
Runeson, P. and M. Höst (2009). “Guidelines for conducting and reporting case study research in
software engineering”. en. In: Empirical Software Engineering 14.2, pp. 131–164.
Shahin, M. (2015). “Architecting for DevOps and Continuous Deployment”. en. In: ACM Press,
pp. 147–148.
Sterman, J. (1989). Modelling Managerial Behavior: Misperceptions of Feedback in a Dynamic
Decision Making Experiment.
VersionOne Inc (2016). 10th Annual State of Agile Report.
Vogel-Heuser, B., A. Fay, I. Schaefer, and M. Tichy (2015). “Evolution of software in automated pro-
duction systems: Challenges and research directions”. en. In: Journal of Systems and Software
110, pp. 54–84.
Voß, P. H., ed. (2015). Logistik – eine Industrie, die (sich) bewegt. de. DOI: 10.1007/978-3-658-10609-6.
Wiesbaden: Springer Fachmedien Wiesbaden.
Waterman, M. G. (2014). “Reconciling agility and architecture: a theory of agile architecture”. In:
Wellbrock, W. (2015). Innovative Supply-Chain-Management-Konzepte. de. DOI: 10.1007/978-3-658-
09181-1. Wiesbaden: Springer Fachmedien Wiesbaden.
Xue, L. (2014). “Governance–knowledge fit and strategic risk taking in supply chain digitization”.
en. In: Decision Support Systems 62, pp. 54–65.
Xue, L., G. Ray, and V. Sambamurthy (2013). “The impact of supply-side electronic integration on
customer service performance”. en. In: Journal of Operations Management 31.6, pp. 363–375.
273
Supply Chain Process Oriented
Technology-Framework for Industry 4.0
Florian Schlüter1 , Endric Hetterscheid1
Digitalization in supply chains is an arising topic. Screening recent publications

in this field delivers no application-oriented classification of potential Industry 4.0
technologies for supply chain processes. The purpose of this paper is to present
an application-oriented Technology-Framework speeding up the development
of digitalization scenarios for supply chain processes.A structured literature re-
view will be carried out to identify and extract relevant technologies within the
field of Industry 4.0. Afterwards, the Process Chain Model will be used to come
up with standardized and application-oriented categories for technologies to
derive a framework. Based on a comparative study, the Process Chain Model was
especially developed for illustrating and transforming supply chain and logistics
processes.The new Technology-Framework supports the transformation process
by giving the user (1) a standardized framework for technology categorization
and (2) by giving a comprehensive overview about existing technologies in the
literature. The framework serves as a basis for future supply chain process trans-
formation. By grouping the technologies around approved categories – originally
defined for describing process chain elements – it speeds up the identification
and selection of appropriate technologies.
Keywords: Industry 4.0; Supply Chain Management; Digitalization; Technology

scouting
275
Supply Chain Process Oriented Technology-Framework for Industry 4.0
1 Introduction
Digitalization plays an important role in many sections of daily life. In the in-
dustrial environment, more and more companies are also pursuing measures to
integrate technologies into business processes and thus to digitize their processes
(Roth, 2016b). In the literature exist further terms with a similar meaning for digi-
talization (Hermann, Pentek and Otto, 2016). Especially the term of Industry 4.0
(I4.0) has determined both the economy and science for some years (Wan, Cai and
Zhou, 2014). Background of the terminology is the increasing customer require-
ments for individualized products. The resulting smaller batch numbers down till
batch size one and shorter product life cycles lead to steadily growing challenges
not only for production and logistics, but throughout the entire supply chain (SC)
(Bauernhansl, 2014). In order to counteract this increasing complexity and to
meet the market requirements with regard to price, quality and time, companies
are forced to increase their technological standard in their SC processes (Kersten,
et al., 2016). Thereby I4.0 pledges improvement in flexibility, product quality,
delivery time and deliverability in value-added networks (Bauernhansl, 2014).
The term of I4.0 is described by the use of existing technologies, such as micro-
computers, broadband internet access, radio frequency identification (RFID), and
stronger networking creating a comprehensive transparency and processing of
information along the entire value chain (Siepmann, 2016a). Additionally, I4.0 is
also to be seen as a new optimization approach, because a faster information flow
results from a reduction in media discontinuities and a stronger technological and
organizational process transparency. As a result, more tightly clocked process
chains emerge (Schlick, et al., 2014).
Although the term of I4.0 has been in existence for several years, there is still no
uniform definition (Siepmann, 2016b). For this paper, the holistic definition of
Roth (2016b) will be used. He defines I4.0 as ”[...] the networking of all human and
machine actors over the entire value chain as well as the digitalization and real-
time evaluation of all relevant information with the aim of making the processes of
value creation more transparent and efficient in order to use intelligent products
and services to optimize customer use”.
The “Dortmund Management-Model for Industry 4.0” by Henke gives a further di-
rection for practitioners and researchers with establishing and formalizing “work-
clusters” in a two-dimensional matrix for transforming value creating activities
(ten Hompel and Henke, 2017). The first dimension describes a company per-
spective with ”Technology”, ”Organization” and ”Process” as characteristics. The
276
1 Introduction
second dimension has a management perspective with ”normative Level”, ”Plan-

ning”, ”Implementation” and ”Monitoring” as characteristics. The following paper
contributes to the work-cluster of ”Technology scouting” within the ”Technology”
and ”Planning” dimensions.
In this context, science is discussing the extent to which a revolution can be ad-
dressed in I4.0. Accordingly to Dais (2014), there are two different groups regarding
interpretation of the term. One group recognizes I4.0 as revolutionary as it brings
innovative improvements with it that have not yet existed. The other group speaks
of evolution, since the technologies are ”nothing new”, but constantly evolving
and improving.
Despite this fact, I4.0 technologies can make a valuable contribution to the digi-
talization and autonomization of processes. However, not many efforts beyond
I4.0 pilot projects are carried out in practice. Oftentimes, I4.0 technologies are
implemented only partially in the logistics or production, so that the desired cost
and benefit effect for companies does not occur (Graef, 2016). The reason for
partial solutions is often insufficient knowledge about the correct use of the right
technology for the respective process. Meanwhile, a large number of technolog-
ical solutions exist to improve business processes (Kersten, et al., 2016) and a
support is required to help companies in the correct use of I4.0 technologies in
existing SC processes.
In recent literature, I4.0 technologies and fields, illustrated with mind maps and
other classification methods, are known (Pfohl, Yahsi and Kurnaz, 2015; Kersten,
et al., 2016). However, these classifications are not very suitable for practice.
Also I4.0 technologies are described in use cases in which these technologies are
often used only in production or logistics and not in the overall context of a SC
(e.g. Zhang, et al., 2014). In addition, there is no framework in the literature that
provides the technologies of I4.0 for the optimization of SCs. Therefore, the aim of
this paper is to develop a Technology-Framework that provides methodological
support for the optimization of SC processes. For this purpose, the following
research questions (RQs) arise:
RQ 1: Which existing technologies in the literature relate to the context of I4.0 and
how can they be classified for practical usage?
RQ 2: How can a relation between the identified technologies and classical process
optimization methods be established?
In order to answer these RQs, the methodological approach will be explained
in section 2. In section 3, the methodology will be executed and an overview of
277
the existing literature in the research field will be provided. The development
of the Technology-Framework by listing existing I4.0 technologies, establishing
a relation between them and process optimization as well as integrating the
results into the optimization framework will take place in section 4. In section 5,
a conclusion and further needs for research follow.
2 Methodology
Today, the whole topic of I4.0 is still an emerging, under-developed and highly
diverse research field (Pfohl, Yahsi and Kurnaz, 2015; Glas and Kleemann, 2016)
and the application of more exploratory research methods is suitable to give
first directions (Stebbins, 2001) instead of a final answer. Three principal ways of
conducting exploratory research are mentioned in the literature (Saunders, Lewis
and Thornhill, 2009): Search of the literature; Interviewing experts in the subject;
and Conducting focus group interviews.
At first a structured literature review (SLR) based on an adopted approach for
systematic literature review from Denyer and Tranfield (2009) has been performed,
to answer the preceded RQs. Reviews that are evidence-based like SLRs allow for a
higher objectivity of the search results while eliminating error or bias issues (Kilubi
and Haasis, 2016). Denyer and Tranfield (2009) defined the following stages: (1)
Question formulation; (2) Locating studies; (3) Study selection and evaluation; (4)
Analysis and synthesis; and (5) Reporting and using the results – a more detailed
description can be found in the literature. For each of those stages the authors
proposed more detailed steps to fulfill each stage. The whole methodology is
summarized in Figure 1.
Stage (1) has already been completed with formulating the RQs in section 1. Stage
(2) and (3) follow in the next section (see section 3). There existing literature will
be located and selected based on criteria. The section ends with an overview
about existing publications which will be used for stage (4) (see section 4). In
stage (4) the selected literature will be screened for relevant technologies and the
Technology-Framework will be developed based on the findings and supported
through group discussions (see section 4).
278
2 Methodology
Step 1: Definition of search terms and data

Stage (1) bases
Question Formulation
Step 2: Definition of evaluation criteria for
study selection
Step 3: Search in data bases

Stage (2)
Locating studies Step 4: Additional search based on citation
review and/or within additional search
engines
Stage (3) Step 5: Overview about existing publications

Study selection and
evaluation Step 1: Screening the literature and
removing of duplicates and/or different
definitions
Stage (4) Step 2: Selection of an appropriate process

Analysis and synthesis design and optimization framework
Step 3: Preparation of documents for group

discussion
Stage (5)
Reporting and using the Step 4: Execution of group discussions
results
Step 5: Integration of results into framework
Figure 1: Proposed methodology
279
3 Research Overview based on located and selected

studies
Step 1: Definition of search terms and data bases
During a brainstorming session the key words “Industrie 4.0” and “Industry 4.0”
have been selected. The authors chose these words to gather as much literature as
possible for collecting mentioned technologies. Both terms are usually connected
through an ”OR” operand for the search engines – this leads to results when at
least one of the terms is included in the title, abstract or key words. The actual way
of keyword combination depended on the used search engine, thus test searches
took place to identify the correct way of formulation.
The SLR uses a broad range of databases based on screening existing literature
reviews like (Ho, et al., 2015; Kilubi and Haasis, 2016). The following databases
have been identified as suitable: IEEExplore, Science Direct, Springer, Emerald,
EBSCOhost and Taylor and Francis. Also the publications should be in English or
German.
Step 2: Definition of evaluation criteria for study selection
The search in databases and study selection took place in April 2017 and is based
on three evaluation criteria, related to the research questions:
Criterion 1: Cluster and/or lists of technologies in the context of I4.0
Criterion 2: Description of a use case in the context of I4.0
Criterion 3: Description of a methodological SC process optimization based on

technologies in the context of I4.0
Criterion 1 and 2 have been chosen because technologies in the context of I4.0
are often mentioned within lists or specific use cases. The third criterion has been
chosen to include existing literature possibly covering the second RQ.
280
3 Research Overview based on located and selected studies
Table 1: Overview of search results

Database Results 1. Check 2. Check
EBSCOhost 69 15 0
Emerald 15 4 0
IEEExplore 319 49 2
ScienceDirect 573 66 0
Springer 1.963 123 7
Taylor & Francis 44 11 0
Sum 2.983 268 9
Step 3: Search in data bases
During the search process, a publication has to fulfill at least one of the mentioned
criteria when checking the abstracts or in case of a book the introduction and
index. Afterwards the remaining literature has been screened in more detail
based on the aforementioned criteria. The number of publications in scientific
databases can be seen in the following table (see Table 1).
As mentioned above the topic of I4.0 is still an emerging research field (Pfohl, Yahsi
and Kurnaz, 2015; Glas and Kleemann, 2016). Which results in only a very limited
number of classified research publications with relation to I4.0 and the formulated
RQs. Publications about I4.0 can mostly be found in scientific magazines and
lower-rated journals or they are studies published by companies or research insti-
tutes (Pfohl, Yahsi and Kurnaz, 2015). That’s why an additional semi-structured
literature search is necessary.
Step 4: Additional search
Because Google Scholar is more complete than other databases (Kilubi and Haasis,
2016), and covers also scientific magazines, lower-rated journals as well as studies
it has been used additionally.
281
Step 5: Overview about existing publications
Applying the aforementioned selection criteria brings additional 23 publications

leading to 32 in total. The 32 publications are evaluated in the Table 2 below.
As it can be seen in Table 2, most of the published literature describes potentials
of certain technologies in the context of I4.0 or a use case where technologies are
already applied in industry. In the third column, a paper is marked if it describes
the link between I4.0 technologies and SC processes. Few are marked and all of
them fulfill this condition only partly. The reason is that they are describing mainly
the potentials of I4.0 technologies in SC processes (Rozados and Tjahjono, 2014;
Pfohl, Yahsi and Kurnaz, 2015; Schrauf and Berttram, 2016). None of the identified
literature describes a methodological approach which links a SC optimization
framework with I4.0 technologies giving guidelines during the SC optimization
process.
282
3 Research Overview based on located and selected studies
Table 2: Overview and evaluation of publications
Methodological
Listing of
Use case link to
Source Technol-
related SC process
ogies
(optimization)
Bauer, et al. (2014) × × (×)
Bauernhansl, et al. (2016) × ×
Bartodziej (2017) ×
Bechtold, et al. (2014) × (×)
Bauernhansl (2014) ×
Bienzeisler, Schletz and Gahle (2014) ×
Bischoff (2015) × ×
Bloching, et al. (2015) ×
Gausemeier, et al. (2016) ×
Geissbauer, Vedso and Schrauf (2016) ×
Hausladen (2016) ×
Hermann, Pentek and Otto (2016) ×
Horvath & Partners (2017) ×
Huber (2016) × ×
Jäger, et al. (2015) ×
Kersten, et al. (2017) × ×
Obermaier (2016) ×
Pfohl, Yahsi and Kurnaz (2015) × (×)
Plattform Industrie 4.0 (2015) ×
Roth (2016a) × ×
Rozados and Tjahjono (2014) × × (×)
Rüßmann, et al. (2015) × ×
Schlaepfer, Koch and Merkofer (2015) ×
Schrauf and Berttram (2016) × (×)
Schwab (2016) ×
Seiter, et al. (2016) × ×
Siepmann (2016b) × ×
Wee, et al. (2015) ×
Wehberg (2015) ×
Wischmann, Wangler and Botthof (2015) ×
Wollschlaeger, Sauter and Jasperneite ×
(2017)
Zillmann and Appel (2016) × × (×)
× = Fully fulfilled; (×) = Partly fulfilled
283
4 Study analysis and synthesis for framework

development
After study selection and thus completing step 2 and 3 of the proposed method-
ology (see section 3 and figure 1) the aim of this section is to identify and select
technologies within the field of I4.0 through literature screening and to develop a
usable framework, in accordance with step 4 of the proposed methodology (see
figure 1).
Step 1: Screening the literature and removing of duplicates/and or

different definitions
By screening the literature, 124 artifacts related to I4.0 could be identified which
can be interpreted as I4.0 technology. The artifacts will be divided between tech-
nologies and I4.0 concepts. Due to the missing definition of I4.0 technology the
authors decided to define such a technology as “directly recognizable entity and
real existing physical hardware or logical software which is financially activat-
able and supports or realizes the principles and ideas of I4.0”. The next step was
to remove duplicate and/or different definitions with the same meaning (see
Figure 1). Also similar technologies have been clustered to technology fields to
make the later combination with the process optimization framework easier. As a
result, the number of technological artifacts could be reduced to 45. Based on
that definition, the authors could divide the artifacts as described in the list below
(see Table 3).
The concepts are more overarching and abstract ideas or principles which be-
come real by using the technologies (e.g. by using Cyber-Physical-Systems and
analytics a predictive maintenance can be realized; using smart devices together
with other devices over the internet creates the Internet-of-Things) while the tech-
nologies are more tangible through certain hardware, software or objects. Only
the technologies on the left side have been used for the later framework. This list
also answers the first RQ.
284
4 Study analysis and synthesis for framework development
Table 3: I4.0 Artifacts classified in Technology and Concept

Technology Concept
3D-Printing Augmented Reality
Actuators Automation
Analytics Cloud Computing
Apps Gamification
Artificial Intelligence (Software) Horizontal Integration
Autonomous Transport Vehicles Internet of Things and
Services
Big Data IT-Security
Blockchain Machine to Machine
Communication
Business Management Software/Systems Predictive Maintenance
(e.g. ERP, APS)
CAx (e.g. CAD, CAM) Smart Factory
Cyber-Physical-System Smart Grids
Data glasses/Head-Mounted Display Smart Logistics
Data Mining Ubiquitous Computing
Digital Shadow Vertical Integration
Embedded Systems Virtual Reality
Human Machine Interaction (e.g. Touch
interfaces)
Identifiers (e.g. Barcodes, RFID, QR-Code)
Image Recognition
Internet Protocol (e.g. IPv6)
Mobile Communication (Infrastructure)
Mobile Devices (e.g. Smart phone, tablet)
Pick-by-Technology
Real-time Data
Robotics
Sensors
Simulation
Smart Objects/Products
Smart Payment (Software)
Social Media
Wearables (e.g. Data glasses, Head mounted
displays)
Wireless Communication (Infrastructure)
Sum: 30 Sum: 15
285
Step 2: Selection of an appropriate process design and optimization

framework
Due to the focus on SC processes an appropriate process design framework comes

from Kuhn (1995). Based on a comparative study only the Process Chain Model of
Kuhn (1995) was especially developed for logistic and SC processes (Nyhuis and
Wiendahl, 2009) and is part of the Process Chain Instrument. The Process Chain
Model enables a holistic visualization and analysis of performance object flows.
Deficiencies in the processes can be clarified across hierarchy levels. Another ad-
vantage is the strong integration of the employees into process optimization. The
process chain representation can be used as a communication medium, since it is
a form which is understandable to all, irrespective of the organizational structure.
This greatly promotes the acceptance of process changes. During optimization,
processes are presented with the design elements system load (sources and sinks),
process flow, steering, structures and resources (Kuhn, 2008).
Sources and sinks form the interfaces between the system and the process chain
to its environment. Through the sources, performance objects enter the system
and leave it transformed over the sink. The entirety of all performance objects
that influence a process is called a system load. The process flow can be subdi-
vided into subordinate processes and these can be further detailed as required.
Together with the source and sink, they represent the process structure. The
process chain elements are subject to various rules and steering rules, e.g. the
decision-making scope of neighboring processes. This is referred as steering
and is divided into five different steering levels. For the transformation of the
performance objects, resources are claimed in the process. The steering levels
are responsible for minimizing resource costs through efficient use in the pro-
cess. The structures describe the classification of process chain elements into
a company. These design elements of a process are divided into 17 different
potential classes. With the help of the potential classes, processes can be pre-
cisely described and investigated for improvement potentials. Through this type
of a ”checklist” action alternatives with their respective effects in the process
optimization can be opened up (Kuhn, 2008). These 17 potential classes will be
explained below (subsequent (Kuhn and Hellingrath, 2002).
286
Process
Source: Sources are the inputs into a process chain element. They define the
input side system load and thus they represent an interface part between the
process chain element and the environment. Sources describe the performance
objects that must be transformed by the process per time unit.
Sink: Sinks represent the counterpart to sources. They describe which perfor-
mance objects (information or material) are transformed to the following process
chain elements. At the sink, the object triggers a retrieval behavior of the subse-
quent process.
Process structure: A process chain element represents the process structure with
the sources and sinks. The process types are processed/checked, transported,
stored or buffered. The process structure can be modified as part of an opti-
mization with the following process chain modulation: Grouping, parallelizing,
extending, shortening, eliminating or exchanging processes.
Resource
Personnel: This includes all employees who are available in a process during
the operating time. On the one hand, the employee is described in terms of his
training, qualification and motivation. On the other hand, the work organization
and the flexibility of working and break times are described. These parameters
significantly influence the performance of a process.
Area: The resource ”area” is used to list all the operational areas for the transforma-
tion of the basic object in a process. It has an impact on investment and operating
costs. Small areas lead to an extension of process times or to the limitation of
action options.
Stock: The stock comprises the number of basic objects in a process. A distinction
is made between materials (e.g. raw materials) and the number of customer or
production orders. Stocks are an important control lever in logistics because they
have a significant impact on a variety of logistical key figures.
Tools and Machinery: This includes all resources that are responsible for the
direct transformation of the object. Typical tools and machinery are production
machines and systems or conveyor and storage systems. In a logistical process,
287
transports which carries out the transfer of the object from a source to a sink can
also be assigned to this category.
Working aids: This category is assigned to all the resources that are required for
the transformation of objects in supportive form. These are e.g. loading aids
such as cranes. Particularly in external logistics is the selection of working tools
significant.
Organizational means: Organizational means are used to summarize all the re-
sources required for information processing in a process. These include informa-
tion carriers (e.g. bar codes), information storage, computer architectures, and
software programs.
Structure
Layout: The layout, which is often referred to as a topology, determines the

arrangement of the work equipment as well as the area and determines the
transport links. Thus, the layout significantly influences area usage. In an optimal
layout, envelope processes are tried to avoid.
Organizational structure: In this case, the process responsibility is determined
within a process chain element. This includes communication and decision-
making structures and thus has a significant influence on the ways of information
and material flow within the process structure. The goal is to share information
more quickly and make resources more flexible to use.
Technical communication structure: This category includes the technical net-
working of all implemented external and internal systems. These are IT and EDV
structures, which are necessary for processing.
Steering levels
Normative: In this category, a company’s superordinate values, objectives and

standards with reference to a process chain element are described. Thereby, the
change of the process is specified. All other potential classes must be able to
measure and evaluate these requirements.
Administrative: This level has the task of coordinating the system load with the
upstream and downstream process chain elements and assessing future loads. It
288
collects orders based on the sources and sinks and remit them to the normative
level. The administrative level also defines current targets, which are aligned with
the requirements of the normative level.
Dispositive: This level aims to analyze the system load of the administrative level
under existing boundary conditions and optimization criteria. The results are
provided to the network level. The orders are assigned to resources and the
sequencing of orders is carried out.
Network: The network level coordinates and synchronizes several related pro-
cesses using autonomous rules to meet customer requirements. The order as-
signment of the dispositive level is taken over and executed according to the
appropriate rules. The difference between the network levels at the administra-
tive level is that the optimized assignment of requests to resources is not taking
place, but rather the autonomous rules for the flexible use of interchangeable
resources.
Control: The control level has the task of executing the specifications from the
network level for a single performance object in a sub-process. The control sys-
tem performs simple decision rules. At this level, the targets are measured and
compared with the relevant agreements.
After selecting an appropriate framework for SC optimization and collecting the
relevant I4.0 technologies, the next step is to connect both aspects methodologi-
cally. This happens in the next step with the help of group discussions.
Step 3: Preparation of documents for group discussion
For assigning I4.0 technologies to the process design framework the authors chose
focus group discussions as an appropriate method. Those focus group interviews
capitalize on communication between research participants to generate data
(Kitzinger, 1995). This method is useful for exploring people’s knowledge and
experiences (Kitzinger, 1995). Two documents were created for the focus group
discussions. The first document contained the aforementioned technologies and
potential classes. Participants should mark for each technology where they see
a direct impact on which potential class. The second document was prepared
for the second discussion round in which the results were clustered based on the
potential classes and participants had to select whether or not they agree on the
technology assignment. In case they disagree they also should mention why, as
289
well as when they think a technology is relevant but was not assigned in the first
round.
Step 4: Execution of group discussions
Two group discussion rounds took place. At first four experts for topics of digi-
talization and I4.0 were selected, who are employees at a research institute. The
I4.0 technologies and potential classes were briefly explained by a moderator.
The objectives of the survey was named, in which the experts should assign the
technologies to the various potential classes. The experts should consider which
technical key feature of a particular technology could yield the greatest benefit
for a potential class with focus on the material flow. It has to be noted that the
Process Chain Model was originally designed to model material, information and
financial flows within SCs and depending on the kind of flow the meaning of the
potential class could change thus the relevance of a technology for a potential
class could also change and makes the assignment harder. It was allowed to
assign an I4.0 technology to several potential classes. The experts received an
overview of the technologies and different potential classes. They were given the
opportunity to talk about the task and to communicate their respective opinions.
The results were evaluated by the moderator and the discrepancy in the results
were presented and discussed. Afterwards, a second focus group discussion took
place, consisting of eight participants from research. The participants were dif-
ferent from those in the first round and their knowledge about I4.0 technologies
differed. This led to more discussions and created more group dynamic. Finally,
the documents were collected by the moderators, evaluated and the improved
results were used to form the following Technology-Framework.
Step 5: Integration of results into framework
The following framework is the result of a structured literature review and two
rounds of focus group discussions. From 17 potential classes 15 were used to
assign technologies. The potential classes ”Normative” and ”Layout” have not
been selected by the participants thus they were neglected from the framework.
The proposed Technology-Framework answers the last RQ and can now be used
within process optimization activities as a supportive tool to stimulate the creativ-
ity of the participants (see figure 2 - figure 3). It has to be noted that there may be
290
Figure 2: Technology-Framework (1/4) – Processes
cases where some technologies are not applicable in their particular potential
class or where additional technologies are suitable.
291
Figure 3: Technology-Framework (2/4) – Structures
292
Figure 4: Technology-Framework (3/4) – Steering levels
293
294 Figure 5: Technology-Framework (4/4) – Resources

5 Conclusion and Further Research
The paper presents a first approach of linking I4.0 technologies with specific ele-
ments of a SC optimization framework based on exploratory research methods.
The purpose is to guide practitioners during optimization activities by generally
giving an overview about technological options they have and also – by locating
improvement potential within the elements of a SC optimization framework –
giving an overview on which technologies suite best for certain problems. Linking
technologies with certain process characteristics makes it also easier to evaluate
the benefits because it is easier to select appropriate process KPIs for measure-
ment. The technologies are methodologically collected from the literature with
a structured literature review (see section 2 and 3) and the link to the SC opti-
mization elements has been created with focus group discussions (see section
4).
5.1 Limitations and further research
There is only limited high-ranked literature about I4.0 and regarding technologies
available. Due to that the authors mainly used non-scientific publications for the
review. The authors also focused only on technologies within the context of I4.0
which may limit the presented technology selection. Also the categorization de-
pends highly on the qualification and number of experts included in focus groups.
The selected experts are mainly experienced researchers and very familiar with
I4.0 technologies as well as with the Process Chain Model. One way to improve
the proposed framework could be to set-up a broad survey or workshops with
practitioners and to assign technologies not only for the material but also for the
information and financial flow. To give direction within SC optimization activities,
the Process Chain Model and its elements have been used due to its focus on
logistic and SC processes. The suitability of other frameworks like Event-driven
Process Chain (EPC) or Supply Chain Operation Reference-Models (SCOR) should
be tested in further research. After creating a valid Technology-Framework, it
should be applied in practice.
However, the framework serves as a basis for future SC process transformation.

By grouping the technologies around approved categories – originally defined
for describing process chain elements – it speeds up the identification and se-
lection of appropriate technologies. The framework is especially designed for
295
practitioners to speed up future transformation processes towards a digitalized

SC.
References
Arnold, D., A. Kuhn, K. Furmans, H. Isermann, and H. Tempelmeier, eds. (2008). Handbuch Logistik.
3., neu bearbeitete Aufl. VDI-Buch. Berlin: Springer.
Bartodziej, C. J. (2017). The Concept Industry 4.0: An Empirical Analysis of Technologies and Applica-
tions in Production Logistics. Wiesbaden: Springer Gabler.
Bauer, W., S. Schlund, D. Marrenbach, and O. Ganschar (2014). Industrie 4.0 - Volkswirtschaftliches
Potenzial für Deutschland. Ed. by BITKOM and Fraunhofer-Institut für Arbeitswirtschaft und
Organisation IAO.
Bauernhansl, T. (2014). “Die Vierte Industrielle Revolution – Der Weg in ein wertschaffendes Pro-
duktionsparadigma”. In: Industrie 4.0 in Produktion, Automatisierung und Logistik. Ed. by T.
Bauernhansl, M. ten Hompel, and B. Vogel-Heuser. Wiesbaden: Springer Fachmedien Wies-
baden, pp. 5–35.
Bauernhansl, T., M. ten Hompel, and B. Vogel-Heuser, eds. (2014). Industrie 4.0 in Produktion,
Automatisierung und Logistik. Wiesbaden: Springer Fachmedien Wiesbaden.
Bauernhansl, T., J. Krüger, G. Reinhart, and G. Schuh (2016). WGP-Standpunkt Industrie 4.0. Ed. by
Wissenschaftliche Gesellschaft für Produktionstechnik WGP e.V. Darmstadt.
Bechtold, J., A. Kern, C. Lauenstein, and L. Bernhofer (2014). Industrie 4.0 - EIne Einschätzung von
Capgemini Consulting: Der Blick über den Hype hinaus. Ed. by Capgemini Consulting.
Bienzeisler, B., A. Schletz, and A.-K. Gahle (2014). Industrie 4.0 Ready Services Technologietrends
2020: Ergebnisse einer Kurzbefragung auf der Messe MAINTAIN 2014. Ed. by Fraunhofer-Institut
für Arbeitswissenschaft und Organisation IAO. Stuttgart.
Bischoff, J. (2015). Erschließen der Potenziale der Anwendung von Industrie 4.0 im Mittelstand.
Ed. by agiplan GmbH 2015.
Bloching, B., P. Leutiger, T. Oltmanns, C. Rossbach, T. Schlick, G. Remane, P. Quick, and O. Shafranyuk
(2015). Die digitale Transformation der Industrie: Was sie bedeutet. Wer gewinnt. Was jetzt zu
tun ist. Ed. by Roland Berger Strategy Consultants GmbH and Bundesverband der Deutschen
Industrie e.V.
Buchanan, D. and A. Bryman, eds. (2009). The Sage Handbook of Organizational Research Methods.
London: Sage Publications.
Bui, T. X. and R. H. Sprague, eds. (2016). Proceedings of the 49th Annual Hawaii International
Conference on System Sciences: 5-8 January 2016, Kauai, Hawaii. Piscataway, NJ: IEEE.
Dais, S. (2014). “Industrie 4.0 – Anstoß, Vision, Vorgehen”. In: Industrie 4.0 in Produktion, Automa-
tisierung und Logistik. Ed. by T. Bauernhansl, M. ten Hompel, and B. Vogel-Heuser. Wiesbaden:
Springer Fachmedien Wiesbaden, pp. 625–634.
Denyer, D. and D. Tranfield (2009). “Producing a Systematic Review”. In: The Sage Handbook of Or-
ganizational Research Methods. Ed. by D. Buchanan and A. Bryman. London: Sage Publications,
pp. 671–689.
296
References
Gausemeier, J., R. Dumitrescu, J. Jasperneite, A. Kühn, and H. Trsek (2016). Auf dem Weg zu
Industrie 4.0: Lösungen aus dem Spitzencluster it’s OWL. Ed. by it’s OWL Clustermanagement
GmbH. Paderborn.
Geissbauer, R., J. Vedso, and S. Schrauf (2016). Industry 4.0: Building the digital enterprise. Ed. by
PwC.
Glas, A. H. and F. C. Kleemann (2016). “The Impact of Industry 4.0 on Procurement and Supply
Management: A Conceptual and Qualitative Analysis”. In: International Journal of Business and
Management Invention 5.6, pp. 55–66.
Graef, N. (2016). “Industrie 4.0-Gesamtkonzept: Zusammenspiel von intelligenten Infrastrukturen,
Paradigmen und technologischen Komponenten”. In: Einführung und umsetzung von Industrie
4.0. Ed. by A. Roth. Gabler, pp. 73–82.
Hans-Christian, P., Y. Burak, and K. Tamer (2015). “The Impact of Industry 4.0 on the Supply Chain”.
In: Innovations and Strategies for Logistics and Supply Chains. Ed. by W. Kersten, T. Blecker, and
C. M. Ringle. Vol. 20. Proceedings of the Hamburg International Conference of Logistics (HICL).
Berlin: epubli GmbH, pp. 31–58.
Hausladen, I. (2016). IT-gestützte Logistik: Systeme - Prozesse - Anwendungen. 3. Aufl. 2016.
Hermann, M., T. Pentek, and B. Otto (2016). “Design Principles for Industrie 4.0 Scenarios”. In:
Proceedings of the 49th Annual Hawaii International Conference on System Sciences. Ed. by T. X.
Bui and R. H. Sprague. Piscataway, NJ: IEEE, pp. 3928–3937.
Ho, W., T. Zheng, H. Yildiz, and S. Talluri (2015). “Supply chain risk management: a literature
review”. In: International Journal of Production Research 53.16, pp. 5031–5069.
Horvath & Partners (2017). Technologische Trends in der Industrie 4.0. Ed. by Horvath & Partners.
Huber, W. (2016). Industrie 4.0 in der Automobilproduktion: Ein Praxisbuch. Wiesbaden: Springer
Viehweg.
Industrie 4.0 - Whitepaper FuE-Themen (2015). http://www.acatech.de/fileadmin/user_upload/
Baumstruktur_nach_Website/Acatech/root/de/Aktuelles___Presse/Presseinfos___News/ab_
2014/Whitepaper_Industrie_4.0.pdf.
Jäger, J. M., D. Görzig, D. Paulus-Rohmer, H. Schatton, S. Baku, M. Weskamp, and D. Lucke (2015).
Industrie 4.0 - Chancen und Perspektiven für Unternehmen der Metropolregion Rhein-Neckar.
Ed. by Industrie- und Handelskammern Rhein-Neckar, Pfalz und Darmstadt Rhein Main Neckar.
Kersten, W., T. Blecker, and C. M. Ringle, eds. (2015). Innovations and Strategies for Logistics and
Supply Chains: Technologies, Business Models and Risk Management. 1st ed. Vol. 20. Proceedings
of the Hamburg International Conference of Logistics (HICL). Berlin: epubli GmbH.
Kersten, W., M. Seiter, B. von See, N. Hackius, and T. Maurer (2017). Trends und Strategien in Logistik
und Supply Chain Management: Chancen der digitalen Transformation. Ed. by Bundesvereini-
gung Logistik. Bremen.
Kersten, W., M. Seiter, B. von See, N. Hackius, C. Rosentritt, C. Böhle, G. Reich, T. Maurer, and R.
Sauter (2016). “Trends und Strategien in Supply Chain Management und Logistik - Chancen
der digitalen Transformation”. In: Den Wandel gestalten. Ed. by T. Wimmer and C. Grotemeier.
Schriftenreihe Wirtschaft & Logistik. Hamburg: DVV Media Group GmbH, pp. 344–361.
Kilubi, I. and H. D. Haasis (2016). “Supply chain risk management research: Avenues for further
studies”. In: International Journal of Supply Chain and Operations Resilience 2.1, p. 51.
Kitzinger, J. (1995). “Introducing focus groups”. In: BMJ 311, pp. 299–302.
Kuhn, A. (1995). Prozessketten in der Logistik: Entwicklungstrends und Umsetzungsstrategien. Dort-
mund: Verlag Praxiswissen.
297
Kuhn, A. (2008). “Das Modellierungsparadigma”. In: Handbuch Logistik. Ed. by A. Dieter, K. Axel,
F. Kai, I. Heinz, and T. H. VDI-Buch. Berlin: Springer, pp. 219–224.
Obermaier, R., ed. (2016). Industrie 4.0 als unternehmerische Gestaltungsaufgabe: Betriebswirtschaftliche,
technische und rechtliche Herausforderungen. Wiesbaden and Wiesbaden: Springer Gabler.
Roth, A., ed. (2016a). Einführung und umsetzung von Industrie 4.0: Grundlagen, Vorgehensmodell
und Use Cases in der Praxis. Gabler.
Roth, A. (2016b). “Industrie 4.0 – Hype oder Revolution?” In: Einführung und umsetzung von Indus-
trie 4.0. Ed. by A. Roth. Gabler, pp. 1–15.
Rozados, I. V. and B. Tjahjono (2014). “Big Data Analytics in Supply Chain Management: Trends
and Related Research”. In: 6th International Conference on Operations and Supply Chain Man-
agement.
Rüßmann, M., Lorenz, Markus, Gerbert, Philipp, M. Waldner, J. Justus, P. Engel, and M. Harnisch
(2015). Industry 4.0: The Future of Productivity and Growth in Manufacturing Industries. Ed. by
The Boston Consulting Group.
Saunders, M., P. Lewis, and A. Thornhill (2009). Research methods for business students. fifth edition.
Harlow: Pearson Education Limited.
Schlaepfer, R. C., M. Koch, and P. Merkofer (2015). Industry 4.0: Challenges and solutions for the
digital transformation and use of exponential technologies. Ed. by Deloitte AG.
Schlick, J., P. Stephan, M. Loskyll, and D. Lappe (2014). “Industrie 4.0 in der praktischen Anwen-
dung”. In: Industrie 4.0 in Produktion, Automatisierung und Logistik. Ed. by T. Bauernhansl, M.
ten Hompel, and B. Vogel-Heuser. Wiesbaden: Springer Fachmedien Wiesbaden, pp. 57–84.
Schrauf, S. and P. Berttram (2016). Industry 4.0 - How digitization makes the supply chain more
efficient, agile, and customer-focused. Ed. by PwC Strategy&.
Schwab, K. (2016). The Fourth Industrial Revolution. Penguin Random House UK.
Seiter, M., C. Bayerle, S. Berlin, U. David, M. Rusch, and O. Treusch (2016). Roadmap Industrie 4.0:
Ihr Weg zur Erfolgreichen Umsetzung von Industrie 4.0. Ulm: tredition.
Siepmann, D. (2016a). “Industrie 4.0 – Fünf zentrale Paradigmen”. In: Einführung und umsetzung
von Industrie 4.0. Ed. by A. Roth. Gabler, pp. 35–46.
Siepmann, D. (2016b). “Industrie 4.0 – Struktur und Historie”. In: Einführung und umsetzung von
Industrie 4.0. Ed. by A. Roth. Gabler, pp. 17–34.
Stebbins, R. A. (2001). Exploratory Research in the Social Sciences. Thousand Oaks, CA: Sage.
ten Hompel, M. and M. Henke (2017). “Logistik 4.0 – Ein Ausblick auf die Planung und das Man-
agement der zukünftigen Logistik vor dem Hintergrund der vierten industriellen Revolution”.
In: Handbuch Industrie 4.0 Bd.4. Ed. by B. Vogel-Heuser, T. Bauernhansl, and M. ten Hompel.
Springer Reference Technik. Berlin, Heidelberg and s.l.: Springer Berlin Heidelberg, pp. 249–
259.
Allgemeine Grundlagen. 2nd ed. Springer Reference Technik. Berlin, Heidelberg and s.l.: Springer
Berlin Heidelberg.
Wan, J., H. Cai, and K. Zhou (2014). “Industrie 4.0: Enabling technologies”. In: 2014 International
Conference on Intelligent Computing and Internet of Things (ICIT), pp. 135–140.
Wee, D., R. Kelly, J. Cattell, and M. Breunig (2015). Industry 4.0: How to navigate digitization of the
manufacturing sector. Ed. by McKinsey & Company.
Wehberg, G. G. (2015). Logistik 4.0: Komplexität managen in Theorie und Praxis. Berlin, Heidelberg:
Springer Gabler.
298
References
Wimmer, T. and C. Grotemeier, eds. (2016). Den Wandel gestalten: 33. Deutscher Logistik-Kongress :
Tagungsband. Schriftenreihe Wirtschaft & Logistik. Hamburg: DVV Media Group GmbH.
Wischmann, S., L. Wangler, and A. Botthof (2015). Industrie 4.0 - Volks- und betriebswirtschaftliche
Faktoren für den Standort Deutschland: Eine Studie im Rahmen der Begleitforschung zum Tech-
nologieprogramm AUTONOMIK für Industrie 4.0. Ed. by Bundesministerium für Wirtschaft und
Energie. Berlin.
Wollschlaeger, M., T. Sauter, and J. Jasperneite (2017). “The Future of Industrial Communication:
Automation Networks in the Era of the Internet of Things and Industry 4.0”. In: IEEE Industrial
Electronics Magazine 11.1, pp. 17–27.
Zhang, Y., G. Q. Huang, S. Sun, and T. Yang (2014). “Multi-agent based real-time production schedul-
ing method for radio frequency identification enabled ubiquitous shopfloor environment”. In:
Computers & Industrial Engineering 76, pp. 89–97.
Zillmann, M. and B. Appel (2016). Keine Industrie 4.0 ohne Digitalisierung der Supply Chain: Intelli-
gente Logistikdienstleistungen für die Fertigungsindustrie. Ed. by Lünendonk GmbH.
299
Cooperation Strategies among SMEs for
Implementing Industry 4.0
Julian Marius Müller1 , Lukas Maier1 , Johannes Veile1 , Kai-Ingo Voigt1
1 – University of Erlangen-Nuremberg
Industry 4.0 is expected to bring several conversions for industrial value creation,
encompassing entire value-added networks. Small and medium-sized enterprises
(SMEs), which play an important role for both the German as well as the Euro-
pean economy, struggle to integrate the concept of Industry 4.0 within their value
creation. However, due to the high importance of SMEs for industrial value cre-
ation networks, their integration is essential to successfully establish Industry
4.0 across value chains. Several SMEs struggle to obtain the resources required
for equipment and machinery or do not possess the required market shares or
market access to establish new business models. Large enterprises are often seen
as too powerful to be a partner for a SME. Nevertheless, cooperation strategies
among SMEs may present a viable alternative to successfully implement Industry
4.0 across the value chain. In this context, literature lacks of a well-founded in-
vestigation of this topic. Therefore, this study attempts to close the present the
present research gap. Due to the exploratory nature of the underlying topic, we
conduct a multiple case study with 68 SMEs in Germany. This paper comes up
with cooperation strategies and presents the interviewees’ answers regarding po-
tentials and challenges of common technology purchasing as well as for common
business models. Subsequently, we present implications for both research and
practice.
Keywords: Industry 4.0; Small and Medium-sized enterprises;

Cooperation strategies; Industrial Internet of Things
301
Cooperation Strategies among SMEs for Implementing Industry 4.0
1 Introduction
Industry 4.0, which is internationally known as the Industrial Internet of Things,

aims to establish an intelligent, self-regulating and interconnected industrial value
creation (Kang et al., 2016) ensuring future competitiveness of the manufacturing
industry (Kagermann et al., 2013; Lasi et al., 2014). Given the topic’s importance
and actuality, research on Industry 4.0 focuses rather on technological develop-
ments related to cyber-physical systems (Brettel et al., 2014; Liao et al., 2017) than
on their organizational implementation (Arnold et al., 2016; Ehret & Wirtz, 2017).
Recently, scholars begin to study the value creation implications of Industry 4.0
(Kans & Ingwald, 2016; Kowalkowski et al., 2013; Rennung et al., 2016). However,
according to Kowalkowsi et al. (2013), researchers investigating this topic mainly
focus on large companies (Radziwon et al., 2014).
Large organizations constitute a minority, as SMEs play an important role in
the overall network of industrial value creation. In the European Union, SMEs
represent over 99% of all companies. Additionally, SMEs employ between 50 and
70% of all European full time equivalents and generate a gross value added share
that encompasses about 50% of the European economy (Airaksinen et al., 2015;
Bundesministerium für Wirtschaft und Energie, 2014).
With this paper we attempt to further examine SMEs in order to fully understand
the mechanisms and implications of Industry 4.0 implementation. The legiti-
macy for our research lies in a lack of fundamental knowledge about Industry 4.0.
along with SMEs’ impact on value creation and their importance for the overall
economy.
Existing literature shows that SMEs and large organizations fundamentally differ
in terms of size, processes, and availability of resources (Ihlau et al., 2013). There-
fore, SMEs require different strategies to successfully implement Industry 4.0 in
comparison to large companies. In this context, an important question is how
the characteristics of a company, such as size and resource base affect its ability
to implement new technologies. For instance, the adoption of ERP systems, a
technological precursor to Industry 4.0, is differently approached in SMEs and was
found to be more challenging in SMEs than in large companies (Buonanno et al.,
2005). Correspondingly, the implementation of Industry 4.0 tools may be more
difficult in SMEs, as such companies often have lower digitization levels, caused
by their operation in niche markets (Knight, 2000), their smaller production series,
as well as due to their smaller production series as well as their limited access to
resources and knowledge. Particularly, SMEs lack resources and knowledge that
302
is, however, critical for the successful implementation of Industry 4.0. Existing lit-
erature shows that SMEs often cooperate with other companies to achieve better
access to financial as well as personnel resources (Ihlau et al., 2013). Thus, using
cooperation strategies appears to be a suitable approach for SMEs to successfully
implement Industry 4.0 within an organization. However, literature provides no
implications on how such cooperation strategies among SMEs should be designed
and how they actually influence the implementation of Industry 4.0. That is why
the aim of this study is to investigate which specific benefits as well as challenges
exist regarding cooperation strategies among SMEs for implementing Industry
4.0.
2.1 Industry 4.0
The term Industry 4.0 encompasses the expectations of politics and corporate
practice that industrial manufacturing heads towards the fourth Industrial Revo-
lution. The previous three Industrial Revolutions have achieved high productivity
increases, driven by a few, fast spreading general-purpose technologies, such
as mechanization, electricity and IT (Veza et al., 2015). These general-purpose
technologies resulted in strong technical improvements and initiated further com-
plementary developments (Bresnahan & Trajtenberg, 1995). The general-purpose
technologies for Industry 4.0 are cyber-physical systems, whose technological
infrastructure are based on the concept of the Internet of Things (Kagermann
et al., 2013; Lasi et al., 2014; Xu, 2012). Cyber-physical systems are intended to
establish an interconnection between the physical world and the cyber-space
(He & Xu, 2015; Lee et al., 2015; Ren et al., 2013). Cyber-physical systems hereby
offer mechanisms for human-to-human, human-to-object and object-to-object
interactions along the entire value-added chain (Wan, 2011). Especially the task
of integrating humans into this concept is perceived to be an enormous chal-
lenge as it faces employees’ resistance (Frazzon et al., 2013; Gorecky et al., 2014;
Hirsch-Kreinsen, 2016; Schuh et al., 2014). Humans’ integration into industrial
manufacturing leads to cyber-physical production systems (Schlechtendahl et
al., 2015). Cyber-physical production systems enable several data-based services,
such as predictive condition monitoring or balancing and reducing energy con-
sumption within production (Lee et al., 2013; Shin et al., 2014; Tao et al., 2011).
303
Manufacturers place high expectations on cyber-physical production systems

because they enable machinery safety, real-time control, self-organization and
self-maintenance, autonomous navigation through production facilities and error
predictability(Meyer et al., 2011; Monostori, 2014) along the entire lifecycle of
machinery and products (Lennartson et al., 2010). Aside from cyber-physical
production systems, Industry 4.0 is driven by technological developments such
as service-oriented architectures (Guinard et al., 2010; Mikusz, 2016; Raja et al.,
2013; Vogel-Heuser et al., 2015), which enable the creation of new services and
product-service bundles (Ehret & Wirtz, 2017).
Those developments in sum result in the concept of smart production, also termed
smart manufacturing (Davis et al., 2012; Feeney & Weiss, 2014; Radziwon et al.,
2014; Wang et al., 2016; Zuehlke, 2010). Smart production has been discussed to
be a core element of smart factories (Radziwon et al., 2014; Zhang et al., 2014).
The latter use flexible and adaptive production processes to dynamically solve the
problems of complex economic environments. Smart Production is characterized
by manufacturing of smart, personalized products as well as high levels of collab-
oration through production networks, among several enterprises (Kagermann et
al., 2013; Lasi et al., 2014; Veza et al., 2015; Xu et al., 2014).
Besides the German initiative Industry 4.0, the EU has initiated a public-private
partnership under the title ”Factories of the Future” to achieve sustainable and
competitive production in the future (European Commission, 2016). In the US,
similar ideas are encouraged through the Industrial Internet Consortium with
founding members such as AT&T, CISCO, GE, IBM and INTEL (Pike, 2014). The
”Internet Plus initiative” in China integrates current technological developments
such as cloud computing and big data enabling state-of-the-art manufacturing
(Keqiang, 2015), while South Korea has introduced the ”Manufacturing Innovation
3.0” (Kang et al., 2016).
2.2 Small and medium-sized enterprises
The term small and medium-sized entreprises refers to companies with less than
50 million Euro in sales and less than 500 employees regardless of their indus-
try (Bundesministerium für Wirtschaft und Energie, 2013; Günterberg & Wolter,
2002). In our paper we investigate SMEs and how to implement Industry 4.0 with
cooperation strategies because of several reasons:
304
3 Method and research design
First, potentials of Industry 4.0 can primarily be expected because of the horizontal
and vertical network of the value chain. In the German industry, SMEs represent
an essential part, as they represent 99,6% of all enterprises generating more than
50% of the GDP. In turn, integrating SMEs is perceived to be key to the success of
Industry 4.0.
Second, existing studies show that SMEs’ specific challenges differ from those
of large companies. Therefore, SMEs require solutions tailored to meet their
specific challenges. Management in SMEs already recognizes the importance of of
Industry 4.0. However, research mainly focuses rather on large enterprises than
on SMEs (Bischoff et al., 2015).
Third, the upper management of SMEs in contrast to that of large companies

seems to be able to keep track of the whole enterprise. Interviewing them may
reveal much information about our research topic. Further, their hierarchical
position allow them to give holistic statements why SMEs qualify for examination.
SMEs’ managers may assess interfaces and provide both an external as well as an
internal perspective (Ihlau et al., 2013).
3 Method and research design
3.1 Multiple case study
Our study’s goal is to investigate potentials and challenges for cooperation strate-
gies among SMEs for implementing Industry 4.0 and to provide an integrative,
systematic, and comprehensive understanding about this topic.
In order to address this goal, we use a qualitative empirical methodology and

conduct a multiple case study. We chose this methodology because it allows us to
investigate the topic in a wider context, to gain a complete and holistic view, and
to derive valid and generalizable results (Bryman & Bell, 2011; Gibbert et al., 2008).
Asthis topic is novel, evolving, and a contemporary phenomenon, a case study
design is the best method that can be used, which is especially true for research
in the setting of operations or IT (Dubé & Paré, 2003; Eisenhardt, 1989; Eisenhardt
& Graebner, 2007; Voss et al., 2012; Yin, 2009). Instead of relying on a single case,
we use multiple cases to increase both the robustness and the generalizability of
our results (Eisenhardt & Graebner, 2007).
305
3.2 Data collection
We use semi-structured expert interviews with qualified and experienced man-

agers as main source following common handling in qualitative research (Mason,
2002). Interviews of this manner allow collecting data structurally while keeping
the openness that is necessary to gather all important information (Yin, 2009).
We interviewed 68 German managers of SMEs between May and July 2016 . 48

of the 68 SMEs have 100 to 500 employees, whereas 20 of the SMEs have up to
100 employees. We chose the industry sectors machine and plant engineering,
electrical engineering and automotive. Our choice is based on the facts that these
industries all contribute a great deal to the German Cross Domestic Product and
the chosen mix well represents the industry landscape in Germany. Furthermore,
the chosen industries are considered to be the ones that are most affected by
and to benefit the most from Industry 4.0 (Kagermann et al., 2013). We chose
Germany because of its representative character for an industrial nation, its lead-
ing economical position within the European Union and its high achievements in
technological and digital development. Using a wide variety of empirical material
helps to counteract negative effects of sample bias in our research (Yin, 2009).
The interviewed manager steam from middle and top management positions
respectively. Those managers know the most about their firm’s cooperation strate-
gies for implementing Industry 4.0, which makes them the best suitable interview
partners in our research. The interviews last between 20 and 60 minutes. In order
to avoid any language barriers, we conduct the interviews in German, the mother
tongue of both the interviewees and interviewers. For confidentiality reasons,
we anonymize detailed case data. Corresponding to the exploratory nature of
this study, the development of the interview guide was inspired by literature but
followed the principle of openness and flexibility to allow unexpected and novel
topics to emerge (Kasabov, 2015).
The interview guideline consists of three parts: First of all, the interviewed man-
agers are questioned about their professional background and their areas of
responsibility. In doing so, we ensure that the interview partners are suitable for
the purpose of the study. Second, we ask the interviewees questions concerning
potentials and challenges of cooperation strategies in technology purchasing.
Third, we question the interview partners concerning potentials and challenges
in jointly run new business models among SMEs . We introduce this differentia-
tion in order to ensure the high importance of new, partner or platform based
306
4 Results
business models within the concept of Industry 4.0 (Kagermann et al., 2013; Wu
et al., 2013).
3.3 Data analysis and reliability of the study
The 68 audio-recorded interviews were transcribed into text material before ana-
lyzing them. We study the transcription applying a qualitative content analysis to
identify and interpret common patterns, themes, and categories (Huber & Power,
1985; Miles & Huberman, 1994). Applying an inductive coding procedure (Char-
maz, 2006; Gioia, Corley, & Hamilton, 2013; Krippendorf, 2013) helpes us not to
restrict our results but generate novel knowledge (Graebner & Eisenhardt, 2004).
We conduct the entire coding process as a team to achieve the best interpreta-
tions and most profound understanding (Weston et al., 2001). An application of a
frequency analysis following Holsti (1969) simplifies the identification of the most
important potentials and challenges for cooperation strategies among SMEs. We
enhance the validity and robustness of our results by applying triangulation of
secondary data from annual reports and company websites to verify the inter-
viewees’ statements (Eisenhardt & Graebner, 2007; Yin, 2009). Furthermore, we
assure the respondents full anonymity and confidentiality addressing potential
key informant bias. The multiple case study approach supported the mitigation
of the negative effects of observer bias (Eisenhardt & Graebner, 2007).
4 Results
4.1 Common technology purchasing regarding Industry 4.0
Table 1 shows the potentials for common technology purchasing regarding Indus-
try 4.0, sorted by their frequency of naming. Our interviews indicate the reduction
of financial commitment to be the most important reason for common technol-
ogy purchasing. Furthermore, interviewees mention the distribution of risks, the
exchange of ideas, and strengthened partnership as further potentials.
Table 2 shows the challenges for common technology purchasing regarding In-
dustry 4.0, sorted by their frequency of naming. In this context, we found trust
between partners, loss of confidential information, and coordination efforts to be
307
the most common challenges for SMEs. Further challenges for common technol-
ogy purchasing are preferred autonomy, increasing dependencies, only near-term
benefits, no suitable partners, the lack of legal conditions, and reluctant behavior
towards Industry 4.0.
4.2 Common business models for Industry 4.0
Table 3 shows the benefits of common business models among SMEs for Indus-
try 4.0. We find three potentials for common business models: 24 participants
mentioned the optimum usage of virtual interconnection, 20 mentioned the de-
crease of existing challenges through Industry 4.0, and three mentioned cost
reductions.
Finally, Table 4 provides an overview of the challenges of common business
models among SMEs for Industry 4.0. Overall, our results indicate nine challenges:
Business model innovations are no a core competence, business model is not
understood, legal uncertainty, lack of resources, no customer demand, preferred
autonomy, coordination efforts, no risk diversification, and loss of flexibility.
Table 1: Potentials for common technology purchasing
Category Frequency Explanation

Reduction of financial 33 Through the generation of com-
commitment pound effects and bargaining
power, SMEs can reduce required
financial capital.
Distribution of risks 19 Commonly purchasing dis-
tributes the risks of misinvest-
ments among partners.
Exchange of ideas 11 Common purchasing leads SMEs
to exchange ideas more openly
and generate new ideas.
Strengthened partner- 4 Through close financial ties ,
ships SMEs are able to establish new or
strengthened partnerships.
308
4 Results
Table 2: Challenges for common technology purchasing

Trust between partners 25 Lacking trust between partners,
e.g. opportunistic or dishonest
behavior, hinders the implemen-
tation.
Loss of confidential infor- 24 In their purchasing strategy,
mation SMEs incorporate confidential
and strategic information that
could be made public.
Coordination efforts 22 High coordination efforts oppose
the compound effects generated
by common purchasing.
Preferred autonomy 9 SMEs, especially owner-run com-
panies, prefer autonomy in their
purchasing activities.
Increasing dependencies 7 Dependencies between partners,
which cannot be reversed are
feared.
Only purchasing benefits 5 Common purchasing is only seen
as beneficial for purchasing, not
for further operations.
No suitable partners 5 The required partners are not
willing to cooperate or no ade-
quate partners exist.
Lacking legal conditions 2 The legal conditions are not fully
determined for data-based pur-
chasing within Industry 4.0.
Reluctant behavior 2 The topic of Industry 4.0 is still
towards Industry 4.0 unfamiliar to many SMEs.
309
Table 3: Potentials for common business models for Industry 4.0

Optimum usage of virtual 24 Common business models
interconnection among SMEs become viable and
easy to coordinate.
Decrease of existing chal- 20 Establishing virtual interconnec-
lenges for SMEs. tion, SMEs may gain the bargain-
ing power and market share of
larger enterprises.
Cost reductions 3 Compound effects may lead to
cost reductions for all partners.
310
4 Results
Table 4: Challenges for common business models for Industry 4.0

Business model innova- 22 The development of business
tions are not a core com- models is not seen as a core com-
petence petence of SMEs what may hin-
der the development of new busi-
ness models within Industry 4.0.
Business model is not un- 22 Up to now, SMEs have not fully
derstood understood the concept of busi-
ness models, and thus they are
mainly process or product ori-
ented.
Legal uncertainty 17 Data-based business models
come into question because of
legal uncertainty regarding data
theft and data property.
Lack of resources 7 As far as resources are concerned,
SMEs are not seen as capable of
establishing new business mod-
els.
No customer demand 7 To date, customers do not de-
mand new business models.
Preferred autonomy 5 SMEs prefer to be independent
and refrain from establishing new
business models with partners.
Coordination efforts 4 The Coordination between part-
ners is perceived to be both time
and cost-intensive.
No risk diversification 3 SMEs become increasingly com-
mitted within a business model,
so they are not able to diversify
which in turn increases risks.
Loss of flexibility 2 SMEs might lose their core
strengths such as providing
individual and fast solutions,
tailored to customer demands.
311
5 Discussion
This chapter discusses cooperation strategies to overcome SMEs’ lack of resources

and knowledge (Ihlau et al., 2013) and thus help to implement Industry 4.0 in their
organization. In this context, commonly purchasing technologies as well as jointly
developing business models are perceived to be adequate strategies, first of all
examining potentials and challenges that come along with them. Keeping the
results of our research in mind, the potentials and challenges can be discussed to
derive implications how such cooperation strategies shall be designed to foster
the implementation of Industry 4.0.
The first strategy investigated is commonly purchasing new technology among

SMEs. About 37% of the interviewees mentioned a lack of trust between partners
and 35% named a potential loss of confidential information, making those aspects
the biggest challenges of common technology purchasing. Therefore, special
attention should be given to these aspects when establishing a cooperation to
make it a success. A lack of trust may stem from perceiving a partner’s high
negotiation power and low opportunity costs to withdraw from a cooperation.
In order to establish a well-based cooperation, trust is the key to success. It is
without saying that communicating trustworthily and acting reliably is the basis of
a trustful cooperation. Bearing in mind not only the financial benefit, but also the
importance of credibility towards partners, SMEs can act more collaboratively in
negotiations. In addition, increasing its own opportunity costs by higher financial
commitments on the partnership may increase the credibility of a partner.
Further, 33 % of the interviewees mention high coordination efforts in maintain-

ing a cooperation to commonly purchase technology. This result indicates the
importance of simplifying any form of cooperation to benefit comprehensively
while suppressing the costs. Therefore, special attention should be drawn to
interfaces in order to avoid any non-value adding processes. Automated com-
munication and interactions, as proposed by Industry 4.0 in general, may lead
to decreasing transaction costs, which in turn decreases coordination efforts.
Additionally, resources may be provided to lower coordination costs for SME. Our
results indicate that SMEs tend to prefer autonomy (13%) and fear increasing
dependencies (10%). When setting up a cooperation, it is important to work to-
gether as closely as necessary for an adequate cooperation while respecting each
partner’s freedom. SME can be integrated into decisions, which may give them a
sense of freedom in their decision-making process. A key aspect in this context is
to create a partner’s awareness for the importance of the cooperation, providing
312
5 Discussion
reasons for both the required efforts and the dependency and respecting the
individual, often owner-based, nature of SMEs.
As a further insight, this study reveals reducing financial commitment as the most
important potential of a cooperation in terms of technology purchasing. 49% of
interviewees name this aspect to be vital. Another potential for common tech-
nology purchasing is the distribution of risks, mentioned by 28% of respondents.
Given these aspects’ importance, the emphasis should be placed on the financial
benefits of the cooperation and its visibility to all partners. In this context, it
seems reasonable to distribute financial benefits among partners in a fair way
so that all partners involved benefit from and value the cooperation. Apart from
financial benefits, an equal distribution may help to increase trust, as explained
in the section before.
Lastly, the exchange of ideas is mentioned to be an important potential when
cooperating among SMEs (16%) to purchase technology. Apart from the financial
gains, our research reveals that one should nevertheless emphasize not primarily
financial aspects. Providing the partner with knowledge about processes and
sharing best-practice cases may offer further incentives to enter into a cooperation.
Reciprocal exchange of knowledge does not only provide benefits for one partner,
but may lead to benefits for all partners in the cooperation
Another strategy analyzed is establishing common new business models between
SMEs.
In our study, we show that SMEs tend to see new business models with particular
caution and there are several reasons to explain this. 33% of our interviewees
state that it is not SMEs’ core competence to develop business models and there-
fore they intend to refrain from doing so. Another reason is that SMEs may not
fully understand the concept of business models. The results therefore reveal
interesting insights how to cooperate among SMEs regarding Industry 4.0. First,
the key to success is to properly share information. Partners with little knowledge
about business models may gain the relevant knowledge just after entering a
cooperation. Sharing information in a cooperation may increase total knowledge
about business models. Second, apart from knowledge sharing, a cooperation
allows to share resources such as working time and human capabilities. One
company may not be able to develop new business models and thus have other
core competences, but sharing resources in a cooperation provides the cooperat-
ing parties with the basis to do so. Last, a cooperation may help to increase the
sense of urgency and the perception of business model’s strategic relevance due
to group effects.
313
Legal uncertainty is an important challenge mentioned by 25% of all interviewees

as issues concerning data security and data property are present. Legal uncer-
tainty indeed is an issue that is widely discussed in research and public opinion
without a silver bullet to solve it. However, entering a cooperation bundles re-
sources, that can be used to invest in data security, to give one example. Further,
data may be stored in a decentral way, reducing the vulnerability and the proba-
bility for data loss. Last, a cooperation has more influence to ask authorities and
legislation to provide SMEs with an environment that guarantees the necessary
ecosystem for Industry 4.0
Our study furthermore reveals that SMEs fear, even when working together in a
cooperation, to not have enough resources to create new business models (10%).
The lack of resources might be an issue but working together summarizes the
resource base of several companies or at least increases the possibilities to get
further resources, such as external capital. Cross-subsidization with financial ben-
efits for instance from commonly purchasing technology may help to overcome
this issue.
As a potential, 35% of our interviewees name the optimum usage of virtual in-
terconnection. 29% of the interviewees believe another potential of common
business models for Industry 4.0 is the decrease of existing challenges for SMEs.
In combination with cost reductions, named by 3% in our sample, especially the
potentials of cost reductions and new value propositions though new business
models should be fostered in a potential cooperation among SMEs to overcome
the challenges named and make benefits more visible and understandable. In a
second step, further benefits of common business models as its understanding
within cooperating SMEs rises, could unfold.
6 Conclusion
The aim of this study was to examine cooperation strategies among SMEs for
implementing Industry 4.0. We find challenges for SMEs and benefits for both
technology purchasing as well as developing business models. We explain this
as our interviewed SMEs are at a rather early stage in the implementation of
Industry 4.0. At this stage, SMEs may not consider cooperation strategies as an
important tool to support their future business. This becomes especially apparent
for the challenges of joint business model development among SMEs regarding
Industry 4.0, which is not seen as a core competence of SMEs as well as the
314
References
concept of business models is not understood entirely by 22 out of 68 SMEs in our

sample. Due to this, relatively early stage of implementation regarding Industry
4.0, where this understanding can still be generated and increased, our study
presents valuable contributions for the current stage.
However, this paper is limited to a short-term perspective of cooperation strate-
gies regarding Industry 4.0 among SMEs. Moreover, further research needs to be
conducted to generalize our results in further cultural context. We recommend to
further investigate cooperation strategies among SMEs within different industry
sectors. A comparison with industries at a more mature stage of implementa-
tion Industry 4.0 for example the IT or software industry related to industrial
production. This allows uncovering industry differences and deriving explanatory
approaches.
For corporate practice, we recommend to develop new business models in context
of Industry 4.0 working together in cooperation. Also, we advise to consider
reorganize corporate culture such as openness to develop business models in
cooperation. Further, we suggest policy makers to provide corporate practice
with legal conditions such as data standards and data property supporting efforts
to work in cooperation.
References
Airaksinen, A., H. Luomaranta, P. Alajaeaeskoe, and A. Roodhuijzen (2015). “Statistics on small and
medium-sized enterprises: Dependent and independent SMEs and large enterprises. [online]
eurostat”. In:
Arnold, C., D. Kiel, and K.-i. Voigt (2016). “How the Industrial Internet of Things Changes Business
Models in Different Manufacturing Industries”. In: International Journal of Innovation Manage-
ment 20, p. 8.
Bischoff, J., C. Taphorn, D. Wolter, N. Braun, M. Fellbaum, S. Goloverov, T. Hegmanns, C. Prasse, M.
Henke, M. ten Hompel, F. Doebbeler, C. Fuss E. Kirsch, B. Maettig, S. Braun, M. Guth, M. Kaspers,
and D. Scheffler (2015). “Erschliessen der Potenziale der Anwendung von Industrie 4.0 0 im
Mittelstand. Studie fuer das Bundesministerium fuer Wirtschaft und Energie”. In:
Bresnahan, T. F. and M. Trajtenberg (1995). “General purpose technologies: Engines of growth?”
In: Journal of Econometrics 65, pp. 83–108.
Brettel, M., N. Friederichsen, M. Keller, and M. Rosenberg (2014). “How Virtualization, Decentraliza-
tion and Network Building Change the Manufacturing Landscape: An Industry 4.0 Perspective”.
In: International Journal of Mechanical, Aerospace, Industrial, Mechatronic and Manufacturing
Engineering 8, pp. 37–44.
Bryman, A. and E. Bell (2011). Business Research Methods. Oxford University Press: Oxford.
315
Bundesministerium fuer Wirtschaft und Energie (2014). “German Mittelstand: Motor der deutschen
Wirtschaft”.
Buonanno, G., P. Faverio, F. Pigni, A. Ravarini, D. Sciuto, and M. Tagliavini (2005). “Factors affecting
ERP system adoption: A comparative analysis between SMEs and large companies”. In: Journal
of Enterprise Information Management 18, pp. 384–426.
Charmaz, K. (2006). Constructing Grounded Theory: A Practical Guide Through Qualitative Analysis.
Los Angeles: Sage.
Commission, E. (2016). “Factories of the Future. [online] European Commission”. In: Available at:
<[Accessed 26, p. 2017.
Davis, J., T. Edgar, J. Porter, J. Bernaden, and M. Sarli (2012). “Smart manufacturing, manufacturing
intelligence and demand-dynamic performance”. In: Computers & Chemical Engineering, ( 47,
pp. 145–156.
Dube, L. and G. Pare (2003). “Rigor in information systems positivist case research: Current prac-
tices, trends, and recommendations”. In: MIS Quarterly 27, pp. 597–636.
Ehret, M. and J. Wirtz (2017). “Unlocking value from Machines: Business models and the industrial
internet of things”. In: Journal of Marketing Management 30, pp. 111–130.
Eisenhardt, K. (1989). “Building theories from case study research”. In: Academy of Management
Review 14, pp. 532–550.
Eisenhardt, K. M. and M. E. Graebner (2007). “Theory building from cases: Opportunities and
Challenges”. In: Academy of Management Journal 50, pp. 25–32.
Feeney, A. B. and B. A. Weiss (2014). “Smart Manufacturing Operations Planning and Control, 1?6”.
In: FY 2014.
Frazzon, E. M., J. Hartmann, T. Makuschewitz, and B. Scholz-Reiter (2013). “Towards Socio-Cyber-
Physical Systems in Production Networks”. In: Procedia CIRP 7, pp. 49–54.
Gibbert, M., W. Ruigrok, and B. Wicki (2008). “What Passes as a Rigorous Case Study?” In: Strategic
Management Journal 29, pp. 1465–1474.
Gioia, D. A., K. G. Corley, and A. L. Hamilton (2013). “Seeking Qualitative Rigor in Inductive Research:
Notes on the Gioia Methodology”. In: Organizational Research Methods 16.1, pp. 15–31.
Gorecky, D., M. Schmitt, M. Loskyll, and D. Zuhlke (2014). “Human-machine-interaction in the
industry 4.0 era”. In: Presented at the 12th IEEE International Conference on Industrial Informatics
(INDIN). Ed. by P. Alegre.
Guenterberg, B. and H. Wolter (2002). Unternehmensgroessenstatistik 2001/2001 Daten und Fakten.
Bonn.
Guinard, D., V. Trifa, S. Karnouskos, P. Spiess, and D. Savio (2010). “Interacting with the SOA-Based
Internet of Things: Discovery, Query, Selection, and On-Demand Provisioning of Web Services”.
In: IEEE Transactions on Services Computing 3, pp. 223–235.
He, W. and L. Xu (2015). “A state-of-the-art survey of cloud manufacturing”. In: International Journal
of Computer Integrated Manufacturing 28, pp. 239–250.
Hirsch-Kreinsen, H. (2016). “Digitization of industrial work: Development paths and prospects”. In:
Journal for Labour Market Research 49, pp. 1–14.
Holsti, O. R. (1969). Content analysis for the social sciences and humanities. Addison-Wesley: Read-
ing.
Huber, G. P. and D. J. Power (1985). “Retrospective reports of strategic-level managers: Guidelines
for increasing their accuracy”. In: Strategic Management Journal 6, pp. 171–180.
Ihlau, S., H. Duscha, and S. Goedecke (2013). Spezifische Merkmale und Besonderheiten von KMU.
Wiesbaden: Springer.
316
References
Institut fuer Mittelstandsforschung (IfM) Bonn (2016). “KMU-Definition des IfM Bonn”.
Kagermann, H., W. Wahlster, and J. Helbig (2013). “Recommendations for implementing the
strategic initiative INDUSTRIE 4.0”. In: Final report of the Industrie Working Group. Acatech:
Frankfurt am Main.
Kang, H. S., J. Y. Lee, S. Choi, H. Kim, J. H. Park, J. Y. Son, B. H. Kim, and S. D. Noh (2016). “Smart
manufacturing: Past research, present findings, and future directions”. In: International Journal
of Precision Engineering and Manufacturing-Green Technology 3, pp. 111–128.
Kans, M. and A. Ingwald (2016). “Business Model Development: Towards Service Management
4.0”. In: Procedia CIRP 47, pp. 489–494.
Kasabov, E. (2015). “Start-Up Difficulties in Early-Stage Peripheral Clusters: The Case of IT in an
Emerging Economy”. In: Entrepreneurship Theory and Practice 39.4, pp. 727–761.
Keqiang, L. (2015). In: Full Text: Report on the Work of the Government.
Knight, G. (2000). “Entrepreneurship and Marketing Strategy: The SME Under Globalization”. In:
Journal of International Marketing 8, pp. 12–32.
Kowalkowski, C., L. Witell, and A. Gustafsson (2013). “Any way goes: Identifying value constellations
for service infusion in SMEs”. In: Industrial Marketing Management 42, pp. 18–30.
Krippendorff, K. (2013). Content analysis. Los Angeles: Sage.
Lasi, H., P. Fettke, H.-g. Kemper, T. Feld, and M. Hoffmann (2014). “Industry 4.0”. In: Business &
Information Systems Engineering 6, pp. 239–242.
Lee, J., B. Bagheri, and H.-a. Kao (2015). “A Cyber-Physical Systems architecture for Industry
4.0-based manufacturing systems”. In: Manufacturing Letters 3, pp. 18–23.
Lee, J., E. Lapira, B. Bagheri, and H.-a. Kao (2013). “Recent advances and trends in predictive
manufacturing systems in big data environment”. In: Manufacturing Letters 1.1, pp. 38–41.
Lennartson, B., K. Bengtsson, C. Yuan, K. Andersson, M. Fabian, P. Falkman, and K. Akesson (2010).
“Sequence Planning for Integrated Product, Process and Automation Design”. In: IEEE Transac-
tions on Automation Science and Engineering 7, pp. 791–802.
Liao, Y., F. De Freitas Rocha Loures, L. F. Pierin Ramos, and F. Deschamps (n.d.). “Past, present
and future of Industry 4.0. a systematic literature review and research agenda proposal.” In:
International Journal of Production Research 2017.
Meyer, G. G., J. C. Wortmann, and N. B. Szirbik (2011). “Production monitoring and control with
intelligent products”. In: International Journal of Production Research 49.5, pp. 1303–1317.
Mikusz, M. (2014). “Towards an Understanding of Cyber-physical Systems as Industrial Software-
Product-Service Systems”. In: Procedia CIRP 16, pp. 385–389.
Miles, M. B. and A. M. Huberman (1994). Qualitative Data Analysis. An Expanded Sourcebook. Sage:
London.
Monostori, L. (2014). “Cyber-physical Production Systems: Roots, Expectations and R&D Chal-
lenges”. In: Procedia CIRP 17, pp. 9–13.
Pike, J. (2014). “AT&T, CISCO, GE, IBM and INTEL Form Industrial Internet Consortium to Improve
Integration of the Physical and Digital Worlds: Technology leaders drive industry ecosystem to
accelerate more reliable access to big data to unlock business value. [online] Industrial Internet
Consortium”. In: p. 2017.
Radziwon, A., A. Bilberg, M. Bogers, and E. S. Madsen (2014). “The Smart Factory: Exploring Adaptive
and Flexible Manufacturing Solutions”. In: Procedia Engineering 69, pp. 1184–1190.
Raja, J. Z., D. Bourne, K. Goffin, M. Cakkol, and V. Martinez (2013). “Achieving customer satisfac-
tion through integrated products and services: An exploratory study”. In: Journal of Product
Innovation Management 30, pp. 1128–1144.
317
Ren, L., L. Zhang, F. Tao, C. Zhao, X. Chai, and X. Zhao (2013). “Cloud manufacturing: From concept
to practice”. In: Enterprise Information Systems 9, pp. 186–209.
Rennung, F., C. T. Luminosu, and A. Draghici (2016). “Service Provision in the Framework of Industry
4.0”. In: Procedia - Social and Behavioral Sciences, pp. 372–377.
Schlechtendahl, J., M. Keinert, F. Kretschmer, A. Lechler, and A. Verl (2015). “Making existing
production systems Industry 4.0-ready”. In: Production Engineering 9, pp. 143–148.
Schuh, G., T. Potente, C. Wesch-Potente, A. R. Weber, and J.-p. Prote (2014). “Collaboration Mech-
anisms to Increase Productivity in the Context of Industrie 4.0”. In: Procedia CIRP 19, pp. 51–
56.
Shin, S.-j., J. Woo, and S. Rachuri (2014). “Predictive Analytics Model for Power Consumption in
Manufacturing”. In: Procedia CIRP, pp. 153–158.
Tao, F., L. Zhang, V. C. Venkatesh, Y. Luo, and Y. Cheng (2011). “Cloud manufacturing: A computing
and service-oriented manufacturing model”. In: Proceedings of the Institution of Mechanical
Engineering. J. Engineering Manufacture, pp. 1969–1976.
Veza, I., M. Mladineo, and N. Gjeldum (2015). “Managing Innovative Production Network of Smart
Factories”. In: IFAC-PapersOnLine 48, pp. 555–560.
Vogel-Heuser, B., A. Fay, I. Schaefer, and M. Tichy (2015). “Evolution of software in automated
production systems: Challenges and research directions”. In: Journal of Systems and Software,
pp. 54–84.
Voss, C., N. Tsikriktsis, and M. Frohlich (2002). “Case research in operations management”. In:
International Journal of Operations and Production Management 22, pp. 195–219.
Wan, J. (2011). “Advances in Cyber-Physical Systems Research”. In: KSII Transactions on Internet
and Information Systems 5, p. 11.
Wang, S., J. Wan, D. Zhang, D. Li, and C. Zhang (2016). “Towards smart factory for industry 4.0: A self-
organized multi-agent system with big data based feedback and coordination”. In: Computer
Networks 101, pp. 158–168.
Weston, C., T. Gandell, J. Beauchamp, L. McAlpine, C. Wiseman, and C. Beauchamp (2001). “An-
alyzing interview data: The development and evolution of a coding system”. In: Qualitative
Sociology 24.3, pp. 381–400.
Wu, J., B. Guo, and Y. Shi (2013). “Customer knowledge management and IT-enabled business
model innovation: A conceptual framework and a case study from China”. In: European Man-
agement Journal 31, pp. 359–372.
Xu, L. D., W. He, and S. Li (2014). “Internet of Things in Industries: A Survey”. In: IEEE Transactions
on Industrial Informatics 10, pp. 2233–2243.
Xu, X. (2012). “From cloud computing to cloud manufacturing”. In: Robotics and Computer-Integrated
Manufacturing 28, pp. 75–86.
Yin, R. K. (2009). Case Study Research: Design and Methods. 4th edn. Sage Publications: London.
Zhang, Y., G. Zhang, J. Wang, S. Sun, S. Si, and T. Yang (2014). “Real-time information capturing
and integration framework of the internet of manufacturing things”. In: International Journal of
Computer Integrated Manufacturing 28, pp. 811–822.
Zuehlke, D. (2010). “SmartFactory: Towards a factory-of-things”. In: Annual Reviews in Control 34,
pp. 129–138.
318
Design of a Cyber-Physical Production
System for Semiconductor
Manufacturing
Sophia Keil1
1 – Technische Universität Dresden
Due to the highly dynamic markets, an increasing complexity, and individualiza-

tion of products, efficient and robust logistical processes are difficult to achieve
through the use of central planning and control approaches. The aim of the contri-
bution is the design of a decentralized, autonomous control system for high tech
production systems. An interdisciplinary perspective was adopted as methods of
artificial intelligence and mechanical as well as electrical engineering were used.
The results are a hardware concept for an intelligent, cyber-physical production
lot and a software concept based on a hierarchical multi-agent architecture. The
basic idea of autonomy and self-control is not new. It can be traced back, for
example, to the ideas of “Evolutionary Management”, or cybernetics.
However, for the first time this contribution shows a practical application for a
complex semiconductor manufacturing system. Until now, the hard and software
concepts have been implemented prototypically. A long-term integration into the
existing IT landscape of a semiconductor factory is planned. A well established and
functioning centralized system should be supplemented by the new decentralized
system, especially in areas in which there is not yet such a high level of automated
processes, e. g. in wafer test facilities.
Keywords: cyber-physical production system; multi-agent system; intelligent

production lot; semiconductor manufacturing
319
Design of a Cyber-Physical Production System for Semiconductor Manufacturing
1 Introduction
Competitive semiconductor companies need to deliver high quality products in a

fast, timely, and cost-efficient manner in the context of highly dynamic markets,
which are characterized by rapidly declining prices, short product life cycles, and
a high frequency of technology changes (Sonar et al., 2013). Short product life
cycles, an increasing complexity, and individualization of products, which are
usually provided in a high number of variants, lead to sophisticated requirements
with respect to logistics processes and systems. Hereby, centrally controlled
systems reach their limits concerning their capabilities to deal with the arising
complexity to plan, control, and monitor changeable systems (Schuhmacher and
Hummel, 2016). Ten Hompel (2010) proposes the adoption of individualization
as a design principle for logistics processes and systems. A key approach in
this context is the concept of self-control. Self-controlled systems promise the
advantages of increased (Brettel et al., 2014)
— flexibility, in terms of changeability of the system;

— robustness, as the decision-making competence is shifted to individual
logistical objects; and
— data availability as well as transparency of complex material flow struc-
tures.
Self-control includes two main concepts: decentralization and autonomy. The

degree of decentralization indicates on which system level control decisions are
made. Self-control is attained when the degree of decentralization reaches the
level of the physical material flow. That means the individual logistical objects
such as the goods to be transported and the load carriers, but also the transport
systems, make autonomous control decisions. The degree of autonomy indicates
how many decision making opportunities the individual logistic object has. A
prerequisite for the realization of self-controlled systems is a certain level of
intelligence, which is realized by ICT technologies and, more importantly, cyber
physical systems (Scholz-Reiter et al., 2005).
Following these thoughts, the main research question of the contribution is as
follows: How should a cyber-physical production system for the semiconductor in-
dustry be designed to enable manufacturing excellence? To answer the question,
the paper is organized in the following manner: after this introduction, Section 2
includes a short overview about the theoretical background of Cyber-Physical
320
Systems and Cyber-Physical Production Systems as well as existing research and

industrial approaches. Section 3 comprises a characterization of the case study
of semiconductor production. Section 4 includes the design of a semiconductor
manufacturing specific CPPS, including with the software and hardware concept
as well as a validation of the concept. Lastly, a summary and outlook for future
tasks are put forth.
2.1 Cyber-Physical Production Systems
Figure 1 shows the structure of a Cyber-Physical System (CPS) schematically.

Within a manufacturing system, an embedded system in the sense of a CPS is
integrated within physical systems, e. g. the machines or production lots. The em-
bedded system includes sensors to gather physical data and electronic hardware
as well as software to save, and analyze data. The results of the data processing
are the foundation for an interplay with other physical or digital systems by means
of actuators (Lee et al., 2015). Furthermore, a CPS comprises a human machine
interface, e. g. for exchange of information and supervision (Geisberger and Broy,
2015). A cyber-physical production system (CPPS) can be formed when numerous
cyber-physical systems are linked and cooperate through digital networks (Seitz
and Nyhuis, 2015).
CPSs are intelligent objects of class 4 as they are characterized by the features:
identification, memory capacity, data processing and interaction/ communica-
tion (Zbib et al., 2008). Thereby, CPSs enable the design of intelligent logistics
systems, where autonomous self-control is one characteristic of intelligent sys-
tems (Reinhart et al., 2013; Ostgathe, 2012). Further characteristics of intelligent
systems are described in the following Section 2.2.
2.2 Characteristics of Intelligent Systems
In order to understand artificial intelligence within systems, it is useful to first

characterize human intelligence, which cannot be described by a single feature.
During a symposium held in 1986, experts provided answers to the question what
321
Figure 1: Structure of a CPS
human intelligence is (Sternberg and Detterman, 1986). These experts most fre-
quently used the following characteristics to define the concept of intelligence:
First, intelligence involves elementary processing processes (perception, sensi-
tivity, attention). Individuals must be able to perceive their environment, have
knowledge, and then reach a higher level of processing, such as logical conclu-
sions, imagination, problem solving, and judgments. In addition, the adaptability
to a changing environment belongs to the concept of intelligence. Kail et al. (1988)
analyzed numerous definitions of human intelligence. They found that most defi-
nitions include the ability of humans to think abstractly and to reason and derive
purposeful actions from it.
The dimensions of thinking and action are also considered in the research field of
artificial intelligence (AI). There are two approaches: First, researchers can try to
understand how human beings think and act, and then model or simulate it on a
computer. Second, the researcher tries to find an optimal approach independent
of how humans would solve the problem, which would be represented by the
rational view. A system is rational when it does the right thing according to its
own knowledge (Russel and Norvig, 1995). Thereby, strong and weak AI are dis-
tinguished. The aim of strong AI is to develop AI to the point where the machine’s
intellectual capability is functionally equal to a human’s or even surpasses it, e.g.
Blue Brain Project. Weak or narrow AI is machine intelligence that is limited to
322
specific application domains, e.g. expert systems (Russel and Norvig, 1995; Ertel,
2009)
In general, intelligent systems should be autonomous, proactive, adaptive, self-
explanatory, fault-tolerant, self-optimizing, adaptive, goal-oriented, flexible, and
cooperative. There is no system worldwide that maintains all of these features; but
if none of these features are present, the system is also not considered intelligent
(Wahlster, 2013).
Autonomy is a main feature of self-controlled systems. Methodological approaches
with respect to self-control have already existed for many years. They can be
regarded from different scientific perspectives. According to the system’s per-
spective, a system can be divided into the levels decision system, information
system, and execution system (Ropohl, 1979). The research with respect to logis-
tics systems includes the self-controlled physical flow of materials and goods and
their accompanying information flow and technology realization as well as the
management of self-organizing logistics processes (Freitag et al., 2004).
The decision system is reflected by the management and organization literature
(Windt, 2006), with a major research area focusing on the Evolutionary Manage-
ment approach. Hereby, researchers transfer approaches of the evolution of
natural organic systems to the evolution of enterprises (Malik and Probst, 1984).
The information and execution system are reflected by research in the areas
of science, technology, engineering, and mathematics. Hereby, two major ap-
proaches are to apply swarm intelligence and multi-agent systems (MAS) (Windt
2006; Scholz-Reiter and Höhns, 2006, Monostori, 2014; Wang et al., 2015). Exam-
ples for the application of swarm intelligence and MAS in the context of CPPS are
the research projects CoCos, InnoCyFer, and SMART FACE (Bundesministerium
für Wirtschaft und Energie, 2016). The research projects show that agent technol-
ogy is a promising approach to implementing a decentralized and autonomous
production controlling system. Therefore, it will be used to achieve the research
goal of designing a self-controlled production system for semiconductor man-
ufacturing (Section 4.1). A characterization of semiconductor manufacturing is
described in the following Section 3.
323
3 Characterization of the Case Study of Semiconductor

Production
3.1 Manufacturing Organization within the examined Factory
Generally, semiconductor fabrication facilities are organized in a job shop (Chien

et al. 2016, Chen et al., 2008; Puffer, 2007). Here, the manufacturing tools are
clustered according to their function. This enables high capacity utilization, but
causes rather long lead times in contrast to the organization according to flow
production where the installation of machines follows the product workflow
(Miltenburg, 2005).
In the investigated factory, the job shop manufacturing organization is reflected

in the bay-chase fab layout. The production floor is structured into different bays,
the “job shops”. Each of these shops comprises similar types of machines, which
accomplish one function (Meyersdorf and Taghizadeh 1998; Chang and Chang
1998). These production bays, which are also called intrabays, are linked to a con-
necting corridor, the interbay (see Figure 2). Between the bays are maintenance
chases (grey room) (Chien et al. 2016). Besides, Figure 2 shows also elements of
the Automated Material Handling System: stocker and lifts, which are described
in the Section 3.2.
3.2 Material Handling System within the examined Factory
The following explanations describe the Automated Material Handling System

(AMHS) of the investigated semiconductor company. This is necessary since
the individual system elements are part of the decentralized control system
to be designed. Within the company it is called the wafer transport system.
Wafers represent the raw material for the production of Integrated Circuits (ICs).
They are formed of highly pure (99.9999999% purity), single crystalline material
(Winzker, 2008).
The task of the wafer transport system is to connect the various production areas
with a material handling technique. For this purpose, a defined transport unit is
transported in carriers from work station to work station. Main components of the
wafer transport system are (Deutschländer et al., 2005; Heinrich and Pyke, 1999):
324
3 Characterization of the Case Study of Semiconductor Production
Figure 2: Bay-Chase Facility Layout
325
Figure 3: Part of the conveyor AHMS – depicting open carrier transportation on

top left (Courtesy of Infineon Technologies Dresden GmbH)
conveyor, carrier, lift, lower and upper buffer, stocker, and the Material Controlling
System (MCS).
Linking the production machines is carried out by a conveyor. On the conveyor
tracks, open carriers move at a speed of 0.23 m/sec. The total length of the con-
veyor is 4.2 km, which is installed on the ceiling of the clean room. The transport
and the machine processing are divided into two different levels. This results
in a high machine density in the clean room area. The height difference is 2.70
m from the floor (machine processing level) to the transport level at the ceiling
(see Figure 3) (Niekisch, 2001). Carriers are used as conveying aids (see top left
of Figure 3). They can hold 25 wafers each. This transport unit corresponds to a
production lot.
Each carrier can be tracked (conveyor path, current location) by means of carrier
identification. The lot number and carrier identification are stored in the mem-
ory chip of the carrier. Lifting and lowering stations connect the transport with
326
3 Characterization of the Case Study of Semiconductor Production
the shop floor level. These technical devices, hereinafter referred to as lifts, are
typically arranged near the production machines to minimize handling efforts
(Niekisch, 2001).
For the purpose of synchronization of consecutive process steps within the manu-
facturing processes, storage devices, so called buffer, are installed in the semicon-
ductor factory. They are arranged in the immediate vicinity of a lift. If the buffer
nearby the machine is being used to capacity, the lots are stored in stockers which
are directly located in front of a production bay (Niekisch, 2001). The existing
central production control system is a Manufacturing Execution System (MES)
based on legacy Workstream, which is a trademark of the firm Applied Materials
(Heinrich and Pyke, 1999).
327
4 Design of a CPPS for Semiconductor Industry
4.1 Multi-agent oriented Software Concept
As described in Section 2.2, the agent technology is a promising approach to im-

plementing a decentralized and autonomous production control system. In agent
technology, a multi-agent system (MAS) is seen as a society of independent actors
which solve different tasks under competition or in cooperation (Bussmann et al.,
2004). A software agent can be described as “a self-contained program capable of
controlling its own decision making and acting, based on its perception of its environ-
ment, in pursuit of one or more objectives” (Jennings and Wooldridge, 1996, p. 1).
A set of interacting agents is referred to as an MAS (Bussmann et al., 2004). In the
case of software agents, the interaction is based in particular on the exchange
of messages, while in robotics the common physical work is also considered
(Scholz-Reiter and Höhns, 2006). A generally accepted definition for MAS and
their applications in complex production systems has not been established to this
day. Rather, the agent is characterized by its role, tasks, and the skills required
for it. Wooldridge and Jennings (1995) name four properties which are a basic
prerequisite for agents:
— Autonomy: agents run sans involvement of people and have control with
respect to their activities and condition (Castelfranchi, 1994);
— Social ability: agents work together with other agents (also people)
through an agent communication language (Genesereth and Ketchpel,
1994);
— Reactivity: agents notice their surrounding conditions and respond to

variations; and
— Pro-activeness: agents are able to show purposeful behavior, grasping

the nettle.
For the design of the MAS, the Process for Agent Societies Specification and Im-
plementation (PASSI) approach was applied which includes five models: systems
requirements model, agent society model, agent implementation model, code
model, deployment model (Cossentino and Seidita, 2014). Exemplarily, the sys-
tem requirements model is described in the following, as this model is the major
input parameter for all the other models. A hierarchical agent society was chosen,
328
Figure 4: Hierarchical Agent Society
as this is the standard approach for production control of complex production

systems, such as semiconductor manufacturing (Mönch and Stehli, 2006).
Figure 4 shows the three main levels of the system pyramid consisting of
— Interface Humans
— Planning/Controlling
— Execution
The top level of the pyramid, “Interface Human”, is used to monitor and control
the entire system by humans. This allows them to access the system directly,
supported by the fab agent. For this purpose, a graphical interface is provided
which visualizes the status of the fabrication facility and offers the possibility
of interacting with the system. The planning and controlling level follows the
thoughts described in Keil et al. (2011). Hereby, the lot agent plays a central
role and is provided with a clock based production schedule and cumulative
quantities (Löding, 2008) by the area agent. Based on the schedule, the lot agent
negotiates with the identified agents, e. g. with the bay agent regarding the
processing capability.
329
Figure 5: Hierarchical Agent Society and roles according to PASSI
The lot agent is able to communicate over all levels of the pyramid. He can act
directly in the event of disruptions or deviations from the plan. In addition, the
lot agent is able to perceive its environment by means of corresponding hardware
at the carrier, which is described in Section 4.2. This enables the lot agent to react
quickly to external influences.
Figure 5 shows exemplarily the negotiation relationships between the agents. At

the interface between the human and the fab agent, the employee physically
releases a production order (lot) (1) and renders the product workflow (2) of the
product to the fab agent. The fab agent creates a software lot agent (3). After-
wards, the fab agent communicates with the area agent and asks him to create
a clock-based production schedule for the new lot (4). The software lot agent
gets the production schedule from the area agent (5). Based on this schedule, the
lot agent asks the area agent which bay agent is responsible for the needed first
production step (6). After receiving the answer, the lot agent asks the responsible
bay agent to offer him production capacity in a certain time frame with respect to
330
Figure 6: FAB-Monitor for visualization of the status characteristics of the semi-

conductor factory
a certain production recipe (7). For example, the bay agent asks for two machines
in his bay which are able to carry out the needed production step (8). Maybe
machine 1 responds: “I am down due to a failure” and machine 2 says: “currently,
I am carrying out production step x, but in 50 minutes I will be available”. Then
the bay agent offers the capacity to the lot agent, who accepts the offer (9). Now,
the lot agent asks the transport agent: “could you transport me to bay x, and lift
y?”. Maybe the transport agent says “yes, in 30 minutes, I can take you to the des-
tination and in the meantime I will transport you to stocker y for an intermediate
storage” (10).
In the background, the transport agent has asked the responsible stocker agent
with respect to his capacity for temporary storage (11). Furthermore, the transport
agent organizes the transport of the lot from the stocker to a buffer nearby a lift
and the transport to the operation level by the respective lift (12).
331
The JADE (Java Agent Development Framework) was selected as an appropriate

framework for implementing the presented MAS as it is in compliance with the
Foundation for Intelligent Physical Agents (FIPA) which offers generic agent tech-
nologies. Thus, the communication with other agent platforms is safeguarded
and the integration effort is decreased (Bellifemine et al., 1999; Bellifemine et
al., 2008). The realization of the MAS occurred in two steps. Hereby, the produc-
tion system was implemented in the first step. This enabled the modeling of a
semiconductor factory and to feed in production orders. For the visualization
of these tasks, the visualization tool ”FAB-Monitor” was designed (see Figure 6),
which directly represents the activities of the MAS. Hereby, at the “interface hu-
mans” the employees can see, for example, which machines are in maintenance
or where system interferences actually occurred. Unlike traditional visualization
tools which are based on a central data structure, the newly deployed decentral-
ized solution does not rely on a higher-level node. The advantages lie in the relief
of the entire system as well as in a maximum topicality of the visualized data.
In the second step, the module for the planning of production orders was devel-
oped. For this purpose, current manufacturing data, such as machine failures,
capacity bottlenecks, and processing times, are used. These data are provided by
the lot agent, the hardware concept of which is described in Section 4.2.
4.2 Hardware Concept for a Cyber-Physical Production Lot
The goal is to develop a cyber-physical production lot. Therefore, a microcom-

puter with communication technology, a sensor system, and a power supply
is integrated into the wafer carrier. For the receiving, evaluation, and further
processing of the sensor data, access points are used. These access points are
supplied with energy by means of Power over Ethernet. The hardware concept,
which is schematically illustrated in Figure 7, is discussed in detail below.
To determine the processing times of the single production steps, it is necessary

to measure the time span from the removal of the first wafer to the completion of
the last wafer. For this purpose, forked light barriers are located in each slot of
the carrier, which can detect the presence of a wafer. The data of the sensors are
collected by an ATMega32 microcontroller from the firm Atmel. This controller
communicates via the SPI interface with another microcontroller, an NA1TR8 of
the firm Nanotron Technologies. This controller is responsible for the evaluation,
conversion, and transmission of the collected sensor data.
332
Figure 7: Hardware Concept of the Cyber-Physical Production Lot (CPPL)
The data are transmitted using the approach Chirp Spread Spectrum (CSS), which
is part of the Multi Dimensional Multiple Access (MDMA) approach. This method
provides high data transfer rates and extremely low power consumption and
was included in the NA1TR8 controller by the company Nanotron Technologies
(Nanotron, 2017).For the purpose of easy handling, all electronics (ATMega32,
NA1TR8, SMD antenna, and other components) are located on one circuit board
with dimensions of 95 x 20 x 4 mm (width x height x depth). An additional circuit
board is required for the sensors and their control electronics.
For the power supply, film batteries are used which have a depth of about 3 mm. In
order to guarantee sufficient energy supply over a period of 50 days (approximate
cycle time of a lot), two batteries with 430 milliampere hours (mAh) capacity each
are required. Since the weight of the batteries together is about 27 grams, and
the cassette must be balanced, the batteries are attached to the opposite carrier
wall.
Access points are available to receive data sent from the sensor nodes. Due to
the absence of high transmission power, it is necessary to install several access
points in the semiconductor factory. The hardware of the access points consists
of a radio module from Nanotron Technologies and a microcomputer (PXA270),
which was provided by the company Phytec. This microcomputer has the full
functionality of a personal computer. On the Linux operating system, a program
for receiving and evaluating data is installed. This data can be forwarded to the
333
requesting agent via Ethernet. The concept of Power over Ethernet is used for
power supply. Hereby, the data is provided and at the same time the voltage is
supplied via a single Ethernet cable. Thus, no further power supply cable with
adaptor is necessary, which significantly increases the flexibility with respect to
the positioning of the access points.
The described hardware concept can be easily expanded if new functions are
required. Further steps include, for example, the function perceiving the sur-
roundings of the carrier by means of further sensor technology. Thus, for example,
”lost” wafers can be detected at an early stage after cleaning operations. This con-
tributes to a stability of production processes, since disturbances can be detected
at an early stage. A task of the microcomputer is the collection of all essential
processing data of the lot. This ensures, on the one hand, the actuality of the data
and, on the other hand, the achievement of a high maturity level of technological
processes. Development, process, maintenance, and product engineers have the
opportunity for easy data evaluation since all collected data (e.g. process temper-
atures, end point times, processing and material flow times) are available for one
lot at one location. This allows quick decisions and reactions of the engineers.
In addition, the ”intelligent” lot, which is equipped with the microcomputer, is
able to carry out the tasks of production scheduling independently on the basis
of predetermined targets.
4.3 Validation of the Concept
The validation of the concept includes three important aspects:
— Design and test of the radio network in the semiconductor factory,

— Integration of the carrier with the new electronic components in the
existing factory, and
— Integration of the MAS in the exiting IT landscape
To point 1: Since the software lot agent runs on a server and not on the physical
production lot carrier like the sensor technology, the communication between
the carrier and the server must take place via a radio network. Therefore, the
construction of a stable radio network is an important prerequisite for the imple-
mentation of the concept. As described in Section 4.2, the data is transmitted via
the MDMA approach, which enables low energy consumption in contrast to other
technologies like WLAN, Bluetooth, or ZigBee (Masini, 2015). This is of critical
334
importance since the physical production lot agent stays in the factory for about
50 days.
In a first step, the transmission quality had to be determined. A measurement
was carried out in the laboratory of the University of Applied Sciences Stralsund
under the following conditions:
— line of sight between transmitter and receiver: distance 80 meters,

— without direct line of sight between transmitter and receiver: distance 40
meters, and
— without a visual connection through a wall: distance 8 meters.
The results of the measurements did show that an acceptable transmission qual-
ity exists at a transmission power of 8 dBm and a distance of 40 m without a
visual connection. Due to tolerances and possible interferences, the distance
between an access point and a sensor node should not exceed 30 meters. Based
on these results, a statement about the positioning of the access points in the
semiconductor factory can be made. As an intrabay has a length of 55.3 meters
(see Figure 2), one access point is integrated in the middle of each intrabay. Ad-
ditional access points in the interbay enable communication between several
intrabays. 38 access points are required for the entire production system.
In a second step an access point was positioned in an intrabay of the regarded
semiconductor manufacturer. The field test in the company did show that some of
the machines emit interfering electromagnetic fields. Therefore, in the following
field tests it is necessary to separately determine the transmission quality in each
intrabay. If there is a reduced transmission quality, additional access points must
be stationed in the respective intrabay. In addition, a possible negative impact of
the radio waves on the machines has to be investigated.
An alternative to the described radio network based on the MDMA approach would
be to use the existing WLAN network of the company or to use the radio network of
a network operator. Then a new energy concept for the cyber physical production
lot would have to be designed, as the data transmission within these networks
would consume more energy. One alternative approach to solving the energy
issue would be to load the carrier during the waiting times in the stocker.
To point 2: Due to clean room requirements, the carrier has to be cleaned ap-
proximately every 100 days. Therefore, in a second step, a watertight box, which
includes the electronic components, was designed. In general, several thousand
335
carriers are used in the production system. When the system is introduced into
the factory, the company would have to ask the carrier supplier to integrate the
electronics into the carrier during the manufacturing of the carrier. Furthermore,
the retrofitting of the system requires rather high investment costs. In general,
there is more than one solution for the technical realization of the cyber physical
production lot. The realization depends on the material flow technique used in
the respective company.
To point 3: The MAS system currently works in a simulation environment. So
far, the focus has been set on the visualization of factory status (see Figure 6).
Until now, about 100 production lots have been introduced into the system. The
next steps are to examine the system behavior with respect to lead times and
capacity utilization at a more realistic number of several thousand production lots.
A huge challenge is the connection of the MAS to the existing material execution
system. As the existing dispatch rules cannot be overridden due to the risk of loss
of production for the firm, a step-by-step approach is required. It is proposed
to use the new cyber physical production system in less automated production
areas. For the semiconductor industry this would be factories that test wafers. In
those factories the functionality of the manufactured products is tested, e.g. with
respect to electrical parameters or through stress tests.
5 Summary and Outlook
The megatrends toward individualization of products and shorter delivery times

together with rising cost and efficiency pressures, lead to an increasing complexity
with respect to the organization, planning and control of production processes.
Through the development and introduction of CPPS, this trend can be coun-
tered. By using the example of semiconductor manufacturing, a decentralized
and autonomous production controlling system based on multi agent technology
supported by a cyber-physical production lot was presented.
The sensor technology within the cyber-physical production lot enables data col-
lection, which enhances the transparency of the production system. For example,
in case of disturbances, employees can react fast and negative effects on other
products can be avoided. This is achieved by the presented fab monitor, which
visualizes the data collected by the lots. In general, the presented lot agent can
enhance the data quality for production planning and control. Nevertheless, new
production control strategies and accompanying technologies do not replace
336
References
the need for designing lean processes. Furthermore, the next step is to design a
control system for production networks.
Although the technology for CPPS is available, production systems of firms cannot
be changed from day to day as it is challenging to integrate advanced produc-
tion control strategies into legacy software systems of firms. Here firm specific
transition concepts are needed.
For the future, it is of major importance to answer the question how the role of
the human is defined in a CPPS as job profiles and the work division between
CPPS and the humans will alter in the face of the technical progress. In addition,
organizational forms have to be changed to facilitate decisions of employees, e.g.
by swarm organizations and flat hierarchies, which reflect the decentralization,
and autonomy in CPPS.
Acknowledgements
The development work was pursued under the leadership of Arthur Deutschländer,
Sophia Keil and Mathias Kuttig at the University of Applied Sciences Stralsund
with the support of Robert Kohn, Andreas Nack, Christoph Nagel, Ricardo Mussehl,
Lars Böcker, Diana Riemer and Bernd Rummler. Thanks a lot to the employees of
the company Infineon Technologies Dresden GmbH, especially Germar Schneider
and Harald Heinrich who supported the work.
References
Bellifemine, F., G. Caire, A. Poggi, G. Rimassa, and A. Jade (2008). “A white paper”. In: Telecom
Italia EXP magazine 3.
Bellifemine, F., A. Poggi, and G. Rimassa (1999). “JADE–A FIPA-compliant agent framework”. In:
Proceedings of PAAM (Vol. 99, No. 97-108, p. 33).
Brettel, M., N. Friederichsen, M. Keller, and M. Rosenberg (2014). “How virtualization, decentraliza-
tion and network building change the manufacturing landscape: An industry 4.0 perspective”.
In: International Journal of Mechanical, Industrial Science and Engineering 8.1, pp. 37–44.
Bussmann, S., N. R. Jennings, and M. J. Wooldridge (2004). Multiagent systems for manufacturing
control: a design methodology. Science & Business Media: Springer.
Castelfranchi, C. (1994). “Guarantees for autonomy in cognitive agent architecture”. In: Interna-
tional Workshop on Agent Theories, Architectures, and Languages. Berlin Heidelberg: Springer,
pp. 56–70.
337
Chang, C. F. and S. K. Chang (1998). “A layer-based layout approach for semiconductor fabrication
facilities”. In: Advanced Semiconductor Manufacturing Conference and Workshop. IEEE/SEMI,
pp. 385–390.
Chen, J. C., R. D. Dai, and C. W. Chen (2008). “A practical fab design procedure for wafer fabrication
plants”. In: International Journal of Production Research 46.10, pp. 2565–2588.
Chien, C. F., C. H. Hu, and Y. F. Hu (2016). “Overall Space Effectiveness (OSE) for Enhancing Fab
Space Productivity”. In: IEEE Transactions on Semiconductor Manufacturing 29.3, pp. 239–247.
Cossentino, M. and V. Seidita (2014). “PASSI: Process for agent societies specification and imple-
mentation”. In: Handbook on Agent-Oriented Design Processes. Ed. by Springer. Berlin Heidel-
berg, pp. 287–329.
Deutschländer, A., G. Brückner, H. Heinrich, and S. Keil (2005). “Materialflussautomatisierung unter
Einsatz von Robotern in der Halbleiterfertigung mit 200 mm-Wafer-Technologie”. In: Logistik
Management, Innovative Logistikkonzepte, pp. 241–255.
Ertel, W. (2009). Grundkurs Künstliche Intelligenz. Wiesbaden.
Freitag, M., O. Herzog, and B. Scholz-Reiter (2004). “Selbststeuerung logistischer Prozesse – Ein
Paradigmenwechsel und seine Grenzen”. In: Industrie Management 20.1, pp. 23–27.
Geisberger, E. and M. ( Broy (2015). Living in a networked world: Integrated research agenda Cyber-
Physical Systems (agenda CPS). Herbert Utz Verlag.
Genesereth, M. R. and S. P. Ketchpel (1994). Software agents. Tech. rep. Center for Integrated
Facility Engineering. Stanford University.
Heinrich, H. and A. Pyke (1999). “The Impact of Conveyor Transports on Factory Performance at
Infineon‘s (Siemens) 200 mm Fab”. In: Semiconductor Fabtech 10, pp. 15–18.
Hompel, M. ten (2010). Individualisierung als logistisch-technisches Prinzip. Internet der Dinge in
der Intralogistik. Berlin Heidelberg: Springer.
Jennings, N. and M. Wooldridge (1996). “Software agents”. In: IEE review 42.1, pp. 17–20.
Kail, R., J. W. Pellegrino, and A. Hildebrandt-Essig (1988). Menschliche Intelligenz. Heidelberg:
Verlag Spektrum der Wissenschaft.
Keil, S., D. Eberts, T. Igel, G. Schneider, K. Wilhelm, R. Lasch, and A. Deutschländer (2011). “In-
novation and manufacturing excellence in mature multi product semi-conductor fabrication
facilities via design for flow by 3”. In: Semiconductor Conference Dresden (SCD), pp. 1–5.
Lee, J., B. Bagheri, and H. A. Kao (2015). “A cyber-physical systems architecture for industry 4.0
based manufacturing systems”. In: Manufacturing Letters 3, pp. 18–23.
Lödding, H. (2008). Verfahren der Fertigungssteuerung: Grundlagen, Beschreibung, Konfiguration.
Berlin: Springer Verlag.
Malik, F. and G. Probst (1984). “Evolutionary management”. In: Self-organization and management
of social systems. Ed. by S. B. Heidelberg, pp. 105–120.
Masini, M. (2015). Störungsunempfindliche Informationsübertragung in dynamischer Produktion-
sumgebung. BoD–Books on Demand.
Meyersdorf, D. and A. Taghizadeh (1998). “Fab layout design methodology: Case of the 300 mm
fabs”. In: Semiconductor international 21.8, pp. 187–196.
Miltenburg, J. (2005). Manufacturing strategy: how to formulate and implement a winning plan.
CRC Press.
Mönch, L. and M. Stehli (2006). “ManufAg: a multi-agent-system framework for production control
of complex manufacturing systems”. In: Information Systems and e-Business Management 4.2,
pp. 159–185.
338
References
Monostori, L. (2014). “Cyber-physical production systems: Roots, expectations and R&D chal-
lenges”. In: Procedia CIRP 17.2014, pp. 9–13.
Nanotron (2017). “Company information”.
Niekisch, T. (2001). “Störungssystematikbestimmte Instandhaltung in einem Wafertransportsys-
tem”. PhD thesis. Technische Universität Dresden.
Ostgathe, M. (2012). System zur produktbasierten Steuerung von Abläufen in der auftragsbezogenen
Fertigung und Montage. Herbert Utz Verlag.
Puffer, W. (2007). “Technisch-ökonomische Effizienzbetrachtungen für die Halbleiterfertigung”.
PhD thesis. Technische Universität München.
Reinhart, G., P. Engelhardt, F. Geiger, T. Philipp, W. Wahlster, D. Zühlke, and P. Stephan (2013).
“PCyber-Physische Produktionssysteme. Produktivitäts-und Flexibilitäts-steigerung durch die
Vernetzung intelligenter Systeme in der Fabrik”. In: wt-online 103.2, pp. 84–89.
Ropohl, G. (1979). “Eine Systemtheorie der Technik”. In: München.
Russell, S., P. Norvig, and A. Intelligence (1995). Artificial Intelligence. A modern approach. Egnle-
wood Cliffs: Prentice-Hall.
Scholz-Reiter, B., M. Freitag, H. Rekersbrink, B. L. Wenning, C. Gorldt, and W. Echelmeyer (2005).
“Auf dem Weg zur Selbststeuerung in der Logistik-Grundlagenforschung und Praxisprojekte”.
In: Intelligente Logistikprozesse: Konzepte, Lösungen Erfahrungen - Begleitband zur 11. Magde-
burger Logistiktagung. Ed. by Logisch-Verlag. Magdeburg: Wäscher, G., pp. 166–180.
Scholz-Reiter, B. and H. Höhns (2006). “Selbststeuerung logistischer Prozesse mit Agentensyste-
men”. In: Produktionsplanung und-steuerung. Berlin Heidelberg: Springer, pp. 745–780.
Schuhmacher, J. and V. Hummel (2016). “Decentralized control of logistic processes in cyber-
physical production systems at the example of ESB Logistics Learning Factory”. In: Procedia
CIRP 54, pp. 19–24.
Seitz, K. F. and P. Nyhuis (2015). “Cyber-physical production systems combined with logistic
models – a learning factory concept for an improved production planning and control”. In:
Procedia CIRP 32, pp. 92–97.
Sonar, A., S. Shinde, and S. Teh (2013). “Automation: Key to Cycle Time Improvement in Semicon-
ductor Manufacturing”. In: IEEE/SEMI Advanced Semiconductor Manufacturing Conference 2013.
Saratoga Springs, NY.
Sternberg, R. J. and D. K. ( Detterman (1986). What is intelligence?: Contemporary viewpoints on its
nature and definition. Praeger Pub Text.
Wahlster, W. (2013). Künstliche Intelligenz: Computer mit Augen, Ohren, Hand und Fuß – aber auch
Verstand? Künstliche Intelligenz: Chance oder Risiko? Tech. rep. Hamburger Akademievorlesun-
gen.
Wang, L., M. Törngren, and M. Onori (2015). “Current status and advancement of cyber-physical
systems in manufacturing”. In: Journal of Manufacturing Systems 37.2, pp. 517–527.
Windt, K. (2006). “Selbststeuerung intelligenter Objekte in der Logistik”. In: Selbstorganisation:
Ein Denksystem für Natur und Gesellschaft. Ed. by B. Verlag. Köln Weimar: Vec, M., Hütt, M. T.,
Freund, A. M.
Winzker, M. (2008). Elektronik für Entscheider: Grundwissen für Wirtschaft und Technik. Springer-
Verlag.
Wirtschaft und Energie), ( für (2016). Autonomik für Industrie 4.0. Ergebnisse. Tech. rep.
Wooldridge, M. and N. R. Jennings (1995). “Intelligent agents: Theory and practice”. In: The knowl-
edge engineering review 10.2, pp. 115–152.
339
Zbib, N., S. Raileanu, Y. Sallez, T. Berger, and D. Trentesaux (2008). From passive products to
intelligent products: the augmentation module concept. 5th International Conference on Digital
Enterprise Technology. Nantes, France.
340
Part III
Risk and Security Management

Cybersecurity in Ports: a Conceptual
Approach
Jenna Ahokas1 , Tuomas Kiiski1 , Jarmo Malmsten1 , Lauri Ojala1
1 – Turku School of Economics at the University of Turku
As the world is becoming increasingly digitalized, the role of cybersecurity on

society is mounting. Recent cyberattacks have showed the vulnerability of critical
infrastructure, including ports. The objective is to describe how cybersecurity is
perceived in ports, as preparedness and regulation for cyberthreats in ports ap-
pears inadequate. The study is a conceptual analysis built upon a comprehensive
literature review. The results show that regardless of the growing awareness of
the issue, much work needs to be done in order to mitigate the cyberthreats in
ports. Situation calls for, among other things, adoption of industry standards and
practical level coordination.
Cybersecurity in general has been a topical subject, while in the context of ports
the theme has thus far been scantly studied. In addition, cybersecurity is currently
not included in International Maritime Organization’s safety and security Con-
ventions relevant to port, such as ISPS or ISM. Hence this study is among the first
openings in its field.
Keywords: Cybersecurity; Maritime Security; Critical Infrastructure; Ports
343
Cybersecurity in Ports
1 Introduction
The world is rapidly becoming digitalized and dependent on efficient communica-

tion systems. The activities, which traditionally have relied on paper documenta-
tion or manual processing, are now increasingly converted into electronic format,
where data is stored in a digital environment called cyberspace (Goldby, 2008;
Fitzgerald, et al., 2013). For business, the implications of these changes are mas-
sive by any standards, including improved data integrity, processing capacity and
emergence of new business models (Boyes, Isbell and Luck, 2016). Some have
even visioned a dawn of the fourth industrial revolution (Lasi, et al., 2014).
Downside of this development is that reliance on technology makes society
susceptible to the functionality of systems (Urcioli, 2015; Carrapico and Bar-
rinha, 2017). Along with natural disasters and terrorism, malicious acts by in-
dividuals through cyberspace on essential systems have been underlined as a
potential threat to the society (Kapto, 2013). The motive behind cyberattacks are
diverse, including excitement, money and political agendas (Ahokas and Kiiski,
2017). Assuring the resiliency of society against cyberattacks requires focusing
on cybersecurity of critical infrastructure. Cybersecurity indicates the security of
cyberspace in terms of access to, and control and storaging of data (Boyes, Isbell
and Luck, 2016).
The concerns relating to cybersecurity have already materialized as the number
of cyberattacks have shown a year-by-year increase, causing substantial financial
losses to society in general and business in particular (Colesniuc, 2013). Hence
cybersecurity has been identified as a top-level priority among policymakers,
businesses, scholars and individual persons (Lewis, 2002). As a result, adoption
of specific cybersecurity strategies have been initiated (ICC, 2015).
Ports are key nodes of global trade — approximately 80 per cent of world trade
is transported by sea — and thus comprise an integral part of critical infrastruc-
ture (UNCTAD, 2016). In addition, ports hold substantial amounts of data, are
involved in a large number of monetary transactions and stakeholders making
them attractive objects for cyberattacks (CyberKeel, 2014; Jensen, 2015).
European Network and Information Security Agency (ENISA 2011) was one of
the first to identify a lack of awareness of cybersecurity in maritime transport
and ports. Despite substantial academic interest shown towards cybersecurity in
general (Hult and Sivanesan, 2013) as well as maritime security (Germond, 2015),
cybersecurity in ports seems far more uncharted an area (Ahokas and Kiiski, 2017).
344
2 Security Aspects in Ports
Fresh empirical evidence from the Baltic Sea Region indicates that preparedness
and regulation for cyberthreats in seaports appears inadequate (Ahokas and
Laakso, 2017).
Concerns about cybersecurity have risen relatively recently, which is evidenced by
the fact that cybersecurity is currently not mentioned in the International Maritime
Organization IMO’s safety and security Conventions relevant to ports, such as
International Ship and Port Facility Security Code (ISPS) or International Safety
Management Code (ISM). Apart from IMO’s Interim guidelines on cybersecurity
published in 2016 in Maritime Safety Committee’s (MSC) Circular MSC.1/Circ.1526,
there are no supranational guidelines how to tackle the issue.
However, IMO is making cyber risk management onboard ships mandatory as
of 1 January 2021, as cited in Resolution MSC.428(98) on Maritime Cyber Risk
Management in Safety Management Systems adopted in June 2017. The resolu-
tion states that an approved safety management system should take cyber risk
management into account in accordance with the objectives and requirements
of the ISM Code. Based on the recommendations in MSC.1/Circ.1526 Guidelines
on maritime cyber risk management, the resolution confirms that existing risk
management practices should be used to address the operational risks arising
from the increased dependence on cyber enabled systems.
Thus, the gap of knowledge on cybersecurity in ports identified above merits
further investigation on the topic. For this purpose, this paper reviews the key
elements and aspects of cybersecurity with the focus on ports. A conceptual anal-
ysis is conducted through a comprehensive literature review aiming to address
the following research question: ”How is cybersecurity perceived in ports?”
The paper is divided as follows: Section 2 contains a discussion regarding security
from the perspective of ports. Section 3 elaborates the key concepts and issues
related to cybersecurity. Section 4 provides insights on cybersecurity in ports.
Section 5 sums up the results and draws conclusions.
2 Security Aspects in Ports
Risk, threat and security are essential concepts when scanning any business
environment. By definition, a risk is a likelihood of an event with potentially
either positive or negative consequences (Prezelj and Ziberna, 2013). According
to Merriam-Webster dictionary a threat is ”an expression of intention to inflict evil,
345
injury, or damage” (Merriam-Webster, 2017). Security is a concept with multidi-

mensional meanings (Brooks, 2010), which is often addressing intentional threats
in contrast to ones with accidental or natural origin (Craigen, Diakun-Thibault,
and Purse, 2014). A difference between risk and security is that the latter involves
also uncertainty (Marlow, 2010).
Maritime community, with its operating field encompassing around the globe, is
susceptible to various types of threats. This turns the focus on maritime security,
a concept of which has recently gained buzzword status among the policymakers
(Bueger, 2015). Maritime security can be seen either as a state of security within the
maritime domain or as a vehicle to mitigate risk of threats, for example, terrorism,
piracy or smuggling, taking place not only on traditional sea or port locations, but
also along the entire supply chain (Helmick, 2008; Germond, 2015).
Characteristics of a port make it challenging object from the security perspective.
A port is a complex and multipart organization with institutions and functions
crossing multiple layers (Baltazar and Brooks, 2007). A port usually consists of
a port authority, port superstructure, for example cranes and conveyors, and
infrastructure, loading and unloading operations, storage facilities, and intra-port
operations (Brooks and Cullinane, 2007; Meersman and Van de Voorde, 2010).
From the national security perspective, ports, together with energy systems, trans-
port infrastructure, health industry, and water supply facilities, comprise critical
infrastructure (Ho and Ho, 2006; Prezelj and Ziberna, 2013). Critical infrastructure
refers to facilities, networks, and assets, which are essential in terms of citizens’
health, safety, security, and economic well-being as well as effective functioning
of society (Carrapico and Barrinha, 2017).
Typical threats of ports include, but are not limited to, financial losses, theft of
cargo or information, and strikes or system malfunctions that can compromise
the operations of a port (Ho and Ho, 2006; Loh and Thai, 2015). Moreover, in
light of the recent surge of activity, threat of terrorism should not be neglected.
One scenario suggests that terrorist may target or use ships at sea or in ports
as weapon to attack passengers and personnel of ports (Eski, 2011). Owing to
technological development, a new kind of threats have emerged as digitalization
has enabled threats coming through cyberspace to become reality (Rittinghouse
and Hancock, 2003; Geers, 2009; Miron and Muita, 2014).
Port security related policymaking is traditionally driven by global shocks. For
example, repercussions of 9/11 terror attacks led to global adoption of IMO’s
ISPS Code. The ISPS aims to enhance maritime security both on ships and in
346
3 Key Concepts and Issues Related to Cybersecurity
ports (Pinto and Talley, 2006; Thai and Grewal, 2007). Adoption of such initia-
tives comes with a price, of which are usually put on shipper’s account (Dekker
and Stevens, 2007). However, policies pursued by the major players persists
fragmented (Papa, 2013). For example, the USA adopted national policies such
as Container Security Initiative, which differs from global standards (Marlow,
2010).
The literature behind cyber-related issues is arguably embedded with various

cyber-prefixed concepts, of which are often used interchangeably (Bayuk, et al.,
2012). In order to make clarity in this complex field, the first part the Section
provides a conceptual image of the process involving concepts related to cyberse-
curity. The remainder of the Section contains elaborate definitions of the concepts
in question.
3.1 Conceptual Illustration
Figure 1 provides a simplified process description of the relevant concepts and

their mutual relationships. In a nutshell, conceptual description of the role of
cybersecurity (C) is as follows. All action is taking place in cyberspace (A), where
system (B; here e.g. the IT systems of a port community) is located. It is protected
by cybersecurity (C). System vulnerabilities (D) together with existing cyberthreats
(F) and the level of cybersecurity (C) comprise the level of cyberrisk (E) at any
given time. In case cybersecurity (C) is not at an adequate level, cyberrisk (E) may
materialize through a cyberattack (G), which targets the system (B) through an
identified vulnerability (D). In practice, a cyberattack (G) can be considered as a
materialized cyberthreat (F), which contains also specific technical methods to
inflict damage.
3.2 Definitions
According to Rantapelkonen and Kantola (2013, p.25) cyberspace is ”the collection

of computing devices connected by networks, in which electronic information is
stored and utilized, and communication takes place”.
347
Merriam-Webster dictionary defines cybersecurity as ”measures taken to protect

a computer or computer system (as on the Internet) against unauthorized access
or attack” (Merriam-Webster, 2017). In other words, the objective of cybersecurity
is a stable condition, where cyberspace is trusted and protected. At this point,
there is also sufficient capacity to proactively control and sustain cyberthreats
(Ministry of Defence of Finland, 2013; Limnéll, Majewski and Salminen, 2015).
Vulnerability refers to a feature or weakness of a computer or data system’s design,
integration, and maintenance (Maurushat, 2013). Vulnerability can either be direct
such as weak passwords that lead to unauthorized access, or indirect such as the
absence of network segregation (IMO, 2016a). Spotting the vulnerabilities requires
great precision as it is estimated that over 90 per cent of attackers are familiar with
the vulnerabilities of their targets (Afful-Dadzie and Allen, 2014; Loukas, 2015).
According to Maurushat (2013) vulnerabilities can be divided into three categories:
1) known vulnerability, 2) zero-day attack, and 3) future threat.
A known vulnerability is noticed in public through some form of communication
such as publication, and refers to failure of existing paradigms for recognizing,
reacting to or mitigating vulnerabilities. A zero-day attack refers to utilization of
a security vulnerability on the same day, when it becomes generally known. A
future threat means a condition that could end in harm as a consequence of a
formerly unknown security vulnerability. Vulnerabilities are increasing e.g. due to
i) shift of society towards on ubiquitous and automated computing environment;
and ii) increased utilization of the Internet (Lewis, 2002).
Cyberrisk refers to a variety of different sources of risk affecting the information
and technology assets of a firm (Biener, Eling and Wirfs, 2015). In more detail,
realized risks may result financial losses, disruption or damage to the reputation
of an organization from some sort of failure of its information technology systems
(IRM, 2014).
The growth of digitalization has entailed increase in frequency, sophistication
and scope of cyberthreats (Chertoff, 2008). A cyberthreat refers to a malicious
attempt in cyberspace, which aims to damage or interrupt a computer network or
system (Boyes, Isbell and Luck, 2016). Five basic types of cyberthreats are hack-
tivism, cybercriminality, cyberespionage, cyberterrorism and cyberwar (table 1).
Each of them has their individual features relating to actors involved, as well as
motivations and objectives behind actions.
Hacktivism refers to operations in cyberspace that make use of various hacking
techniques to invade into web pages and on computers, and create pressure
348
Table 1: Types and elements of cyberthreats

Cyberthreats Actors Motivations Objectives
Hacktivism Hacktivists Egoism Attention
Hackers Political Disruptions
Individuals Reputation Knowledge
Cybercriminality Individuals Economical Cargo
Industrial spies Informational Digital assets
Organized crime Organizational data
Cyberespionage Industrial spies Ideological Digital assets
Governments Informational Knowledge
Organized crime Political Organizational data
Cyberterrorism Governments Ideological Disruptions
Political National institutions
Terrorists Religious Critical Infrastructure
Social
Cyberwar Governments Egoism Military
Political National
Terrorists Religious Critical
Social
349
on a certain object (Limnéll, Majewski and Salminen, 2015; Boyes, Isbell and
Luck, 2016). A hacker uses own quick programming skills for invading into a
computer network file and seeks recognition for his/her technological capabilities
(Christou, 2016). Hackers can be divided into three different groups (Rittinghouse
and Hancock, 2003; Kapto, 2013):
1. White-hat hacker aims to promote security with his/her actions
2. Grey-hat hacker, often with criminal background, seeks gaps and vulner-
abilities
3. Black-hat hackers, i.e. a hacktivist, has criminal intentions
Cybercriminality refers to criminal activities that involve computer and infor-
mation systems either as a primary tool or as a primary target (Christou, 2016;
Carrapico and Barrinha, 2017). The aim is to gain financial benefits or to in-
flict personally motivated harm such as revenge or bullying (Gross, Canetti and
Vashdi, 2017). The economic benefit of cybercriminality can include criminal dam-
age, robbery of cargo, or identity thefts (European Commission, 2013; Boyes, Isbell
and Luck, 2016). Cybercriminality can be divided into four categories (Limnéll,
Majewski and Salminen, 2015; Luppicini, 2014):
1. Actions endangering confidentiality, integrity and availability of data and

systems
2. Forgery or identity thefts
3. Illicit gambling or spreading false information
4. Copyright or brand violations
Cyberespionage refers to illegal access to secret and delicate information such

as company strategy, private information, or intellectual capital, and it aims for
getting competitive advantage (Rittinghouse and Hancock, 2003; Boyes, Isbell
and Luck, 2016). Five different losses can be seen as a consequences of cyberes-
pionage (Platt, 2011; Fitzpatrick and Dilullo, 2015):
1. Loss of intellectual property, business and customer information

2. Extra costs due to interrupted business plans and competitive exercises
3. Loss of profits and efficiency
4. Damage to company reputation
350
5. Increased IT related security costs
Cyberterrorism is defined by Limnéll, Majewski and Salminen (2015, p.131) as ”a

deliberate politically motivated attack against information, computer systems,
computer software, and databases in the form of a violent invasion by interna-
tional groups or secret agents”. Cyberterrorist is an individual, who is specialized
in hacking into computer systems and is competent in organizing individual cy-
berattacks on global networks (Kapto, 2013).
Cyberwar is a part of modern information war between nations, during which
cyberattacks are made against opponents computer networks, which are relevant
from the military perspective (Lewis, 2002; Ministry of Defence of Finland, 2011;
Kapto, 2013). Cyberwar employs malicious software and viruses to disable military
targets (Gross, Canetti and Vashdi, 2017).
In case cyberrisks are realized, a cyberattack will take place, which has the basic
elements of cyberthreats in relation to actors, motivations and objectives. The
exact methods used by cyberattackers vary, while the most common ones are
phishing, malicious software and Denial-of-Service attack. (Colesniuc, 2013; Cy-
berEdge Group, 2016). Phishing is an attempt to gain discrete information by
imitating a reputable enterprise or person in e-mail or other communication chan-
nel. Malicious software or malware is a harmful program to steal, encrypt, delete
or change data, hijack or monitor users of target computer (Kendrick, 2010). A
Denial-of-Service (DoS) attack is an attempt to overtake a network by blocking it
with huge amount of communication (Fok, 2015).
351
Figure 1: Simplified process chart of concepts related to cybersecurity
352
4 Cybersecurity in Ports
4 Cybersecurity in Ports
ENISA’s study in 2011 on cybersecurity in maritime transport and ports identified

a clear lack of awareness on cybersecurity issues (ENISA, 2011). However, this has
not resulted in any European-wide strategy or coordinated action on the topic. As
mentioned in Section 2, critical infrastructure, including ports, constitutes a likely
target for cyberattacks given its significance on the functionality of societies. In
addition, what makes ports particularly vulnerable to cyberthreats relates to their
basic characteristics: dependency of data systems, handling massive volumes of
cargo or passengers , high monetary values, immense number of transactions,
numerous stakeholders involved, as well as non-transparent ownership of goods
end equipment (see e.g. Jensen, 2015). The geographical location may also be
influential as cyberattackers have been noted to target operators inside ports,
where the level of preparation tends to be low (Miron and Muita, 2014).
Potential consequences of cyberattacks against ports can be harmful in many

ways. The most common ones are scenarios, where cyberattackers gain access to
one or more of the following: i) overtake control of a ship, ii) shut down the entire
port, iii) delete or alter operational data, or iv) access to delicate information
(CyberKeel, 2014).
In response to a growing pressure for countermeasures against cyberthreats, pol-

icymakers throughout the world have started the adoption of multilevel general
cybersecurity strategies. Examples of these is the United Kingdom’s National
Cyber Security Strategy 2016–2021 (UK, 2016) at a national level, and the Cy-
ber security strategy of European Union: an Open, Safe and Secure Cyberspace
(European Commission, 2013) at a supranational level.
Similarly, work has begun among maritime authorities and international organi-
zations to develop strategies and standards for port facilities and ships against
cyberthreats. However, there are some challenges in this process. When design-
ing maritime specific guidelines, the globalized nature of the business and large
number of stakeholders set requirements for policy development. For example,
operations of a large container shipping company can easily involve over 100 coun-
tries, and its fleet size be measured in several hundreds of vessels (Jensen, 2015).
Global coordination and standardization of practices are essential elements in
this regard. So far none of the maritime specific guidelines are not mandatory by
nature, which may hinder the adoption process.
353
In 2015, the United States Coast Guard (2015) introduced its Cyber strategy (for
critical maritime infrastructure). The Institution of Engineering and Technology
(IET) introduced in 2016 the Code of Practice (Boyes, Isbell and Luck, 2016).
In 2016, the Baltic and International Maritime Council (BIMCO), the International
Chamber of Shipping (ICS), INTERCARGO, INTERTANKO and the Cruise Lines In-
ternational Association (CLIA) published ”Guidelines on Cyber Security Onboard
Ships” (BIMCO, et al., 2016). The guidelines introduced a six-step approach, which
is dedicated to cybersecurity and cyberthreats:
1. Identification of external and internal cyberthreats

2. Identification of vulnerabilities
3. Assessment of risk exposure
4. Development of protection and detection measure
5. Establishment of contingency plan
6. Response to cybersecurity incidents
In 2016, IMO published the ”Interim Guidelines on Maritime Cyber Risk Manage-
ment”, which underlines that cyberrisk management should be complementary
to existing security and safety risk management requirements, like ISM and ISPS
Codes (IMO, 2016a). The objective of IMO’s guidelines is to keep cyberrisks at
a reasonable level by using multilevel approach that involves all relevant port
actors (IMO, 2016a).
By and large, the number of reported cyberattacks against ports has remained
on a very low level thus far. The only case, which has received wider attention,
was the attack against port of Antwerp in late-2013 (Boyes, Isbell and Luck, 2016).
The exact reasons behind absence of attacks can only be speculated. Similarly,
the number of attempts is fairly uncertain given that they may not be reported
or noticed. It should be remembered that the security situation is constantly
evolving — what was adequate yesterday may not hold today.
There is very limited amount of publicly available information about contempo-
rary cybersecurity related practices in ports, which is presumably due to discre-
tionary nature of the subject. However, initial empirical evidence from the Baltic
Sea Region indicate that neither ports nor regulation seem to be well prepared to
cyberthreats (Ahokas and Laakso, 2017).
354
5 Results and Conclusions
Moreover, it appears that other maritime industry sectors are not neither that
well prepared against cyberthreats. In mid-2017, there was a cyberattack against
the world’s largest shipping line, Maersk, which temporarily crippled the entire
company (Knowsler, 2017). The episode has explicitly showed that there is still
room for improvement in this sector as well.
The notion of ports lesser role in terms of cybersecurity considerations receives
support when looking at academic literature, as the number of articles concerning
cybersecurity in ports is scarce. In addition, the topic appears recent, i.e. pub-
lished after 2011. Apart from two peer-reviewed journal articles by Kouwenhoven
(2014) and Jones (2015), other dedicated reports are predominantly industry,
policy or consultancy papers (Ahokas and Kiiski, 2017). The topic’s novelty is con-
vergent with the body of literature, while the scarcity observation is contrasting
given the reported influx of studies covering cybersecurity and maritime security
in general (Germond, 2008; Jensen, 2015).
5 Results and Conclusions
The recent growth of cyberattacks and subsequent increased awareness of cy-

bersecurity, in which ports appeared to be somewhat neglected, provided the
ultimate inspirations for this paper. A conceptual analysis, which was based on
comprehensive literature review, was conducted. Objective of the paper was to
describe cybersecurity in ports by answering to specific research question: ”How
is cybersecurity perceived in ports?” This was approached by first establishing
port’s dual role in security as being part of both maritime and national security
considerations. After this, the relevant terminology and concepts related to cyber-
security were scrutinized. Finally, state-of-the-art situations about cybersecurity
in ports were mapped.
The results show that regardless of the growing awareness of the issues, much
work needs to be done in order to mitigate the cyberthreats in ports. The matter
is both novel and of great urgency as cyberattacks are becoming more common
with pervasive impacts on society. Maritime sector and ports in particular are
no exception in this regard as recent attacks against Maersk and port of Antwerp
have showed.
There is limited amount of information available about the contemporary cy-
bersecurity related practices in ports, which presumably is due to discretionary
355
nature of the subject. However, there are indications suggesting that the ports
current level of preparation and regulations are not adequate.
Over the past five years, policymakers and other stakeholders have become ac-
tively engaged in cyberthreats by adopting cybersecurity strategies and guidelines,
for example, IMO (2016a) and BIMCO, et al. (2016). However, mandatory global
standards are yet to be introduced, which, among other things, could expedite the
adoption process. Owing to the global scale and large number of parties involved,
coordinated efforts are needed to ensure adoption of adequate practices and
regulations throughout the industry. This supports Helmick’s (2008) call for exten-
sive cybersecurity framework. Here, IMO’s Resolution adopted in June 2017 to
make cyber risk management onboard ships mandatory as of 1 January 2021 is a
significant, yet belated step ahead. Similar steps for seaports are still pending.
Unlike popular research streams of maritime security and cybersecurity in general,
the port environment in a cyber context appears to have received scant exposure.
Only few journal papers appear to have dealt with the topic, while the majority of
publications consist of consultancy or policy related papers.
The terminology behind cybersecurity appears far from being harmonized as the
use of various concepts with different meanings is common (see also IMO, 2016b).
This finding supports previous arguments by Craigen, Diakun-Thibault and Purse
(2014) and Hult and Sivanesan (2013). Especially the relationship between cy-
berthreat and cyberattack is a cumbersome (Kadivar, 2014; Loukas, 2015). In order
to provide input to this issue, a conceptual map was introduced that delineates
the relationships between different concepts.
This paper contains limitations that needs to be taken into consideration. The
major limitation compounds from the novelty of the topic as there is only limited
amount of publications and empirical data available. Future research should
study port cybersecurity strategies in more detail, for example, by establishing
a suitable typology and/or a taxonomy on these preparation plans. In addition,
more information is needed about how these strategies have been implemented
empirically and how effective they are in terms of mitigating cyberthreats.
References
Afful-Dadzie, A. and T. T. Allen (2014). “Data-Driven Cyber-Vulnerability Maintenance Policies”. In:
Journal of Quality Technology 46.3, pp. 234–250.
356
References
Ahokas, I. and K. Laakso (2017). Deplhi study on safety and security in the Baltic Sea Region ports.
URL: https://blogit.utu.fi/hazard/.
Ahokas, J. and T. Kiiski (2017). Cybersecurity in ports.
Baltazar, R. and M. R. Brooks (2007). “Port Governance, Devolution and the Matching Framework: a
Configuration Theory Approach”. In: Devolution, Port Governance and Port Performance. Ed. by
M. R. Brooks and K. Cullinane. London: Elsevier, pp. 379–403.
Baltic and International Maritime Council (BIMCO), International Chamber of Shipping (ICS),
INTERCARGO, INTERTANKO, and Cruise Lines International Association (CLIA) (2016). The
Guidelines on Cyber Security Onboard Ships. Version 1.1. Bagsvaerd.
Bayuk, J. L., J. Healey, P. Rohmeyer, M. H. Sachs, J. Schmidt, and J. Weiss (2012). Cyber Security
Policy Guidebook. New Jersey: Wiley.
Biener, C., M. Eling, and J. H. Wirfs (2015). “Insurability of Cyber Risk: an Empirical Analysis”. In:
The Geneva Papers 40, pp. 131–158.
Boyes, H., R. Isbell, and A. Luck (2016). Code of Practice: Cyber Security for Ports and Port Systems.
Stevenage.
Brooks, D. J. (2010). “What is Security: Definition through knowledge categorization”. In: Security
Journal 23.3, pp. 225–239.
Brooks, M. R. and K. Cullinane (2007). “Governance Models Defined”. In: Devolution, Port Gover-
nance and Port Performance. Ed. by M. R. Brooks and K. Cullinane. London: Elsevier, pp. 405–
435.
Bueger, C. (2015). “What is maritime security?” In: Marine Policy 53, pp. 159–164.
Carrapico, H. and A. Barrinha (2017). “The EU as a Coherent (Cyber)Security Actor?” In: Journal of
Common Market Studies, pp. 1–19.
Chertoff, M. (2008). “The cybersecurity challenge”. In: Regulation and Governance 2.4, pp. 480–484.
Colesniuc, D. (2013). “Cyberspace and Critical Information Infrastructure”. In: Informatica Econom-
ica 17.4, pp. 123–132.
Commerce (ICC), I. C. of (2015). Cyber Security Guide for Business. Paris.
Craigen, D., N. Diakun-Thibault, and R. Purse (2014). “Defining Cybersecurity”. In: Technology
Innovation Management Review 4.10, pp. 13–21.
CyberKeel (2014). Maritime Cyber-Risks. Copenhagen.
Cybersecurity in the European Union: Resielience and Adaptability in Governance Policy (2016).
Hampshire: Palgrave Macmillan.
Cybersecurity Strategy of the European Union: An Open, Safe and Secure Cyberspace (2013). URL:
https://ec.europa.eu/digital-single-market/en/news/eu-cybersecurity-plan-protect-open-
internet-and-online-freedom-and-opportunity-cyber-security.
Defense, M. of (2011). Security Strategy for Society - Government Resolution 16.12.2010. Helsinki:
Ministry of Defense of Finland.
Defense, M. of (2013). Finland’s Cyber Security Strategy - Government Resolution 24.1.2013. Helsinki:
Ministry of Defense of Finland.
Dekker, S. and H. Stevens (2007). “Maritime security in the European Union — empirical findings on
financial implications for port facilities”. In: Maritime Policy & Management 34.5, pp. 485–499.
Eski, Y. (2011). “”Port of Call”: Towards a criminology of port security”. In: Criminology & Criminal
Justice 11.5, pp. 415–431.
Fitzgerald, M., N. Kruschwitz, D. Bonnet, and M. Welch (2013). Embracing Digital Technology: A
New Strategic Imperative. 2, pp. 1–12.
357
Fitzpatrick, W. M. and S. A. Dilullo (2015). “Cyber Espionage and the S.P.I.E.S. Taxonomy”. In:
Competition Forum 13.2, pp. 307–336.
Fok, E. (2015). “Cyber Security Challenges: Protecting Your Transportation Management Center”.
In: Institution of Transportation Engineers Journal 85.2, pp. 32–36.
Geers, K. (2009). “The Cyber Threat to National Critical Infrastructures: Beyond Theory”. In: Infor-
mation Security Journal: A Global Perspective 18.1, pp. 1–7.
Germond, B. (2015). “The geopolitical dimension of maritime security”. In: Marine Policy 54,
pp. 137–142.
Goldby, M. (2008). “Electronic bills of lading and central registeries: what is holding back progress?”
In: Information & Communications Technology Law 17.2, pp. 125–149.
Gross, M. L., D. Canetti, and D. R. Vashdi (2017). “Cyberterrorism: its effects on psychological
well-being, public confidence and political attitudes”. In: Journal Of Cybersecurity 3.1, pp. 49–
58.
Group, C. (2016). 2016 Cyberthreat Defense Report - North America, Europe, Asia Pacific, and Latin
America. Annapolis.
Helmick, J. S. (2008). “Port and maritime security: A research perspective”. In: Journal of Trans-
portation Security 1.1, pp. 15–28.
Ho, M. W. and K. H. ( Ho (2006). “Risk Management in Large Physical Infrastructure Investments:
The Context of Seaport Infrastructure Development and Investment”. In: Maritime Economics &
Logistics 8.2, pp. 140–168.
Hult, F. and G. Sivanesan (2013). “Introducing cyber”. In: Journal of Business Continuity & Emergency
Planning 7.2, pp. 97–102.
International Maritime Organization (IMO) (2016a). IMO Multilingual Glossary on Cyberterms. Lon-
don.
International Maritime Organization (IMO) (2016b). Interim Guidelines on Maritime Cyber Risk
Management. London.
Jensen, L. (2015). “Challenges in Maritime Cyber-Resilience”. In: Technology Innovation Manage-
ment Review 5.4, pp. 35–39.
Jones, S. (2015). “Addressing cyber security risks at ports and terminals”. In: Port Technology
International Journal 62.
Kadivar, M. (2014). “Cyber-Attack Attributes”. In: Technology Innovation Management Review 4.11,
pp. 22–27.
Kapto, A. S. (2013). “Cyberwarfare: Genesis and Doctrinal Outlines”. In: Herald of the Russian
Academy of Science 83.4, pp. 357–364.
Kendrick, R. (2010). Cyber Risks for Business Professionals: A Management Guide. Cambridgeshire:
IT Governance Publishing.
Knowsler, G. (2017). Maersk cyber attack forces carrier to put cargo bookings on hold. URL: http:
//fairplay.ihs.com/article/4288541/maersk-cyber-attack-forces-carrier-to-put-cargo-bookings-
on-hold.
Kouwenhoven, N. (2014). “The implications and threats of cyber security for ports”. In: Port Tech-
nology International Journal 61.
Lasi, H., P. Fettke, H.-G. Kemper, T. Feld, and M. Hoffmann (2014). “Industry 4.0”. In: Business &
Information Systems Engineering 6.4, pp. 239–242.
Lewis, J. A. (2002). “Assessing the Risks of Cyber Terrorism, Cyber War and Other Cyber Threats”.
In:
Limnéll, J., K. Majewski, and M. Salminen (2015). Cyber Security for Decision Makers. Docendo.
358
References
Loh, H. S. and V. V. Thai (2015). “Management of disruptions by seaports: preliminary findings”. In:
Asia Pacific Journal of Marketing and Logistics 27.1, pp. 146–162.
Loukas, G. (2015). Cyber-Physical Attacks: A Growing Invisible Threat. Oxford: Elsevier Science.
Luppicini, R. (2014). “Illuminating the Dark Side of the Internet with Actor-Network Theory: An
Integrative Review of Current Cybercrime Research”. In: Global Media Journal - Canadian Edition
7.1, pp. 35–49.
Marlow, P. B. (2010). “Maritime security and update of key issues”. In: Maritime Policy & Management
37.7.
Maurushat, A. (2013). Disclosure of Security Vulnerabilities: Legal and Ethical Issues. London:
Springer.
Meersman, H. and E. Van de Voorde (2010). “Port Management, Operation and Competition: a
Focus on North Europe”. In: The handbook of Maritime Economics and Business. Ed. by C. T.
Grammenos. London: Lloyd’s List, pp. 891–906.
Merriam-Webster Dictionary (2017). URL: https://www.merriam-webster.com/.
Miron, W. and K. Muita (2014). “Cybersecurity Capability Maturity Models for Providers of Critical
Infrastructure”. In: Technology Innovation Management Review 4.10, pp. 33–39.
Papa, P. (2013). “US and EU strategies for maritime transport security: A comparative perspective”.
In: Transport Policy 28, pp. 75–85.
Pinto, C. A. and W. K. Talley (2006). “The Security Incident Cycle of Ports”. In: Maritime Economics &
Logistics 8.3, pp. 267–286.
Platt, V. (2011). “Still the fire-proof house? An analysis of Canada’s cyber security strategy”. In:
International Journal: Canada’s Journal of Global Policy 67.1, pp. 155–167.
Prezelj, I. and A. Ziberna (2013). “Consequence-, time- and interdependency-based risk assess-
ment in the field of critical infrastructure”. In: Risk management 15.2, pp. 100–131.
Rantapelkonen, J. and H. Kantola (2013). “Insights into Cyberspace, Cyber Security, and Cyberwar
in the Nordic Countries”. In: The Fog of Cyber Defense. Ed. by J. Rantapelkonen and M. Salminen.
Helsinki: National Defense University, pp. 24–36.
Risk Management (IRM)l, I. of (2014). Cyber Risk: Executive Summary. London.
Rittinghouse, J. and W. M. Hancock (2003). Cybersecurity Operations Handbook. London: Elsevier
Digital Press.
Thai, V. v. and D. Grewal (2007). “The Maritime Security Management System: Perceptions of the
International Shipping Community”. In: Maritime Economics & Logistics 9.2, pp. 119–137.
Trade, U. N. C. on and Development (2016). Review of Maritime Transport 2016. New York: United
Nations Publication.
United Kingdom, H. G. (2016). National Cyber Security Strategy 2016-2021. London.
Urcioli, L. (2015). “Cyber-Resilience: A Strategic Approach for Supply Chain Management”. In:
Technology Innovation Management Review 5.4, pp. 13–18.
359
Smart Supply Chain Risk Management -
A Conceptual Framework
Florian Schlüter1 , Michael Henke1
Screening existing literature on Supply Chain Risk Management (SCRM) shows

that only sporadic attention is paid on real data driven SCRM. Most tools and
approaches lead to an expert knowledge based SCRM. Due to the arising topic of
digitalization in supply chains, leading to Industry 4.0 (I4.0), there is huge poten-
tial in building a data driven, smart SCRM. To speed up research in this direction it
is worthwhile to define a new research framework giving direction. To create a
consistent framework and define smart SCRM in more detail a literature review will
take place to select appropriate dimensions like SCRM phases, readiness stages of
Digitalization/ I4.0 and SC perspectives describing the degree of SC collaboration.
Afterwards the SCRM and I4.0 dimensions will be put into focus describing what
impact I4.0 will have on SCRM leading to future requirements. The new frame-
work serves as a basis for future SSCRM research. It helps to categorize research
projects through multiple dimensions and to identify potential research gaps.
The developed SSCRM requirements framework is a practical tool guiding the
requirement specification when designing a company specific SSCRM system.
Keywords: Supply Chain Risk Management; Industry 4.0; Digitalization
361
Smart Supply Chain Risk Management - A Conceptual Framework
1 Introduction
There are many example cases in the literature, like Ericsson (Chopra and Sodhi,
2004; Norrman and Jansson, 2004), Toyota (Pettit, Crocton and Fiksel, 2013) and
Land Rover (Tang and Tomlin, 2008), which show that a supply chain disrup-
tion and a resulting glitch can have serious cascading effects on all supply chain
members and their performance.
To lower the impact of such glitches firms usually establish a supply chain risk
management (SCRM) which became a critical supply chain management disci-
pline in the past due to the increasing number of events causing supply chain
disruptions (Hillman and Keltz, 2007). In the past usually historical company and
external data are used in the traditional SCRM concept (Güller, et al., 2015). The
limitation of these practices is that information is not available timely enough
and they don’t provide a real-time view of the entire supply chain operations
(Güller, et al., 2015). Faisal, Banwet and Shankar (2006) have empirically shown
the benefit of information sharing of supply chain members to understand the
different risks which could have an impact on the supply chain. While supply
chain risk information has been identified as crucial, the importance of a firm’s
information processing capability to its SCRM effort has received little attention
in the literature (Fan, et al., 2016). A system which processes SC risk (SCR) infor-
mation would help firms to respond in a timely manner (Fan, et al., 2017) and
enables recognition, analysis and assessment of negative trends to manage risks
inside and outside of the SC (Zweig, et al., 2015).
Due to the arising topic of digitalization in supply chains (Pfohl, Yahsi and Kur-
naz, 2015; Kersten, et al., 2016) there is huge potential in building a data driven,
smart SCRM (Schröder, Indorf and Kersten, 2014). Available real-time information
and data-processing tools bring new opportunities for companies to react more
quickly to changing conditions within the supply chain (Güller, et al., 2015). The
new principles and components of Industry 4.0 (I4.0) (e.g. Hermann, Pentek and
Otto, 2016; Siepmann, 2016a; 2016b) lead also to a SCRM based on different prin-
ciples compared to classical SCRM (Schröder, Indorf and Kersten, 2014; Schlüter,
Diedrich and Güller, 2017). Therefore and to speed up research it is worthwhile
to define Smart Supply Chain Risk Management (SSCRM) as a sub-research field
within the field of SCRM and to come up with a new research framework giving
direction.
The purpose of this paper is to create a consistent framework based on existing
literature, serving as a basis for future SSCRM research. It helps to categorize
362
2 Research Overview
research projects to identify potential research gaps. Afterwards a guiding design

instrument for individual SSCRM requirement definitions will be derived. This
leads to a practical tool supporting the design process for a company specific
SSCRM system.
After a research overview in section 2 the research questions will be defined. In
section 3 the framework will be developed and SSCRM will be defined in more
detail in section 4. The paper closes in section 5 with a conclusion, an outlook for
further research and managerial implications.
2 Research Overview
For an appropriate definition of a SSCRM research framework it is necessary to

define SCRM and give insights about digitalization and related concepts. The
section ends with an overview about related research and the research questions
which will be answered throughout the rest of the paper.
2.1 Supply Chain Risk Management
SCRM can be seen as an emerging critical and cross-functional discipline between

Supply Chain Management (SCM), corporate strategic management and Enter-
prise Risk Management (ERM) (Hillman and Keltz, 2007; Zsidisin and Ritchie, 2009).
In their literature review, Ho, et al. (2015) stated that the proposed definitions
of SCRM in the literature usually focus on specific elements of SCRM and do not
span the SCRM processes completely or differ in their SCRM methods and types
of events. Given this, the authors also follow Ho, et al. (2015) in their definition of
SCRM as: “an inter-organizational collaborative endeavor utilizing quantitative
and qualitative risk management methodologies to identify, evaluate, mitigate
and monitor unexpected macro and micro level events or conditions, which might
adversely impact any part of a supply chain”.
2.2 Digitalization and related concepts
A digitalized SC makes potential risks visible, allows companies to monitor mate-

rial flows in real time and to develop future plans (Goh, et al., 2013). The integra-
363
tion of Cyber-Physical-Systems (CPS) in existing or new supply chain processes

leads to a convergence of the physical world and the virtual world (Wan, Cai
and Zhou, 2015) and are the foundation of an I4.0 (Bischoff, et al., 2015). CPS
are physical objects, equipped with embedded systems, sensors and actuators
adding intelligence and the ability for self-control, cross-linking with other CPS
and for interaction with their environment (Bischoff, et al., 2015). Beside the term
Digitalization there are other definitions in the literature with a similar meaning,
like Industrial Internet, Internet of Things, Integrated Industry, Smart Industry,
Smart Manufacturing and I4.0 (Hermann, Pentek and Otto, 2016). Especially the
term Industry 4.0 or Industrie 4.0 is widely used in German speaking literature
and slowly makes its way into Anglo-Saxon literature (e.g. Wan, Cai and Zhou,
2015 or Qin, Liu and Grosvenor, 2016). The main characteristic of the I4.0 is auton-
omization based on cross-linked systems which communicate with each other
via Internet (Roth, 2016). For this paper the term Digitalization is defined as a nec-
essary action on the road to I4.0 and will be used synonymously at some points.
More information about I4.0, Digitalization and its components can be found in
the literature (Bauernhansl, ten Hompel and Vogel-Heuser, 2014; Bischoff, et al.,
2015; ten Hompel and Henke, 2017).
2.3 Smart Supply Chain Risk Management
The integration of CPS into supply chains leads to a smart supply chain manage-
ment, which combines multiple independent data analytics models, historical
data repositories, and real-time data streams (Wang and Ranjan, 2015). Through
this embedded intelligence, supply chain management moves from support-
ing decisions to delegating them and, ultimately, to predicting which decisions
need to be made (Butner, 2010). The main drivers for the digitalization of supply
chain processes are typically an increase in flexibility and reaction rate of indus-
trial/logistic systems (ten Hompel and Henke, 2017). Another perspective is to
improve the supply chain robustness by using this available data from digitalized
supply chain processes and CPS in SCRM, leading to a smart SCRM. Making the
supply chain smarter from a risk management perspective can be described as
“SCRM digitalization”, thus as: “the integration of technology (sensors, actors,
connectivity, analytics) along supply chain processes to improve supply chain
risk identification, analysis, assessment, mitigation and monitoring through pro-
cessing real time supply chain risk information – which comprises supply chain
risk information sharing and analysis” (Schlüter, Diedrich and Güller, 2017).
364
2 Research Overview
2.4 Existing work
In the literature various works for both SCRM and Digitalization can be found
which try to guide researchers as well as practitioners in their effort to find and
define new research gaps and projects.
In SCRM most of this work is done over the past years via structured literature
reviews (SLRs), which are usually based on statistical analysis of the existing
literature at the time of release. They are either focusing more on specific SCRM
sub-topics (e.g. Tang and Musa, 2011; Fahimnia, et al., 2015; Heckmann, Comes
and Nickel, 2015; Kilubi and Haasis, 2016) to show there are specific research
gaps or they are more generalized to show multiple research gaps in different sub-
topics of SCRM (e.g. Jüttner, Peck and Christopher, 2003; Ritchie, 2007; Singhal,
Agarwal and Mittal, 2011; Ho, et al., 2015). Also using the same steps, Schlüter,
Diedrich and Güller (2017) performed a literature review to identify literature
mentioning how digitalization and the usage of data driven tools will somehow
affect SCRM in the near future. While doing so these publications are not focusing
on creating a SSCRM, their main purpose is to bring new/better tools into the
classic SCRM procedure. A similar publication comes from Schröder, Indorf and
Kersten (2014) who postulate briefly how I4.0 will change the steps of SCRM and
they propose new risks arising in the I4.0. The above mentioned literature usually
focuses on different variations of a SCRM framework, describing the process of
SCRM while none attention is paid defining a research framework to structure
work in the field of SCRM. To the authors knowledge only the three-dimensional
framework by Lindroth and Norrman (2001) can be used to structure work within
the field of SCRM. The framework has been further developed and extended by
Norrman and Lindroth (2004).
Within the field of Digitalization some framework approaches to cluster current
and future research are available. Those are usually focusing on I4.0 in general.
Pfohl, Yahsi and Kurnaz (2015) designed a matrix to categorize research within
four research-fields, based on the two dimensions “Confirmatory Quantitative
vs. Exploratory Qualitative” and “Analysis on management-level vs. Analysis on
technology- and process-level”. To give direction for practitioners and researchers
the “Dortmund Management-Model for Industry 4.0” by Henke, establishes and
formalizes “work-clusters” for transforming value creating activities into the I4.0
(ten Hompel and Henke, 2017).
A recent publication of Lu (2017) gives insights about the development of publi-
cation numbers of I4.0 literature and the author has clustered 88 selected publi-
365
cations in categories like “Concepts and perspectives of Industry 4.0”, “Key tech-
nologies of Industry 4.0” or “Applications of Industry 4.0”.
Until now there is no known research framework available in the literature con-
necting SCRM and Digitalization into a SSCRM research framework, which helps
to classify above mentioned research and to identify new research directions
and helps to develop a guiding tool for developing company specific SSCRM
requirements.
This leads to the following research questions:
RQ1: How does a research framework based on those dimensions look like?.
RQ2: What are appropriate fist definitions of SSCRM maturity steps and role
definitions?
3 Developing a research framework
The advanced framework of Norrman and Lindroth (2004) serves as basis for the
SSCRM.
In one dimension Norrman and Lindroth (2004) introduced five units of analysis
in SCRM: single logistical activity within a company (single logistics); logistical
activities of the whole company (company logistics); logistical activities between
two companies (dyads logistics); logistical activities between companies linked
to a chain (supply chain logistics) and logistical activities between companies
linked to a network (supply chain network). These different scopes of SCM should
be included to reflect the different levels of collaboration presented in the litera-
ture.
For the second dimension Norrman and Lindroth (2004) present four SCRM stages
but they give no information on why these stages have been selected. Screening
existing literature (see section 2) shows that there is a diverse understanding.
For a comprehensive framework it is necessary to screen available literature
about SCRM steps and build a SCRM procedure based on the findings. A research
overview about paper explicitly describing models and frameworks of SCRM
comes from Ponis and Ntalla (2016). de Oliveira, et al. (2017) performed a similar
approach by screening 27 publications for SCRM steps and comparing them with
the ISO 31000-SCRM procedure (see e.g. Curkovic, Scannel and Wagner, 2013).
Based on their exhaustive literature review the following SCRM stages will be
366
3 Developing a research framework
Table 1: Industry 4.0 Maturity Stages (Schuh, et al., 2017)

Maturity stage Description
Computerisation Support through IT-Systems and worker will be disbur-
dened from repetitive work
Connectivity Systems are structured and connected
Visibility Digital Shadow available and management decisions
are data-based
Transparency Companies understand why things happen
Predictive capacity Companies know what might happened and decisions
are based on future scenarios
Adaptability Systems react and adapt autonomously
implemented in the SSCRM framework instead of the original stages of Norrman

and Lindroth (2004): risk identification (identification of risks and sources); risk
analysis (measurement of risk consequences and identification of risk factors);
risk assessment (evaluation of risks); risk treatment (proposal of strategies and
mitigation of risks) and risk monitoring (measurement of results, control of risks
and ongoing improvement process).
The last dimension of the original framework describes the type of risk: opera-
tional, tactical and strategical (Norrman and Lindroth, 2004).
Because a SSCRM connects principles of I4.0 and SCRM the development goes
along the same I4.0 maturity stages as mentioned in the literature. Recently the
German acatech – National Academy of Science and Engineering published a
study to provide companies with I4.0 maturity stages to help them identifying
their current maturity stage and also to achieve a higher stage in order to maximize
the economic benefits of I4.0 and digitalization (Schuh, et al., 2017). The maturity
stages are described in the table below (see table 1).
When speaking about I4.0, the focus is usually on the technological aspects and
the important role of CPS, but often neglected is the fact that CPS-based pro-
duction systems are socio-technical systems (Hirsch-Kreinsen, 2014), consisting
of a technical and social subsystem which are interlinked (Bostrom and Heinen,
1977). The social sub system focuses on the role of people using the technology
while the technical sub system focuses on the available technology and its role
(Bostrom and Heinen, 1977). Because future publications are not necessarily
367
Figure 1: SSCRM Research Framework
focusing on the technological aspects of SSCRM but more on the social aspects
and the dynamics between people and technology, the framework should also
reflect this.
To speed up research in SSCRM an additional research-outcome dimension is

required. The field of Design Science Research (DSR) has strongly formalized
possible research outcome in the form of artifacts. DSR has the goal to develop
practical solutions that can be used by professionals in their field (Lessard and Yu,
2012). A comprehensive list of how artifacts can look like comes from Hevner, et
al. (2004), who mention theories, frameworks, instruments, constructs, models,
methods and instantiations.
Combining the above mentioned dimensions delivers the following SSCRM frame-
work, which answers the first research question (see figure 1).
The proposed framework now allows the classification of existing and current
research regarding the field of SSCRM. A detailed explanation of the combined
368
4 SSCRM Requirements Framework
perspectives “Risk Management” and “Industry 4.0 Maturity” follows in the next
chapter.
To support practitioners in designing SSCRM systems requirements it is helpful

to define how the aforementioned I4.0 maturity stages become manifest within
SCRM. While the general descriptions of the I4.0 maturity stages (see section 3
and Schuh, et al., 2017) and SCRM (see section 3 and de Oliveira, et al., 2017) are
given, the authors defined how the traditional SCRM phases could look like in
SSCRM. The goal is to give practitioners a guiding instrument for deriving indi-
vidual requirements for SSCRM systems throughout the SCRM process, for the
technology as well as for the people using the technology.
Due to some similarities between SCRM steps and I4.0 maturity stages (e.g. Risk
Analysis and Transparency) it is necessary to specify the objectives within the
SCRM phases in an I4.0 context. Due to the digitalization of SC processes, future
SC can be understood as data points in a Digital Shadow (Leveling, et al., 2014),
leading to the following general SCRM objectives (just for orientation - degree of
automatization depends on SSCRM maturity level) (see table 2):
Based on this specification and in combination with the I4.0 maturity stage defi-
nitions a framework could be derived, giving overview about potential SSCRM
development stages (see table 3 – table 7). For the framework development pre-
vious work by Schlüter, Diedrich and Güller (2017) has been used to get an idea
how the traditional SCRM phases will change through digitalization. These ideas
have been further refined in brainstorming sessions and discussions with other
researchers for this working paper. In the future additional research will take
place to verify the authors’ ideas.
This first attempt of describing the roles of technology and people within a SSCRM
answers the second research question. As it can be seen in the first and second
maturity stage there are only marginal difference between the SCRM phases. The
reason is that accordingly to (Schuh, et al., 2017) in these stages there is not yet an
I4.0 environment and thus there is no infrastructure for a SSCRM available. The
SC IT systems serve more as general data sources for risk managers who use the
data supportive to traditional SCRM methods. After stage three the technology
gets smart and more flexible and can be fitted to SCRM tasks.
369
Table 2: Overview SCRM objectives in I4.0 context
SCRM phase Objective

Risk identification Identification of new risk relevant data points and iden-
tification of relations between these points
Risk analysis Identification of directions within the relations to create
cause-and-effect relationships
Risk assessment Realistic calculation of risks, based on the relationships
between data points and objective quantitative data
Risk treatment Identification of activities through data points and
proposing of suitable activities based on effect analysis
Risk monitoring Monitoring of known risks and risk relevant data points
as well as chosen activities to ensure risk reduction
370
Table 3: SSCRM-Framework – Identification

I4.0 Readiness Stage Technical System (Role) Social System (Role)
Computerisation Isolated usage of IT sys- Manager use the IT sys-
tems in SC processes al- tem records to identify
lows the collection of potential risks in addi-
structured and unstruc- tion to workshops and
tured process data; Risk expert interviews
relevant information can
be stored locally
Connectivity Connection of IT sys- Manager can use a
tems allows the uni- or broader range of SC
bidirectional exchange process data to identify
of process data; Risk potential risks in addi-
relevant information can tion to workshops and
be stored in a process expert interviews
wide data base
Visibility All SC processes are rep- The user has to recog-
resented and character- nize critical relations be-
ized through data points tween the data points
within a Digital Shadow and define them as risks
via real-time data by himself
Transparency Analytic systems recog- Found relations have to
nize relations between be checked for criticality
data points (undirected by the user
relations) and they
search for similar
relations
Predictive capacity Simulation of scenar- Changes of relations and
ios within the Digital future relations have to
Shadow shows future be checked for criticality
relations between by the user
data points and how
identified relations will
change
Adaptability System recognizes po- User monitors the re-
tential critical relations ports and uses the re-
between data points, ag- ported potential risky re-
gregates them to risks lations and aggregated
and reports them to the risks as basis for further
user assessment and treat-
ment 371
Table 4: SSCRM-Framework – Analysis

Computerisation Isolated usage of IT sys- Manager use different
tems in SC processes al- IT system records, in
lows the collection of addition to workshops
structured and unstruc- and expert interviews to
tured process data; Risk identify risk causes and
relevant information can patterns
be stored locally
Connectivity Connection of IT sys- Manager can use a
tems allows the uni- or broader range of SC
bidirectional exchange process data, in addition
of process data; Risk to workshops and expert
relevant information can interviews to identify
be stored in a process risk causes and patterns
wide data base
Visibility Every data point within The user has to interpret
each risk offers a range the data of each data
of qualitative, structured point to recognize cause-
and semi-structured and-effect relations (di-
data rected relations)
Transparency The system recognizes The user has to check
the cause-and-effect re- the data point ranking
lations based on the real- and chooses the points
time information from which should stay in fo-
each data point within cus for future risk treat-
and without the identi- ment
fied risk cluster; Data
points get ranked based
on their influence on
other points
Predictive capacity Forecasting of changes The user has to check
of direction of the cause- the improved data point
and-effect relations ranking and chooses the
points which should stay
in focus for future risk
treatment
Adaptability Based on previous capa- User checks the results
bilities the system gives for plausibility and uses
suggestions on which the results for further as-
372 data points should sessment and treatment
stay in focus for risk
treatment
Table 5: SSCRM-Framework – Assessment

Computerisation Isolated usage of IT sys- In addition to expert es-
tems in SC processes al- timations available pro-
lows the collection of cess data can be used
structured and unstruc- to calculate more reli-
tured process data; Risk able values for probabil-
relevant information can ity and impact
be stored locally
Connectivity Connection of IT sys- In addition to expert es-
tems allows the uni- or timations the available
bidirectional exchange broader range of process
of process data; Risk data can be used to cal-
relevant information can culate more reliable val-
be stored in a process ues for probability and
wide data base impact
Visibility Every data point within Assessment of risks
each risk offers a range based on the KPIs
of quantitative data and and comparison with
KPIs reference values
Transparency Connecting of KPIs Manual comparison of
and thus improved actual risk values with
calculation of impact; reference values
Calculation of proba-
bility based on a large
number of historical
and real-time data;
Development of new
risk measures where
necessary
Predictive capacity Prediction-based risk Manual comparison of
calculation gives an actual and prediction-
overview about the based risk values with
actual and potential risk reference values
development
Adaptability Comparison of actual User checks the calcu-
and potential risk lated values and reacts
development with when the system gives a
reference values and report/warning
automatic report in case
of significant deviation 373
Table 6: SSCRM-Framework – Treatment

Computerisation Isolated usage of IT sys- Treatment actions are
tems in SC processes al- developed in workshops
lows the collection of and improved cost-value
structured and unstruc- ratio calculation based
tured process data; Risk on process data.
relevant information can
be stored locally
Connectivity Connection of IT sys- Treatment actions are
tems allows the uni- or developed in workshops
bidirectional exchange and improved cost-value
of process data; Risk ratio calculation based
relevant information can on broader range of
be stored in a process available process data.
wide data base
Visibility Treatment activities are Based on the character-
characterized through ized risks the user has to
data points within the identify suitable mitiga-
Digital Shadow tion actions through the
data points as well as to
select and initiate them
Transparency The system recognizes Found actions and their
relations between risk effects have to be esti-
data points and treat- mated, selected and ini-
ment activity data points tiated by the user
and clusters them to po-
tential actions as well as
reports them to the user
Predictive capacity Potential actions will be User chooses the actions
simulated and the re- with the best possible
sults serve as decision outcome or with the
support. Additionally least side effects
potential negative im-
pacts on other risk data
points can be recognized
in advance
Adaptability After the evaluation User checks the chosen
the system decides au- actions and intervenes/-
tonomously about the corrects if necessary
374 initiation of mitigation
actions
Table 7: SSCRM-Framework – Monitoring

Computerisation Isolated usage of IT sys- At discrete points in time
tems in SC processes al- manager come together
lows the collection of and discuss about moni-
structured and unstruc- tored risks and initiated
tured process data; Risk treatment actions, sup-
relevant information can ported by available pro-
be stored locally cess data
Connectivity Connection of IT sys- At discrete points in
tems allows the uni- or time manager come
bidirectional exchange together and discuss
of process data; Risk about monitored risks
relevant information can and initiated treatment
be stored in a process actions, supported
wide data base by a broader range of
available process data
Visibility Identified risks and User has to recognize
where necessary initi- plan deviations by him-
ated actions appear as self and identify the rea-
individual entities in the sons
Digital Shadow
Transparency In case of a plan devia- Based on the reports the
tion the system tries to user has to adapt the ini-
identify the reasons and tiated actions or choose
reports them to the user other options reactively
Predictive capacity Due to a projection Based on the reports
of the digital shadow the user has to adapt
into the future poten- the initiated actions or
tial plan deviations choose other options
and reasons can be proactively
recognized in advance;
Corrective actions and
their effectiveness can
be simulated in advance
Adaptability Autonomous correction User supervises the sys-
of actions in case of a po- tem and corrects actions
tential plan deviation in case when plan de-
viations cannot be con-
tained through the sys-
tem 375
The paper presents a first approach of establishing SSCRM as sub-research field

of SCRM by proposing a specific research framework. The framework has been
created by reviewing literature from the field of SCRM and I4.0. Afterwards a
framework with SSCRM requirements has been developed as a guiding design
instrument.
5.1 Limitations and further research
The proposed framework is only one way to define the sub-research field of SS-
CRM and was created by combining an established SCRM research framework
with recent literature about I4.0. When more literature is available, further re-
search can be suggested for testing if these dimensions are sufficient or additional
dimensions have to be considered. Also the SSCRM phases and based on that
the roles of technology and people within the SSCRM maturity stages have been
postulated by the authors. Additional research will be undertaken to verify the
authors’ ideas, leading to generalized roles for individual requirement derivation.
It also has to be noted that some framework dimension combinations may not al-
low some of the artifacts as a research outcome. This issue can be solved through
screening future literature focusing on their position in the research field and
their outcome.
5.2 Managerial Implications
The smart SCRM developed here is a good basis for a proactive SCRM which in
the literature is discussed on a conceptual basis for many years (e.g. Henke, 2009)
but up to now it has rarely been realised in business practice. In the age of Big
Data, digitisation and autonomisation today we have sufficient data as well as
the technologies (such as blockchain), which can allow a proactive management
of such data along supply chains. The transparency in value-added networks
exists end-to-end so that in the future risks can be avoided or reduced at an earlier
stage than today. For a practical application of such a SSCRM it is also necessary
that there is a structured approach from application-oriented research to core
elements of a cycle of SSCRM.
376
References
References
Bauernhansl, T., M. ten Hompel, and B. Vogel-Heuser, eds. (2014). Industrie 4.0 in Produktion,
Automatisierung und Logistik: Anwendung - Technologien - Migration. Wiesbaden: Springer
Fachmedien.
Bischoff, J., C. Taphorn, D. Wolter, N. Braun, M. Fellbaum, A. Goloverov, S. Ludwig, T. Hegmanns,
C. Prasse, M. Henke, M. ten Hompel, F. Döbbeler, E. Fuss, C. Kirsch, B. Mättig, S. Braun, M. Guth,
M. Kaspers, and D. Scheffler (2015). Erschließen der Potenziale der Anwendung von ’Industrie
4.0’ im Mittelstand: Studie im Auftrag des Bundesministeriums für Wirtschaft und Energie (BMWi).
Ed. by J. Bischoff. Mülheim an der Ruhr.
Bostrom, R. P. and J. S. Heinen (1977). “MIS Problems and Failures: A Socio-Technical Perspective:
PART I: THE CAUSES”. In: MIS Quarterly 1.3, pp. 17–32.
Brindley, C., ed. (2004). Supply chain risk. Hanpshire, Burlington: Ashgate Publishing Ltd.
Bui, T. X. and R. H. Sprague, eds. (2016). Proceedings of the 49th Annual Hawaii International
Conference on System Sciences: 5-8 January 2016, Kauai, Hawaii. Piscataway, NJ: IEEE.
Butner, K. (2010). “The smarter supply chain of the future”. In: Strategy & Leadership 38.1, pp. 22–
31.
Chopra, S. and M. S. Sodhi (2004). “Managing Risk To Avoid Supply-Chain Breakdown”. In: MIT
Sloan Management Review 46.1, pp. 53–61.
Curkovic, S., T. Scannel, and B. Wagner (2013). “ISO 31000:2009 - Enterprise and Supply Chain Risk
Management: A Longitudinal Study”. In: American Journal of Industrial and Business Manage-
ment 3, pp. 614–630.
de Oliveira, U. R., F. A. S. Marins, H. M. Rocha, and V. A. P. Salomon (2017). “The ISO 31000 standard
in supply chain risk management”. In: Journal of Cleaner Production 151, pp. 616–633.
Fahimnia, B., C. S. Tang, H. Davarzani, and J. Sarkis (2015). “Quantiative Models for Managing
Supply Chain Risks: A Review”. In: European Journal of Operational Research 247.1, pp. 1–15.
Faisal, M. N., D. K. Banwet, and R. Shankar (2006). “Supply chain risk mitigation: modeling the
enablers”. In: Business Process Management 12.4, pp. 535–552.
Fan, H., T. C. E. Cheng, G. Li, and P. K. C. Lee (2016). “The Effectiveness of Supply Chain Risk Infor-
mation Processing Capability: An Information Processing Perspective”. In: IEEE Transactions on
Engineering Management 63.4, pp. 414–425.
Fan, H., G. Li, H. Sun, and T. Cheng (2017). “An information processing perspective on supply chain
risk management: Antecedents, mechanism, and consequences”. In: International Journal of
Production Economics 185, pp. 63–75.
Goh, R. S. M., Z. Wang, X. Yin, X. Fu, L. Ponnambalam, S. Lu, and X. Li (2013). “RiskVis: Supply chain
visualization with risk management and real-time monitoring”. In: IEEE International Conference
on Automation Science and Engineering (CASE), 2013. Piscataway, NJ: IEEE, pp. 207–212.
Güller, M., E. Koc, T. Hegmanns, M. Henke, and B. Noche (2015). “A Simulation-based Decision
Support Framework for the Real-time Supply Chain Risk Management”. In: International Journal
of Advanced Logistics 4.1, pp. 17–26.
Heckmann, I., T. Comes, and S. Nickel (2015). “A critical review on supply chain risk: Definition,
measure and modeling”. In: Omega 52, pp. 119–132.
Henke, M. (2009). Supply Risk Management: Planung, Steuerung und Überwachung von Supply
Chains. Berlin: Erich Schmidt.
377
Hermann, M., T. Pentek, and B. Otto (2016). “Design Principles for Industrie 4.0 Scenarios”. In:
Proceedings of the 49th Annual Hawaii International Conference on System Sciences. Ed. by T. X.
Bui and R. H. Sprague. Piscataway, NJ: IEEE, pp. 3928–3937.
Hevner, A. R., S. T. March, J. Park, and S. Ram (2004). “Design Science in Information Systems
Research”. In: MIS Quarterly 28.1, pp. 75–105.
Hillman, M. and H. Keltz (2007). Managing Risk in the Supply Chain - A Quantitative Study. Ed. by
AMR Research.
Hirsch-Kreinsen, H. (2014). “Wandel von Produktionsarbeit - ”Industrie 4.0””. In: WSI Mitteilungen
6, pp. 421–429.
Ho, W., T. Zheng, H. Yildiz, and S. Talluri (2015). “Supply chain risk management: a literature
Jüttner, U., H. Peck, and M. Christopher (2003). “Supply Chain Risk Management: Outlining an
Agenda for Future Research”. In: International Journal of Logistics: Research and Applications 6,
pp. 197–210.
Kabashkin, I. V. and I. V. Yatskiv, eds. (2014). Proceedings of the 14th International Conference
"Reliability and Statistics in Transportation and Communication" (RelStat’14).
Kersten, W., T. Blecker, and C. M. Ringle, eds. (2015). Sustainability in logistics and supply chain man-
agement: New designs and strategies. 1st ed. Vol. 21. Proceedings of the Hamburg International
Conference of Logistics (HICL). Berlin: epubli GmbH.
Kersten, W., M. Seiter, B. von See, N. Hackius, C. Rosentritt, C. Böhle, G. Reich, T. Maurer, and R.
Sauter (2016). “Trends und Strategien in Supply Chain Management und Logistik - Chancen
der digitalen Transformation”. In: Den Wandel gestalten. Ed. by T. Wimmer and C. Grotemeier.
Hamburg: Deutscher Verkehrs-Verlag, pp. 346–361.
Kilubi, I. and H. D. Haasis (2016). “Supply chain risk management research: Avenues for further
studies”. In: International Journal of Supply Chain and Operations Resilience 2.1, p. 51.
Lessard, L. and E. Yu (2012). “Using Design Science Research to Develop a Modeling Technique for
Service Design”. In: DESRIST 2012, LNCS 7286. Ed. by K. Peffers, M. Rothenberger, and B. Kuechler.
Berlin, Heidelberg: Springer-Verlag, pp. 66–77.
Leveling, J., A. Schier, F. Luciano, and B. Otto (2014). “Konzeption eines proaktiven Risikomanage-
ments in Logistiknetzwerken”. In: Logistics Journal : Proceedings 2014.01.
Lindroth, R. and A. Norrman (2001). “Supply Chain Risks and Risk Sharing Instruments - An Illustra-
tion fromo the Telecommunication Industry”. In: Proceedings of the Logistics Research Network
6th Annual Conference. Ed. by McKinnon, pp. 297–307.
Lu, Y. (2017). “Industry 4.0: A Survey on Technologies, Applications and Open Research Issues”. In:
Journal of Industrial Information Integration.
McKinnon, ed. (2001). Proceedings of the Logistics Research Network 6th Annual Conference.
Norrman, A. and U. Jansson (2004). “Ericsson’s proactive supply chain risk management approach
after a serious sub-supplier accident”. In: International Journal of Physical Distribution & Logistics
Norrman, A. and R. Lindroth (2004). “Categorization of Supply Chain Risk and Risk Management”.
In: Supply chain risk. Ed. by C. Brindley. Hanpshire, Burlington: Ashgate Publishing Ltd., pp. 14–
27.
Peffers, K., M. Rothenberger, and B. Kuechler, eds. (2012). DESRIST 2012, LNCS 7286. Berlin, Heidel-
berg: Springer-Verlag.
Pettit, T. J., K. L. Crocton, and J. Fiksel (2013). “Ensuring Supply Chain Resilience: Development
and Implementation of an Assessment Tool”. In: Journal of Business Logistics 34.1, pp. 46–76.
378
References
Pfohl, H.-C., B. Yahsi, and T. Kurnaz (2015). “The Impact of Industry 4.0 on the Supply Chain”. In:
Sustainability in logistics and supply chain management. Ed. by W. Kersten, T. Blecker, and
C. M. Ringle. Vol. 21. Proceedings of the Hamburg International Conference of Logistics (HICL).
Berlin: epubli GmbH, pp. 31–58.
Ponis, S. T. and A. C. Ntalla (2016). “Supply Chain Risk Management Frameworks and Models: A
Review”. In: International Journal of Supply Chain Management 5.4, pp. 1–11.
Proceedings of 2015 International Conference on Intelligent Computing and Internet of Things: ICIT
2015 : January 17-18, 2015, Harbin, China (2015). Piscataway, NJ: IEEE.
Qin, J., Y. Liu, and R. Grosvenor (2016). “A Categorical Framework of Manufacturing for Industry
4.0 and Beyond”. In: Procedia CIRP 52, pp. 173–178.
Ritchie, B. (2007). “Supply chain risk management and performance: A guiding framework for
future development”. In: International Journal of Operations & Production Management 27.3,
pp. 303–322.
Roth, A., ed. (2016a). Einführung und Umsetzung von Industrie 4.0: Grundlagen, Vorgehensmodell
und Use Cases aus der Praxis. Berlin, Heidelberg: Springer-Verlag.
Roth, A., ed. (2016b). Einführung und Umsetzung von Industrie 4.0: Grundlagen, Vorgehensmodell
und Use Cases aus der Praxis. Berlin, Heidelberg: Springer-Verlag.
Roth, A. (2016c). “Industrie 4.0 - Hype oder Revolution?” In: Einführung und Umsetzung von Industrie
4.0. Ed. by A. Roth. Berlin, Heidelberg: Springer-Verlag, pp. 1–15.
Schlüter, F., K. Diedrich, and M. Güller (2017). “Analyzing the Impact of Digitalization on Supply
Chain Risk Management”. In: 2017 IPSERA Conference.
Schröder, M., M. Indorf, and W. Kersten (2014). “Industry 4.0 and its Impact on Supply Chain Risk
Management”. In: Proceedings of the 14th International Conference "Reliability and Statistics in
Transportation and Communication" (RelStat’14). Ed. by I. V. Kabashkin and I. V. Yatskiv, pp. 114–
125.
Schuh, G., R. Anderl, J. Gausemeier, M. ten Hompel, and W. Wahlster (2017). Industrie 4.0 Maturity
Index: Managing the Digital Transformation of Companies. acatech STUDY. München: Herbert
Utz Verlag.
Siepmann, D. (2016a). “Industrie 4.0 - Fünf zentrale Paradigmen”. In: Einführung und Umsetzung
von Industrie 4.0. Ed. by A. Roth. Berlin, Heidelberg: Springer-Verlag, pp. 35–46.
Siepmann, D. (2016b). “Industrie 4.0 - Technologische Komponenten”. In: Einführung und Umset-
zung von Industrie 4.0. Ed. by A. Roth. Berlin, Heidelberg: Springer-Verlag, pp. 47–72.
Singhal, P., G. Agarwal, and M. L. Mittal (2011). “Supply chain risk management: review, classifica-
tion and future research directions”. In: International Journal of Business Science and Applied
Tang, C. and B. Tomlin (2008). “The power of flexibility for mitigating supply chain risks”. In:
Tang, O. and S. N. Musa (2011). “Identifying risk issues and research advancements in supply chain
risk management”. In: International Journal of Production Economics 133.1, pp. 25–34.
ten Hompel, M. and M. Henke (2017). “Logistik 4.0 – Ein Ausblick auf die Planung und das Man-
agement der zukünftigen Logistik vor dem Hintergrund der vierten industriellen Revolution”.
In: Handbuch Industrie 4.0 Bd.4. Ed. by B. Vogel-Heuser, T. Bauernhansl, and M. ten Hompel.
Springer Reference Technik. Berlin, Heidelberg and s.l.: Springer Berlin Heidelberg, pp. 249–
259.
379
Allgemeine Grundlagen. 2nd ed. Springer Reference Technik. Berlin, Heidelberg and s.l.: Springer
Berlin Heidelberg.
Wan, J., H. Cai, and K. Zhou (2015). “Industrie 4.0: Enabling technologies”. In: Proceedings of 2015
International Conference on Intelligent Computing and Internet of Things. Piscataway, NJ: IEEE,
pp. 135–140.
Wang, L. and R. Ranjan (2015). “Processing Distributed Internet of Things Data in Clouds”. In: IEEE
Cloud Computing 2.1, pp. 76–80.
Wimmer, T. and C. Grotemeier, eds. (2016). Den Wandel gestalten: Driving Change. Hamburg:
Deutscher Verkehrs-Verlag.
Zsidisin, G. A. and B. Ritchie (2009a). “Chapter 1: Supply Chain Risk Management - Developments,
Issues and Challenges”. In: Supply Chain Risk. Ed. by G. A. Zsidisin and B. Ritchie. Vol. 124.
International Series in Operations Research & Management Science. Boston MA: Springer-
Verlag US, pp. 1–12.
Zsidisin, G. A. and B. Ritchie, eds. (2009b). Supply Chain Risk: A Handbook of Assessment, Manage-
ment, and Performance. Vol. 124. International Series in Operations Research & Management
Science. Boston MA: Springer-Verlag US.
Zweig, M., A. Agawal, B. Stall, C. Bremer, P. Mangers, A. Beifus, and M. Chauhan (2015). Globalize or
customise: finding the right balance: Global steel 2015-2016. Ed. by EY.
380
Framework for Digitalized Proactive
Supply Chain Risk Management
Katharina Diedrich1
1 – Technical University of Dortmund & Volkswagen Commercial Vehicles
In order to initiate proactive measures and to prevent the actual occurrence of risk
events a digitalized and proactive supply chain risk management is needed.Exist-
ing approaches for digitalized and proactive SCRM are investigated by a system-
atic literature review and completed by a requirements engineering. Furthermore
frameworks for SCRM are reviewed and the fulfillment of the requirements is
checked. These results are applied to develop a digitalized and proactive SCRM
framework.The findings in this paper are a profile of requirements and a frame-
work, which enables a proactive management of forthcoming risks. Therefore an
application of risk prediction and quantitative risk assessment in context of digi-
talized SCRM is used. The developed framework includes an adapted proactive
SCRM process containing digitalized SCRM phases and digitalized SCRM methods,
inter alia early knowledge and early assessment.
Keywords: Digitalized Proactive SCRM; Risk Information Integration; Risk Pre-

diction; Risk Assessment
381
Framework for Digitalized Proactive Supply Chain Risk Management
1 Introduction
Due to high outsourcing rates manufacturing companies are highly dependent on

the performance of suppliers and supply chains (SC). Supply chain risks (SCR) can
cause disruptions which lead to glitches. Glitches result in demand and supply
mismatches and can have propagating effects on the entire SC. SCR is defined
as the likelihood and impact of unexpected risk events which influence any part
of the SC (Ho et al. 2015 p. 5035). A company’s SC performance is depending on
the degree of implementation of a supply chain risk management (SCRM) system
(Thun & Hoenig 2011 pp. 246) and of assignment of responsibilities. Due to the
crucial impacts and thereby high importance the management board should
bear the responsibility for SCRM system planning, implementing and operating.
The impact and probability of SCR occurrences for companies can be minimized
with proactive SCRM systems and a high efficiency of company’s SC processes
can be enabled. The need for proactive SCRM has been identified by scientific
researchers and managers in companies practice. Several recent published find-
ings which cover the topic are available, while in the previous years before 2010
just a few papers were published. Due to the digitalization of SCs as well as
the development of new technologies and SC partner collaborative approaches
new potentials for proactive SCRM arise. In order to remain competitive and for
economical production, companies have to deliberately take risks and accept
a certain insecure risk situation. This publication investigates proactive SCRM
approaches which accept the risk situation and focus on the management of
individual forthcoming risk events.
1.1 SCRM Research Needs in the Field of Digital Transformation
Although the SCRM research field has gained a lot of attention during the last
decade, the field of digitalized SCRM and proactive SCRM development is not
considered adequately (Kırılmaz & Erol 2017, Kache & Seuring 2017 p. 20). Besides
the scientific research gap, the need for proactive SCRM approaches in companies
practice is urgent due to high number of risk events and the limitation to reactive
firefighting systems (He et al. 2015 p. 1005, Kilubi & Haasis 2016 p. 66, Tang &
Musa 2011 pp. 31). There have been some analyses conducted reviewing the
SCRM literature in the recent past, which have a specific focus and thereby don’t
serve the objective of this publication. The past reviews have been focused on
the purpose of SCRM studies (Vanany, Zailani & Pujawan 2009, Sodhi, Son & Tang
382
1 Introduction
2012), definition of specific SCRM terms (Ho et al. 2015, Singh & Wahid 2014), risk
classification (Qazi, Quigley & Dickson 2015, Tang & Musa 2011, Singhal, Agarwal &
Mittal 2011), the SCRM process (Ho et al. 2015, Qazi, Quigley & Dickson 2015), risk
sources (Kilubi & Haasis 2016), strategic changes in the field of SCRM (Ghadge, Dani
& Kalawsky 2012), quantitative models for SCRM (Heckmann, Comes & Nickel 2015,
Fahimnia et al. 2015) and relationship between SCRM and performance (Kilubi
& Haasis 2016). The SCRM procedure has become a key issue of this research
field. Most literature review publications show an analysis of SCRM publications
in general, whereas the focus on proactive SCRM and the Digital Transformation
of SCRM is missing. There is a research need for quantitative and model based
proactive SCRM approaches (Kırılmaz & Erol 2017 pp. 62, Tang & Musa 2011 p.
32). These models should take interdependencies between dynamic risks and
the risk propagation along the SC into account (Qazi, Quigley & Dickson 2015). In
addition to material flow, for SCRM information flow and financial flow ought to
be considered in depth (Tang & Musa 2011 p. 32). In addition to the analysis of
Schlüter, Diedrich & Güller 2017 for digitalized approaches for the SCRM phases a
comprehensive framework is needed which includes digitalized approaches and
applies them for a proactive SCRM procedure.
1.2 Objective and Research Methodology
The objective of this paper is the development of a proactive SCRM framework

which takes the Digital Transformation of the environment and of SCs that is
currently taking place as well as the arising potentials for SCRM into account. In
order to meet this objective, a Systematic Literature Review (SLR) is conducted.
Due to the small number of publications regarding proactive SCRM, the keyword
search of the SLR is extended to proactive SCRM and SCRM Frameworks. This
paper explores the above described research gap regarding scientific approaches
for proactive SCRM.
Hence, a SLR was conducted, targeting to collect and analyze relevant publications
regarding proactive SCRM approaches. This SLR is used for the analyzation of
proactive SCRM approaches and for the description of the research gap concerning
proactive SCRM. For the following development of a proactive SCRM framework a
requirements gathering has to be conducted. Requirements were derived from
SLR findings and the results are applied to develop a digitalized and proactive
SCRM framework. In addition specific features of risk assessment will be outlined
and suggestions for further research conclude this paper.
383
2 Perspectives in Proactive SCRM
2.1 Digitalized SC Function as an Enabler for Advanced SCRM
New technologies enable the digitalization of SCs. These technologies are for
example near field communication (NFC), radio frequency identification (RFID),
barcode and machine sensors. New technologies are applied in research fields
which are part of Digital Transformation and Industry 4.0, such as cyber-physical
systems, cloud computing, smart factory and internet of things (Kersten, Schröder
& Indorf 2015). From these applied technologies in supply chain management
and logistics new data sources arise which enable a data driven management
of SCRs by application in appropriate methods. For efficient SCRM an advanced
SC information management is required, which contains a supplier data base
including master data of n-tier supplier, production sites, and further specific
knowledge. The knowledge management and visualization of SC and the SCR
information processing in the context of Digital Transformation and the associated
data generation is still an unexplored research field (Fan et al. 2017, Biswas & Sen
2016). Current key challenges of the Digital Transformation of SC and SCRM are
data islands, silo mentality and insufficient cooperation let alone collaboration
of data availability. The Supply Chain Shadow as a virtual representation of a
SC is enabled by the digitalization of SCs and the accompanying availability of
real-time data (Henke 2017). The concept of the Supply Chain Shadow uses the
available data in a virtual information and knowledge transfer.
2.2 Digitalized Proactive SCRM Approaches
The digitalization of SCRM is based on digitized SCs and enhanced technologies

and methods. Available real-time data regarding SCRs result from applied tech-
nologies. New potentials for digitalized SCRM have been analyzed by Schlüter,
Diedrich & Güller 2017, their findings are summarized in an overview of existing
approaches which take digitalized methods for the phases identification, assess-
ment, treatment and monitoring & control and thereby a digitalization of the
SCRM procedure into account. A resulting change in the SCRM procedure is the
parallelized pass through of risks in the digitalized procedure (Schlüter, Diedrich
& Güller 2017 p. 6). Thereby the phases don’t have to be processed sequentially
by long-term committees anymore and risks can be handled simultaneous in
different phases.
384
3 Systematic Literature Review for Proactive SCRM
Figure 1: SLR method
The SLR conducted in this paper is based on the methodical approach described by
Seuring & Gold 2012, as can be seen in figure 1. A SLR is a transparent, reproducible
scientific method which targets to cover the state of research with a specific
research focus defined by the following research questions (RQ):
RQ1: Do the proactive SCRM frameworks, methods or approaches correspond
with the introduced definition for proactive SCRM?
RQ2: Do the approaches take new potentials for proactive SCRM arising from
the Digital Transformation into account?
RQ3: Which requirements for a proactive SCRM Framework can be raised?
Found publications are selected as relevant by use of assessment criteria (AC):
AC1: Future-oriented methods for the phases of the SCRM process
AC2: Management of forthcoming risks while accepting the risk situation
AC3: Potentials for SCRM which result from the Digital Transformation of SCs
& SCRM itself
AC4: Further aspects of proactive SCRM
The first AC shall bring attention to forthcoming risks. The AC2 has to be described
further for a deeper insight: The bearable risk situation of a company depends on
the SC dependency, buffers, flexibility and further aspects, whereat no guidelines
exist that describe how much risk a company should accept. The higher the ac-
cepted risks are, the more chances can be taken and the higher is the impact of
potential risks. The company’s risk situation is accepted to gain chances by ac-
cepting risks and this paper targets at not changing the risk situation and existing
385
Figure 2: Number of references
structures. Therefore the optimization of the risk situation by preventive man-

agement and not taking the risks is explicitly not looked for. The third AC refers
to the application of new technologies and methods for deeper risk knowledge
and management improvement. AC4 targets the open search for further aspects
or approaches which serve the proactive SCRM. SCRM framework publications
are assessed based on in-depth description of reactive SCRM frameworks. After
selection of papers the full paper review process and the content analysis have
been carried out. These findings are introduced in the following section.
3.1 Analytical Description of Publications
For answering these research questions appropriate publications are reviewed

with the year of publication between 2000 and June 2017. The applied keywords
for database search are proactive supply chain risk management and framework
supply chain risk management. The database SCOPUS.COM was most suitable for
this search, as Fahimnia et al. 2015 has described before. SCOPUS includes e.g.
Wiley, Taylor & Francis, Emerald, IEEE Xplore and Springer. Thereby the search is
open to several journals and conferences and limited by the application of the
defined keywords.
The resulting numbers of publications are shown in figure 2, whereby the first
number refers to proactive SCRM and the second number to SCRM framework
publications. The analysis of the distribution of suitable publications over the year
386
Figure 3: Years of publication of suitable publications
Figure 4: Authors countries of suitable publications
387
of publication is shown in figure 3. This analysis shows a significant research in-

terest in proactive SCRM since 2008. There are just a small number of approaches
for proactive SCRM in scientific literature available, which confirms the research
gap for proactive management.
The suitable publications are published in Journals (15/18) and conferences (7/9).
This shows that conferences are an important research platform for SCRM and
shouldn’t be neglected in SLR. The countries of publishing authors are visualized
in figure 4.
3.2 Content Analysis of Proactive SCRM Publications
The suitable SCRM framework publications show SCRM procedure approaches,

but none of these are proactive approaches. In literature the terms reactive, pre-
ventive and proactive SCRM are often used, but the definition and understanding
of the term proactive SCRM differs in the literature even though many publica-
tions refer to proactive SCRM (Qazi, Quigley & Dickson 2015). Many publications
use the term ”proactive” but actually don’t meet the ACs for proactive SCRM
and are rather preventive approaches. The term proactive SCRM is often applied
to approaches which should be assigned to preventive SCRM, because these
approaches change SC structures, apply alternative strategical approaches, for
example dual sourcing, and by not taking on risks any longer improve the risk
situation. None of the reviewed publications give a definition for proactive SCRM.
The author’s understanding and definition of proactive SCRM is as follows: Proac-
tive SCRM approaches target to gain action time for management of individual
forthcoming SCR and thereby can be described as ex ante management while
accepting the risk situation of a company in a SC network. A common, but not
defined, understanding can be found in applicable publications, such as Rotaru,
Wilkin & Churilov 2014 and Leveling et al. 2014. The remaining suitable publica-
tions meet at least one AC (8), whereas no publication meets all ACs, as can be
seen in figure 5.
To gain the ability of proactive SCRM a high level of technical abilities has to be
achieved. Furthermore the research field of digitalized and proactive SCRM is
fast growing, so that the existing reviews are no longer up to date, which was
confirmed by the conducted review.
388
Figure 5: Analysis of proactive SCRM approaches
389
3.2.1 Proactive SCRM Approaches (RQ1)
Real-time risk monitoring and SC visualization is proposed by Goh et al. 2013

for early alert to achieve proactive SCRM. This might enable in-time knowledge
regarding SC structures and dependencies but no future oriented knowledge
which is required for proactive management. Leveling et al. 2014 describe data
integration and Schlegel 2015 underlines the importance of immediate risk infor-
mation and knowledge which can be gained by big data and predictive analytics,
nonetheless a concrete description of application of analytics for proactive SCRM
and inclusion in SCRM phases is not considered by found approaches.
Most publications which refer to a SCRM process describe the phases identifi-
cation, assessment, mitigation and control (Ho et al. 2015). Other publications
variate these phases to special requirements for proactive SCRM which contain
for example a prioritization based on risk evaluation, risk criteria and the differen-
tiation between tolerable and not acceptable risk events (Kırılmaz & Erol 2017 p.
57). Kırılmaz & Erol 2017 describes the phases of identification, measurement,
evaluation, mitigation and control.
The proactive approach of Kırılmaz & Erol 2017 is based on shifting orders among
suppliers and contains a quantitative model for proactive SCRM, which takes
costs and risk aspects for procurement into account. Approaches for selection
and combination of action measures can be found in for example Ghadge, Dani &
Kalawsky 2012 p. 324.
3.2.2 Proactive SCRM Potentials Arise from Digital Transformation (RQ2)
For answering this research question lots of cross referencing and further infor-
mation gathering had to be carried out. Only seven of the located publications
found by the SLR referred to new technologies, Digital Transformation of SCs and
risk management, data mining or predictive analytics, which shows the research
need for Digital Transformation of scientific approaches for proactive SCRM. A re-
search gap aspect is the consideration of available information, the uncertainties
involved and the information integration into the SCRM procedure (Rotaru, Wilkin
& Churilov 2014 p. 12, Fan et al. 2017). This future research need includes readily
available, timely and accurate risk information and can be gained by application
of information and communication technologies such as RFID, GPRS and social
networks (Ghadge, Dani & Kalawsky 2012 p. 328, Chae 2015). The application of
390
Figure 6: SLR findings regarding proactive SCRM
391
big data analytics can provide SC visibility and thereby support decision making
for SCRM (Seele 2017, Wang et al. 2016, Ghosh 2015, Schoenherr & Speier-Pero
2015, Leveling, Edelbrock & Otto 2014). This small number of publications refers
to the application and advantages of data-driven analytical approaches in SCRM,
but they rarely refer to risk management processes. The application of data min-
ing and predictive analytics for SCRM is still a research gap. Future implications
for SCRM based on digitalized approaches and digitalized approaches for conven-
tional SCRM phases have been analyzed and summarized by Schlüter, Diedrich &
Güller 2017.
3.2.3 Requirements for a Proactive SCRM Framework (RQ3)
SCRM phases are rather complex process phases which interact with each other
and require a high level of technical abilities. A high number of publications
describing SCRM process phases exist in scientific literature. Within this research
question SCRM frameworks with characteristics for proactive management are
analyzed.
Ponis & Ntalla considered 16 SCRM framework models as adequate, of which
none is published after 2013 (Ponis & Ntalla 2016 p5). During the SLR of this publi-
cation there have been found 43 SCRM framework publications with a publication
date of 2013 or newer, which have to be examined for requirement fulfillment to
complement the SLR conducted by Ponis & Ntalla.
The framework designed by Aqlan & Lam 2015 includes risk prediction, but further
description of risk prediction methods and time dependency of risk analysis are
missing (Aqlan & Lam 2015).
He et al. 2015 show the potential of risk prediction, risk simulation & evaluation
and risk mitigation as key process components to support the SCRM process, but
a transfer to a comprehensive proactive SCRM framework and proactive SCRM
procedure is still missing. The key process components are executed sequentially
(He et al. 2015 p. 1008).
SCRM as part of SCM calls for technical competence and for new technological
solutions to enable new SCRM approaches based on big data (Leveling et al. 2014
p. 6).
Mangla, Kumar & Kumar Barua 2014 have identified important factors for SCRM
and SC sustainability. Some of these factors can be understood as requirements
392
Figure 7: Requirements for proactive SCRM
for proactive SCRM, inter alia knowledge and understanding about SCRs, strategic
risk planning, SC integration, security issues knowledge, network and global com-
plexion understanding, SC visibility and mutual transparency as well as decision
and information sharing (Mangla, Kumar & Kumar Barua 2014 pp. 127). A research
project supported by the BMWi, Germany showed the timely identification of SCR
as a key issue for proactive SCRM (Cirullies & Kamphues 2014).
The profile of requirements for a proactive SCMR framework is summarized in

figure 7. The requirements for proactive SCRM have been derived from existing
proactive approaches and supplemented by requirements which have beforehand
been raised in a workshop with SCRM research experts.
393
Figure 8: Digitalized SCRM components
4 Conceptual Framework
The terms approach, model, method, concept, procedure and framework are often
synonymously applied for the description of a complex management approach
(Vogel & Lasch 2015 pp. 102). The objective of this section is the development of
a proactive management approach. Due to digitalization of SC and SCRM, the use
of data, tools and analytical approaches is steadily increasing. Components for a
digitalized and proactive SCRM framework are shown in figure 8.
The requirements for proactive SCRM can be fulfilled by the development of
new methods for an advanced early-warning-system, which contains innovative
aspects of the proactive SCRM framework, including smart risk analytics and risk
prediction methods. As figure 9 shows, in addition to the early warning system an
adaption of SCRM phases for management of future risks in a proactive procedure
is developed.
The risk identification phase and risk assessment phase should be connected
in a single comprehensive phase, because the identification of a risk requires
knowledge regarding the risk assessment and especially risk impact for managers
394
Figure 9: Digitalized proactive SCRM framework
to be able to recognize an event as a risk with a negative impact to a company or

a SC network. Therefore these phases are connected in the early warning system,
as can be seen in figure 9.
Risk prediction and quantitative risk assessment for future SCRs are core compe-
tences of the framework. In the following the use of data mining, big data and
predictive analytics for proactive SCRM is referred to as smart risk analytics.
The collection of company and SC data from a wide range of internal and external
data sources, inter alia social networks, natural disaster prediction, transportation
data, stock data, material flow data and internal process and planning data, is
a prerequisite for smart risk analytics. Natural disaster risks information, for
example earthquakes, can be combined with industrial data and geographical
data for risk sensing, assessment and predicting future events (Seele 2017 p.
684, Rajesh 2016, He et al. 2015 p. 1009, Schlegel & Trent 2015, Schlegel 2015).
Smart risk analytics enables an intelligent and proactive SCRM based on the
digitalization of SCs and risk management processes targets at concrete risk
information for forthcoming SCR. Smart risk analytics can help assessing and
managing SCR and distinguishing between risks that should be avoided and risks
that should be taken (Wang et al. 2016 p. 101, p. 103). Information processing for
395
Figure 10: Risk information layer and data sources
SCRM, risk analysis and assessment as well as sharing of risk information should
be processed proactively (Fan et al. 2017 p. 67).
For achieving smart risk analytics, the key function of the framework risk data
management is needed, which leads to a high level of risk information and knowl-
edge. Collected data has to be stored, made available and maintained in real-time
to achieve potentials of SC digitalization. This enables a rapid analysis of in-
formation streams to generate real-time risk information and by application of
predictive analytics risk forecasting abilities.
According to the information layer of the described Supply Chain Shadow, devel-
oped by Henke 2017, a risk information layer is described in figure 10. A variety
of internal and external data sources have to be integrated in risk data manage-
ment.
There is also need for research regarding the evaluation and further development
of risk assessment, inter alia based on multi-criteria evaluation methods (Rotaru,
Wilkin & Churilov 2014 p. 12). The extended functions of an adapted procedure
396
Figure 11: Adapted SCRM Phases
are core competencies of the framework, as can be seen in figure 9, and are further
described in figure 11. The functions are based on SLR findings and completed
with companies requirements in practice and own developed approaches. Espe-
cially the risk assessment phase has to be processed beforehand and thereby ex
ante to risk events. This phase enables the triggering of the following proactive
procedure.
Rapid risk assessment can be achieved by development of appropriate tool solu-

tions (Aqlan 2016 pp. 110). Approaches of quantitative risk assessment can be
found in Fahimnia 2015 and Dong & Cooper 2016. The forward-looking method is
a key function for proactive SCRM procedure, hereafter referred to as dynamic risk
397
assessment. The method of an ex ante risk assessment is based on knowledge

and information of forthcoming SCR.
Risk events are linked to SC processes and performance, which themselves can
trigger further risk events and decrease the SC performance. Besides this, simul-
taneous occurring events can increase or diminish the probability and impact.
The dynamic time-dependent risk behavior, risk interdependencies and the prop-
agation of SCR along the entire SC has to be considered for risk assessment and
proactive approaches. This research field is relatively unexplored, but first ap-
proaches can be found in for example Ghadge, Dani & Kalawsky 2012 p. 329 and
Qazi, Quigley & Dickson 2015 p. 7.
Further research is required to include risks which do not occur frequently and for
which no historical data is available. The methodical approach of risk aggregation
should also take risk propagation along the SC into account. Depending on the
risk propagation along the SC an individual risk measure for the SC partner as a
viewing point has to be aggregated.
The combination of smart risk analytics and dynamic risk assessment enables
a smart forward looking risk action and is the functionality of the described
early warning system component of the proactive SCRM framework. It can be
inferred from this, that the potential of proactive management and the ability to
generate action time before a risk event occurs is enabled by the combination of
existing SCRM methods, the adaption of SCRM phases and the development of
time-depending and forward-looking innovative methods.
The prioritization of risks is needed by risk managers to detect the most urgent
need for action and to focus on these risks (Aqlan & Lam 2015 p. 55), and thereby
should be part of a proactive SCRM system. Those risks should be high prioritized,
which are most likely to stop or disturb the production, and those risks which
might lead to high financial losses for the manufacturing company. However, in
order to evaluate the whole dimension of forthcoming risks, the risk prioritization
requires the consideration of risk interdependencies and risk aggregation as
described above.
For proactive SCRM approaches, the prioritization has to be linked to information

which gives deeper insights into the risk sources, remaining action time and po-
tential measures to handle the risk events. Companies need reliable information
to choose which risk should be handled with cost-intensive proactive measures.
A visualization approach shows identified risks with need for proactive action in a
398
5 Conclusion and Discussion
risk heatmap (Singh & Wahid 2014 p. 64), which needs to be further developed to
include dynamic time dependent information.
5 Conclusion and Discussion
The methodical approach of this paper ensures a reproducible compilation of

scientific research and publications of the research fields of proactive SCRM
and SCRM frameworks. For further improvement empirical validation in compa-
nies practice is still pending. The main limitations of the SLR result from used
databases and keywords. This SLR includes only publications with respect to
proactive SCRM approaches. Access to scientific publications is often limited. For
this paper the access to SCOPUS.COM via university license was used and enabled
access to many publications. Nonetheless a few publications were not available
in digital form or not accessible. Open Access is a rather new and slowly spreading
way to ensure access to scientific publications.
The major findings of this paper are an overview of existing approaches in the field
of proactive SCRM and digitalization, comprehensive requirements for proactive
SCRM as well as the development of a digitalized proactive SCRM framework.
Individual risk prediction, defined as smart risk analytics, and time depending
risk assessment, defined as dynamic ex ante risk assessment, are key components
of this proactive SCRM framework. These components are enabled by the Digital
Transformation of SCs.
Risk mitigation through collaboration has a high need for implementation (Ghadge,
Dani & Kalawsky 2012 p. 328). The need for collaborative approaches of SC part-
ners regarding collaborative decision making and strategies is high due to multi-
ple dependencies and the risk propagation along the entire SC, which has to be
initiated by management. Besides this, companies are reluctant to implement
proactive SCRM processes because as Repenning & Sterman 2001 stated ”nobody
ever gets credit for fixing problems that never happened” (Tang 2006 p. 480).
Moreover the key function of SCR prioritization has to be implemented - other-
wise critical risks would simply be drowned by the flood of predicted risk events
and made targeted action impossible. Figure 12 shows managerial implications
of this paper.
From the findings of this paper a future research need for risk prediction and
forward looking risk assessment with concrete time reference can be derived.
399
Figure 12: Significance of Digital Transformation and future research directions

of SCRM
Both research gaps are correlated to higher information integration into the SCRM
process. The high level information integration enables an objective and specific
assessment of potential forthcoming risks. Furthermore research needs result
from the limitations of the conducted SLR, so that smart risk analytics and early
warning are further fields of research which have not been exploited with this
analysis. Future research should be focused on risk assessment of forthcoming
risks, smart risk analytics and early warning.
References
Aqlan, F. (2016). “A software application for rapid risk assessment in integrated supply chains”. In:
Expert Systems with Applications 43, pp. 109–116.
Aqlan, F. and S. S. Lam (2015). “A fuzzy-based integrated framework for supply chain risk assess-
ment”. In: International Journal of Production Economics 2015.161, pp. 54–63.
Biswas, S. and J. Sen (2016). “A Proposed Architecture for Big Data Driven Supply Chain Analytics”.
In: International Journal of Supply Chain Management, pp. 1–24.
Chae, B. (2015). “Insights from hashtag #supplychain and Twitter Analytics”. In: International
Journal of Production Economics 165, pp. 247–259.
400
References
Cirullies, J. and J. Kamphues (2014). Integrierte Terminierung und Transportplanung unterstützt

durch kollaboratives Risikomanagement in der Automobilindustrie: Schlussbericht. Fraunhofer
IML, Bundesministerium für Wirtschaft und Energie.
Dong, O. and O. Cooper (2016). “An orders-of-magnitude AHP supply chain risk assessment
framework”. In: International Journal of Production Economics 2016.182, pp. 144–156.
Fahimnia, B., C. S. Tang, H. Davarzani, and J. Sarkis (2015). “Quantitative models for managing
supply chain risks”. In: European Journal of Operational Research 2015.247, pp. 1–15.
Fan, H., G. Li, H. Sun, and T. Cheng (2017). “An information processing perspective on supply chain
risk management: Antecedents, mechanism, and consequences”. In: International Journal of
Production Economics 185, pp. 63–75.
Ghadge, A., S. Dani, and R. Kalawsky (2012). “Supply chain risk management: present and future
scope”. In: The international journal of logistics management 23.3, pp. 313–339.
Ghosh, D. (2015). “Big Data in Logistics and Supply Chain Management”. In: 2015 International
Symposium on Advanced Computing and Communication (lSACC).
Goh, R. S. M., Z. Wang, X. Yin, X. Fu, and L. Ponnambalam (2013). “RiskVis: Supply Chain Visual-
ization with Risk Management and Real-time Monitoring”. In: IEEE International Conference on
Automation Science and Engineering (CASE), pp. 207–212.
Grötsch, V., C. Blome, and M. Schleper (2013). “Antecedents of proactive supply chain risk man-
agement”. In: International Journal of Production Research 51.10, pp. 2842–2867.
He, M., H. Ji, Q. Wang, C. Ren, and R. Lougee (2015). “Big data fueled process management of
supply risks”. In: Proceedings of the 2014 Winter Simulation Conference, pp. 1005–1013.
Heckmann, I., T. Comes, and S. Nickel (2015). “A critical review on supply chain risk - Definition,
measure and modeling”. In: Omega 2015.52, pp. 119–132.
Henke, M. (2017). Digitalisierung als Chance - Kollaboratives Management und Wissenschaftskoop-
erationen für Smart Logistics. Hannover: Presentation.
Ho, W., Z. Zheng, H. Yildiz, and S. Talluri (2015). “Supply chain risk management: a literature
Jung, K., Y. Lim, and J. Oh (2011). “A model for measuring supplier risk”. In: British Journal of
Management 2011.22.
Kache, F. and S. Seuring (2017). “Challenges and opportunities of digital information at the in-
tersection of Big Data Analytics and supply chain management”. In: International Journal of
Operations & Production Management 37.1, pp. 10–36.
Kersten, W., M. Schröder, and M. Indorf (2015). “Supply Chain Risikomanagement für die Industrie
4.0”. In: Industrie 4.0 Management: Gegenwart und Zukunft industrieller Geschäftsprozesse 31.3,
pp. 36–40.
Kilubi, I. and H.-D. Haasis (2016). “Supply chain risk management research: avenues for further
studies”. In: International Journal of Supply Chain and Operations Resilience 2.1, pp. 51–71.
Kırılmaz, O. and S. Erol (2017). “A proactive approach to supply chain risk management: Shifting
orders among suppliers to mitigate the supply side risks”. In: Journal of Purchasing and Supply
Leveling, J., M. Edelbrock, and B. Otto (2014). “Big Data Analytics for Supply Chain Management”. In:
IEEE International Conference on Industrial Engineering and Engineering Management, pp. 918–
922.
Leveling, J., A. Schier, F. Luciano, and B. Otto (2014). “Konzeption eines proaktiven Risikomanage-
ments in Logistiknetzwerken”. In: Logistics Journal: Proceedings.
401
Li, Y., M. R. Kramer, A. Beulens, and J. van der Horst (2010). “A framework for early warning and
proactive control systems in food supply chain networks”. In: Computers in Industry 2010.61,
pp. 852–862.
Li, Y. and Y. Guo (2016). “Be proactive for better decisions”. In: International Journal of Disaster Risk
Reduction 2016.19, pp. 75–83.
Mangla, S. K., P. Kumar, and M. Kumar Barua (2014). “Flexible Decision Approach for Analysing
Performance of Sustainable Supply Chains Under Risk”. In: Global Journal of Flexible Systems
Mojtahedi, M. and B. L. Oo (2017). “Critical attributes for proactive engagement of stakeholders in
disaster risk management”. In: International Journal of Disaster Risk Reduction 2017.21, pp. 35–
43.
Park, Y.-B. and H.-S. Kim (2016). “Simulation-based evolutionary algorithm approach for deriving
the operational planning of global supply chains from the systematic risk management”. In:
Computers in Industry 2016.83, pp. 68–77.
Ponis, S. T. and A. C. Ntalla (2016). “Supply Chain Risk Management Frameworks and Models: A
Review”. In: International Journal of Supply Chain Management 5.4.
Qazi, A., J. Quigley, and A. Dickson (2015). “Supply Chain Risk Management: Systematic liter-
ature review and a conceptual framework for capturing interdependencies between risks”.
In: Proceedings of the 2015 International Conference on Industrial Engineering and Operations
Management. Dubai, United Arab Emirates.
Rajesh, R. (2016). “Forecasting supply chain resilience performance using grey prediction”. In:
Electronic Commerce Research and Applications 20, pp. 42–58.
Repenning, N. P. and J. D. Sterman (2001). “Nobody ever gets credit for fixing problems that never
happened”. In: California Management Review 43.4, pp. 64–88.
Rotaru, K., C. Wilkin, and L. Churilov (2014). “Proactive Risk Identification in Cross-organizational
Business Processes”. In: 25th Australasian Conference on Information Systems, Auckland, New
Zealand 2014, pp. 1–14.
Schlegel, G. L. (2015). “Utilizing Big Data and Predictive Analytics to Manage Supply Chain Risk”.
In: Journal of Business Forecasting 2014-2015.Winter, pp. 11–17.
Schlegel, G. L. and R. J. Trent (2015). Supply chain risk management: An emerging discipline. Boca
Raton, Florida: CRC Press.
Schlüter, F., K. Diedrich, and M. Güller (2017). “Analyzing the Impact of Digitalization on Supply
Chain Risk Management”. In: IPSERA Conference: International Purchasing and Supply Education
and Research Association, Budapest / Balatonfüred, Hungary.
Schlüter, F. and P. Sprenger (2016). “Mitigation framework for decentralised and proactive risk iden-
tification in a steel supply chain via industry 4.0 technologies”. In: 5th International Symposium
and 27th National Conference on Operation Research, pp. 85–91.
Schoenherr, T. and C. Speier-Pero (2015). “Data Science, Predictive Analytics, and Big Data in
Supply Chain Management: Current State and Future Potential”. In: Journal of Business Logistics
36.1.
Schröder, M., M. Indorf, and W. Kersten (n.d.). “Industry 4.0 and its impact on supply chain risk
management”. In: Reliability and statistics in transportation and communication 2014.
Seele, P. (2017). “Predictive Sustainability Control: A review assessing the potential to transfer
big data driven ‘predictive policing’ to corporate sustainability management”. In: Journal of
Cleaner Production 2017.153, pp. 673–686.
402
References
Seuring, S. and S. Gold (2012). “Conducting content-analysis based literature reviews in supply
chain management”. In: Supply Chain Management: An International Journal 17.5, pp. 544–555.
Sherwin, M., H. Medal, and S. Lapp (2016). “Proactive cost-effective identification and mitigation
of supply delay risks in a low volume high value supply chain using fault-tree analysis”. In:
Singh, G. and N. A. Wahid (2014). “Supply Chain Risk Management: A Review”. In: International
Journal of Supply Chain Management 3.3, pp. 59–67.
Singhal, P., G. Agarwal, and M. L. Mittal (2011). “SCRM: review, classification and future research
directions”. In: International Journal of Business Science and Applied Management 6.3, pp. 15–
42.
Sodhi, M. S., B.-G. Son, and C. S. Tang (2012). “Researchers’ Perspectives on SCRM”. In: Production
and Operations Management 21.1, pp. 1–13.
Tang, C. S. (2006). “Perspectives in supply chain risk management”. In: International Journal of
Production Economics 2006.103, pp. 451–488.
Tang, O. and S. N. Musa (2011). “Identifying risk issues and research advancements in supply chain
risk management”. In: International Journal of Production Economics 2011.133, pp. 25–34.
Thun, J.-H. and D. Hoenig (2011). “An empirical analysis of supply chain risk management in
the German automotive industry”. In: International Journal of Production Economics 2011.131,
pp. 242–249.
Vanany, I., S. Zailani, and N. Pujawan (2009). “Supply Chain Risk Management: Literature Re-
view and Future Research”. In: International Journal of Information Systems and Supply Chain
Vogel, W. and R. Lasch (2015). “Approach for Complexity Management in Variant-rich Product
Development”. In: Proceedings of the Hamburg International Conference of Logistics (HICL) -
Operational Excellence in Logistics and Supply Chains 2015, pp. 97–140.
Wang, G., A. Gunasekaran, E. W. T. Ngai, and T. Papadopoulos (2016). “Big data analytics in logistics
and supply chain management: Certain investigations for research and applications”. In: Int.
J.ProductionEconomics 176, pp. 98–110.
Wiendahl, H.-P., A. Slaouto, and R. Nickel (2008). “Proactive supply chain management in the
forging industry”. In: Prod. Eng. Res. Devel. 2008.2, pp. 425–430.
403
An Adaptive Supply Chain Cyber Risk
Management Methodology
Stefan Schauer1 , Martin Stamer2 , Claudia Bosse2 , Michalis Pavlidis3 , Haralambos

Mouratidis3 , Sandra König1 , Spyros Papastergiou4
1 – AIT Austrian Institute of Technology
2 – Fraunhofer CML
3 – University of Brighton
4 – University of Piraeus Research Center
Maritime information infrastructures have developed to highly interrelated cyber

ecosystems, where ports as well as their partners are connected in dynamic In-
formation and Communication Technology (ICT)-based maritime supply chains.
This makes them open and vulnerable to the rapidly changing ICT threat land-
scape. Hence, attacks on a seemingly isolated system of one business partner
may propagate through the whole supply chain, causing cascading effects and
resulting in large-scale impacts. In this article, we want to present a novel risk
management methodology to assess the risk level of an entire maritime supply
chain. This methodology builds upon publicly available information, well-defined
mathematical approaches and best practices to automatically identify and as-
sess vulnerabilities and potential threats of the involved cyber assets. This leads
to a constantly updated risk evaluation of each business partner’s cyber assets
together with their cyber interconnections with other business partners. The
presented risk management methodology is based on qualitative risk scales,
which makes the assessment as well as the results more intuitive. Furthermore, it
enables a holistic view on all of the integrated ICT-systems as well as their interde-
pendencies and thus can increase the security level of both a whole supply chain
and every participating business partner.
Keywords: IT security; cyber risk management; cyber risk assessment;

maritime supply chains
405
An Adaptive Supply Chain Cyber Risk Management Methodology
1 Introduction
For an organization, participating in a maritime supply chain implies not only

the need to cooperate with other stakeholders at business level, but due to the
ongoing digitalization also to set up interfaces in their information and commu-
nication technology (ICT) infrastructure for the ICT systems of their business
partners. Hence, these supply chains have become highly interrelated cyber
ecosystem, where the complexity and degree of networking of connected digital
assets beyond company borders increases. Nevertheless, every data interface
also represents a potential threat in form of a possible entry point for unplanned
access to the networks and the systems located behind it.
A global study among risk managers and risk experts rated cyber incidents as the
third highest business risk worldwide for all sectors and are expected to become
the highest business risk in the future. In Europe, cyber risks are rated already as
the second highest and in Germany as the highest business risk (Allianz Global
Corporate & Specialty SE, 2017).
So far, the number of disclosed cyber incidents in the transportation sector is
not very high and thus can be considered to be even smaller in maritime supply
chains (Verizon, 2017). However, companies might not report every attack due
to fears of reputational damage or - even worse - the attacks weren’t noticed
due to a lack of awareness and knowledge (Wingrove, 2017; Kotchetkova, 2015).
Considering the damage potential, vessels and ports might become an appealing
target for attackers in the future. The following incidents from the past illustrate
the bandwidth of possibilities: (a) Drugs were hidden in containers and these
containers were misled without early recognition (Bateman, 2013); (b) Customs
systems were shut down, stopping operations for hours, probably to extort ransom
(Port of Rotterdam, 2016); (c) Disruption of the GPS-signal stopped operations
of vessels as well as of terminal cranes that store and locate containers basing
on GPS for the same reason (Wagstaff, 2014; Scott, 2015; Hayes, 2016); (d) Piracy
attacks use AIS-signals to identify vessels and hack into the shipping companies
systems to identify their loaded goods (Allianz Global Corporate & Specialty SE,
2016); (e) Global ransomware campaign known as “WannaCry” and detected on
May 12, 2017, affected various organizations with tens of thousands of infections
in over 150 countries (US-CERT, 2017a).
Just a couple of weeks after the ”WannaCry” attack, on June 27, 2017, another
major global cyberattack (at some point linked to the existing ransomware ”Petya”,
but later on due to its additional features also referred to as ”NotPetya”) was
406
2 Regulations and Standards for Port Security
launched, using among other attack vectors the same exploit as ”WannaCry” (US-
CERT, 2017b; Fox-Brewster, 2017). It exploited a vulnerability in a Ukrainian tax
preparation software update mechanism to propagate and attack entire networks
(e.g. Cimpanu, 2017). Besides several Ukrainian ministries, banks and metro
systems, large companies became also victim of the attack. Among many others,
Beiersdorf AG, A. P. Moller-Maersk Group, Merck Sharp & Dohme (e.g. Holland,
2017) and India’s largest container terminal JNPT (e.g. PTI, 2017) were affected
and, as a consequence, had to deal with business interruptions. The malware’s
attack path leading from a Ukrainian software update to several international
company networks shows how malware can propagate among the connected ICT
systems in supply chains.
Due to these incidents, the general awareness for the need of cyber security and
cyber risk management increases and will rise further with every new mayor
security incident. Nevertheless, state-of-the-art risk management methodolo-
gies for maritime environments pay limited attention to cyber-security and do
not adequately address security processes for international supply chains. Moti-
vated by these limitations, we introduce the MITIGATE methodology, a novel risk
management approach, which will empower stakeholders’ collaboration for the
identification, assessment and mitigation of risks associated with cyber-security
assets and supply chain processes. This collaborative system will boost trans-
parency in risk handling, while enabling the generation of unique evidence about
risk assessment and mitigation.
The paper is structured as follows: Section 2 presents general regulations and
standards for port security. Section 3 provides a short overview on the MITIGATE
project while one of the project’s main outputs, the MITIGATE risk management
methodology, is described in Section 4. The key concepts of the MITIGATE method-
ology are sketched in Section 5 followed by a discussion, while section 7 concludes
the paper.
2 Regulations and Standards for Port Security
ICT systems of ports are classified as “Critical Information Infrastructures” (CII),

because ports are of crucial importance for the unrestricted supply, trade and
economy of a country. The EU adopted in July 2016 the Network and Information
System (NIS) Directive (EU, 2016). The directive aims to reach a common level of
security for NIS in the EU. This process will be supported by the European Union
407
Agency for Network and Information Security (ENISA) and protected by Computer
Security Incident Response Teams (CSIRT) all over Europe.
There are already several security guidelines in place, e.g., from the Baltic and In-
ternational Maritime Council BIMCO (BIMCO, 2017). They provide effective advice,
and awareness-rising posters for the use on board showing the need for security
measures. Further, they indicate how to avoid the biggest part of incidents by
giving striking rules for the use of passwords and private communication devices.
The International Maritime Organization (IMO) issued the ”Interim Guidelines on
Maritime Cyber Risk Management” in 2016 (IMO, 2016) and the U.S. promotes
”Information Sharing and Analysis Organizations” (ISAO), e.g., the ”Maritime &
Port Security Information Sharing and Analysis Organization” (MPS-ISAO, 2017).
Finally, the International Association of Classification Societies (IACS) in ship-
ping reacts to cyber threats with a ”Cyber Systems Panel” that was installed in
2016 (IACS, 2015). The focus of this panel lies on the early development of cyber
resilient onboard systems.
Beside these guidelines, there are also several standards and regulative which
address security and cyber security issues in maritime supply chains. Among
them, the most important is the International Ship and Port Facilities Security
(ISPS) Code (International Maritime Organization, 2003). The ISPS Code is a com-
prehensive set of measures to enhance the security of ships and port facilities,
focusing mainly on topics from the field of physical security and object protection.
Hence, a major drawback is the lack of specific tools, distinct measures or general
role descriptions tailored to the ICT security for port infrastructures. The main
objectives of the ISPS-Code with regards to ICT infrastructures are to ensure that
security communication is easily available and to prevent unauthorized deletion,
destruction or amendment of the security plans. Security plans may be saved in
an electronical format and therefore need to be protected.
An international standard specifically tailored to the field of ICT security is the

ISO/IEC 27001:2013 (International Standardization Organization, 2013). The ISO/
IEC 27001 is a commercial standard, representing a collection of best practices
and guidelines, describing how to establish, implement, maintain, monitor and
improve an Information Security Management System (ISMS). The standard is
generic in a way that the specified ISMS is applicable to organizations of various
types, sizes as well as different industries and markets. It should be noted that
ISO/IEC 27001 is actually not a risk management methodology, but rather a com-
pliance standard, reporting a list of controls for good security practices and the
408
3 MITIGATE Project
requisites that an existing method should have to be standard-compliant. Specif-

ically, it provides generic requirements that the risk analysis and management
needs to fulfill and references the ISO/IEC 27005 (International Standardization
Organization, 2011) (and further the ISO 31000 (International Standardization
Organization, 2009)) as a possible risk management methodology.
Although the ISO/IEC 27001 is applicable to several domains, the transportation

and logistics industry has introduced a common security management standard,
the ISO 28001:2007 (International Standardization Organization, 2007). Whereas
the ISO/IEC 27001 or the ISPS are focused on a single organization, the security
of the overall supply chain is the main objective of the ISO 28001. Therefore,
the standard includes the specific requirements to improve the security of all
aspects of the supply chain, including financing, manufacturing, information man-
agement and the facilities for packing, storing and transferring goods between
modes of transport and locations. As a specialty of the ISO 28001, all partners
involved in the supply chain need to sign a security declaration specifying their
currently implemented security measures to ensure a common security level over
the whole supply chain.
3 MITIGATE Project
As described in the previous Section, there are several standards and guidelines
at hand to prepare for cyber attacks and incidents. Nevertheless, a framework
dedicated to the assessment and management of cyber risks of maritime supply
chains has not been developed, yet. The ICT infrastructure of ports is particularly
vulnerable, due to comprising hard- and software assets of the companies en-
gaged in transport and goods handling in the maritime supply chain. Ports are
located at the interface of information flows from many different users and coun-
tries, which have to offer access and exchange capabilities for digital information.
However, all these interfaces also represent possible entry points for attackers.
The ongoing digitalization will result in even more complex and a higher degree
of networked ICT systems and so will the number of electronic interfaces to busi-
ness partner systems in supply chains increase, which cannot be supervised and
controlled by the single company.
In order to ensure that these processes and interconnections don‘t allow malware
to shut down operations or allow manipulation of data for illegal purposes, a
409
solution to identify threats along the supply chain and beyond company bound-
aries is urgently needed. The H2020 project MITIGATE (MITIGATE, 2016) is looking
in particular into security issues within the supply chain and aims at providing
tailored solutions for these problems. MITIGATE will introduce, integrate, validate,
evaluate and commercialize a risk management system for port infrastructures,
which will be able to deal with port CIIs and ICT systems, as well as their impact
on dynamic maritime supply chains. MITIGATE will emphasize the collaboration
of various stakeholders in the identification, assessment and mitigation of risks
associated with cyber-security assets and international supply chain processes.
This collaborative approach will boost transparency in risk handling by the various
stakeholders, while it will also generate unique evidence about risk assessment
and mitigation.
The collaborative approach of the project will be empowered by the MITIGATE
Open Simulation Environment enabling the participants to model, design, execute
and analyze attack-oriented simulation experiments using novel simulation pro-
cesses. Particular emphasis will be laid on the estimation of the cascading effects,
as well as on the prediction of future risks (based upon common metrics across
sectors). Relying on evidence–based simulations, port operators, decision makers
and other stakeholders will be able to select cost effective countermeasures and
compile holistic port security policies going beyond the ports’ CII isolated domain
to ensure the ports’ supply chain security.
Furthermore, the tools will be equipped with real-time decision support systems,
which will aim at automating the process of estimating risk and enacting risk
mitigation measures. MITIGAGE will integrate open source intelligence data (in-
cluding data from social networks and crowd-sourcing) towards enhancing its
threat assessment and prediction functionalities. At the heart of the MITIGATE
system will be a range of mathematical instruments, which will be used for threat
and vulnerability analysis, as well as for the assessment of contingency plans and
their cost-effectiveness.
4 Risk Management Methodology
As a core result of the MITIGATE project, the MITIGATE Risk Management Method-
ology has been developed. It aims at estimating the cyber risks for all assets of all
business partners involved in a maritime supply chain service (SCS) and represent
the basis for the MITIGATE system. The MITIGATE methodology is compliant with
410
Figure 1: Overview of the different steps of the MITIGATE methodology
the main standards for port security, the ISPS Code (IT Section), ISO 27001 and ISO
28001, which have been briefly described in the previous Section 2. Accordingly,
the six steps of the methodology (cf. Figure 1) represent the main steps also de-
scribed in these standards. In the following, we will present a high-level overview
on the different steps of the methodology going into detail on the central features
later on in Section 5.
4.1 SCS Analysis
In this first step, the scope of the risk assessment is defined. Therefore, the
business partners involved in the SCS under examination are identified. All the
business partners agree on the goals and the desired outcome of the risk assess-
ment. Further, the SCS under examination is decomposed and inspected in detail
by the business partner’s risk assessors who initiated the risk assessment. They
identify the participants of the SCS involved from their perspective, i.e., within
their organizations.
For each participant of the risk assessment, the main cyber and/or physical pro-
cesses (i.e., controlled/monitored by a cyber system) that comprise the examined
411
SCS are collected. The MITIGATE methodology is focusing in particular on the in-
terdependencies among these cyber assets. Therefore, these interdependencies
are further classified based on different types (e.g., whether they are installed
on the same system, communicating of network interfaces, etc.) describing the
relationship between the cyber assets in more detail.
The SCS analysis results in a list of all business partners together with their cyber
assets relevant for the SCS. Further, a graph of all cyber assets connected based
on their interdependencies is created.
4.2 SCS Cyber Threat Analysis
Based on the list of cyber assets created in the first step, all potential threats
related to these cyber assets are identified in the second step of the MITIGATE
methodology. Due to today’s rapidly changing threat landscape, the list of threats
needs to be as exhaustive and up-to-date as possible. To achieve that, the MIT-
IGATE methodology foresees the integration of multiple source of information,
i.e., online threat repositories like the National Vulnerability Database (NVD)(NIST,
2017), crowd sourcing and social media as well as the business partners’ experts.
This makes the methodology highly adaptive to novel attack strategies and at-
tacker behavior. The multitude of different data sources helps to increase the
quality of the whole risk assessment.
When the list of relevant threats is established, the likelihood of occurrence is

estimated for each of them. Also for this step, various sources of information
are combined: information from online repositories and social media is taken
into consideration as well as historical data and expert opinions. Instead of just
use one of these sources (e.g., relying only on historical data or expert opinions),
this approach offers the advantage of integrating a more diverse and complete
overview on the topic. Thus, the assessor obtains a more realistic estimation
of the threat likelihood. The resulting likelihoods are expressed using a semi-
quantitative, five-tier scale and all the gathered information is integrated. Finally,
a Threat Level (TL) based on this likelihood is assigned to each threat.
412
4.3 Vulnerability Analysis
Similar to the identification of threats in the previous step, in this step a list of
vulnerabilities of the cyber assets of the SCS under examination is compiled. In the
context of the MITIGATE methodology, a vulnerability is understood as a defective
state of a cyber assets due to a poor configuration, the lack of security patching,
etc. A threat can manifest in the SCS by exploiting a vulnerability of one of the
involved cyber assets.
The MITIGATE methodology differences between two main types of vulnerabilities:

confirmed vulnerabilities and potentially unknown or undisclosed vulnerabilities.
In more detail, vulnerabilities which are already know in the community and
are listed in online repositories or by specific Computer Emergency Response
Teams (CERTs) are understood as confirmed vulnerabilities. On the other hand,
there are vulnerabilities in software systems which are not publicly known, yet.
Such unknown or undisclosed vulnerabilities are more dangerous since security
experts are not aware of them but they can be (easily) exploited by adversaries.
A core feature of the MITIGATE methodology is to take these unknown and/or

undisclosed vulnerabilities into account. In this context, the data coming from
various information sources (online repositories, social media, expert knowledge,
etc.) is collected and processed to estimate the existence of unknown vulnerabili-
ties. In more detail, the analysis is carried out over all time scales in the available
dataset (e.g., by empirically characterizing the distribution of a vulnerability’s
lifespan) or determining the number of vulnerabilities publicly announced for a
specific period of time (e.g., using the rate of vulnerability announcements in the
NVD).
To characterize both confirmed and unknown/undisclosed vulnerabilities within

one methodology and make them comparable, the Common Vulnerability Scoring
System (CVSS) (Mell and Scarfone, 2007) is applied. For each vulnerability, the
Individual Vulnerability Level (IVL) is specified by assessing the Access Vector,
Access Complexity and Authentication. The scores for these three values are
coming from the online database NVD and are mapped onto a qualitative, five-tier
scale for further processing. The details on this mapping are given in section 5.1.
Additionally, the MITIGATE methodology is not only looking at the immediate ef-
fects of an attack exploiting a specific vulnerability but is also taking the respective
cascading effects into account. Therefore, the concepts of a Cumulative Vulnera-
bility Level (CVL) and a Propagated Vulnerability Level (PVL) are introduced. They
413
are described in detail in the following Section 5.2. Accordingly, the vulnerability
analysis results in a list of all vulnerabilities together with their respective IVL, CVL
and PVL.
4.4 Impact Analysis
After the vulnerability analysis done in the previous step, the MITIGATE methodol-
ogy is also looking at the potential impact an exploitation of these vulnerabilities
might have. To stay consistent with the vulnerability analysis, the CVSS (more
specifically, the three security criteria Confidentiality, Integrity and Availability) is
applied for assessing the impact. Accordingly, the scores for the security criteria
are also mapped onto the same qualitative, five-tier scale as the vulnerabilities
(cf. Section 5.1).
Furthermore, the notion of cascading effects is carried on for the impact analysis,
resulting in the concepts of Individual Impact Level (IIL), Cumulative Impact Level
(CIL) and Propagated Impact Level (PIL). Details on these impact levels are also
discussed in further detail in Section 5.2.
4.5 Risk Assessment
The risk assessment in the MITIGATE methodology is loosely based on the general
approach risk = likelihood × impact (Oppliger, 2015). Hence, in our context the
threat level (as described in Step 2, Section 4.2), vulnerability level (as described in
Step 3, Section 4.3) and impact level (as described in Step 4, Section 4.4) contribute
to the risk level. Further carrying on the notion of cascading effects, the MITIGATE
methodology describes three risk levels: Individual Risk Level (IRL), Cumulative
Risk Level (CRL) and Propagated Risk Level (PRL). This leads to the following
formula
IRL = T L × IV L × IIL
for the Individual Risk Level; the other two risk levels (CRL and PRL) are computed
accordingly. The overall result is then again mapped onto a qualitative, five-tier
scale.
414
5 MITIGATE Key Concepts
4.6 Risk Mitigation
In the final step of the MITIGATE methodology, the main results of the risk assess-
ment are compared against specific thresholds, which have been set and agreed
by all business partners. If some of the results exceed these predefined thresholds,
additional security controls need to be implemented by the business partners
and by the SCS (as a whole) to lower the respective risk levels. To identify the best
choice of mitigation actions out of a set of possible controls, a game-theoretic
approach is applied. This represents a mathematically sound method to find a
way to minimize the expected damage caused by an attack that exploits multiple
vulnerabilities.
To formalize the game, the possible actions taken by the adversary (i.e., a mali-
cious party performing an attack) and the defender (i.e., all business partners in
the supply chain) need to be identified. Any combination of these attack and de-
fense strategies yields a particular damage (i.e., the risk level), which is interpreted
as the respective payoff for this combination. Minimizing over all these damages
(i.e., the game’s payoff matrix) leads to the three main outcomes of this step: an
optimal attack strategy, an optimal defense strategy and the maximum risk level
for the case the attacker and defender both follow their optimal strategies.
The optimal defense strategy indicates which mitigation actions should be chosen
by all the business partners to minimize the damage to the entire SCS. Due to the
mathematical basis of game theory, it can be shown that even if the adversary
deviates from the optimal attack strategy, the business partners don’t have to
change their defensive strategy; a deviation by the adversary only manifests in a
lower maximum risk level as long as the defender plays his optimal strategy. We
describe this approach in more detail in section 5.3.
The MITIGATE methodology builds on three major concepts for the assessment of
the cyber risks within the SCS, which also represent the main research results of
the MITIGATE project. Further, the combination of these concepts also represents
the main difference and advantage of the MITIGATE methodology over existing
solutions. In the following, we will describe these three concepts in more detail.
415
Figure 2: Mapping of the CVSS metric ”Exploitability” onto the IVL
5.1 Semi-Automated Vulnerability Analysis
As already pointed out in previous sections, threats and attacks on cyber systems
have evolved drastically over the last years. An increasing number of more and
more complex attacks have been carried out and large companies as well as criti-
cal infrastructures have fallen victim to those attacks. One major reason for that
is the large number of vulnerabilities in software systems, which can be exploited
by malicious parties to circumvent security systems and infiltrate an organiza-
tion’s infrastructure. As mentioned in Section 4.3, unknown vulnerabilities are
the most critical ones in this context, because neither the users nor the creators
of a software system are aware of their existence.
Most of today’s risk assessment methodologies and frameworks are not able to
keep up this speed of evolving attacks and are not aware of the vulnerabilities
within examined systems. The MITIGATE methodology is able to adapt to this
fact and to build the risk assessment on top of a constantly updated vulnerability
database, i.e., the NVD. It is maintained by the National Institute of Standards
and Technology (NIST) (NIST, 2017) and updated frequently with the most current
information on numerous software systems. Further, the NVD applies the CVSS
to assess each vulnerability, providing an estimation of a specific vulnerability’s
relative importance, which further allows setting up a prioritization later on.
As described in Section 4.2 above, all the assets relevant for a specific SCS are
collected during Step 2 of the MITIAGTE methodology. In addition, information
on existing vulnerabilities is imported from the NVD on a daily basis and checked
against the identified assets. This results in a list of assets together with the latest
version of their vulnerabilities. Furthermore, the CVSS scoring given in the NVD
is mapped onto a five-tier scale, ranging from ”Very Low” to ”Very High”. The
resulting score represents the above mentioned Individual Vulnerability Level
(IVL) and is automatically assigned to every vulnerability of every asset in the SCS
(cf. Figure 2).
416
Figure 3: Mapping of the CVSS metric ”Impact” onto the IIL
Since the CVSS also estimates the consequences exploiting a specific vulnera-
bility may have on the Confidentiality (C), Integrity (I) and Availability (A) of the
underlying asset, this information is integrated into the Individual Impact Level
(IIL) by applying a similar mapping (cf. Figure 3).
Furthermore, also currently unknown vulnerabilities can be defined in the MITI-

GATE methodology for each asset. As already mentioned in Section 4.3 above,
this information is usually found by involving expert knowledge or interpreting
contributions in news feeds or social media. The MITIGATE methodology sup-
ports this activity by an automated search of the respective online sources and
highlighting potential relevant topics. Nevertheless, the assessment has to be
carried out by an expert but can be done using the CVSS metrics (or the five-tier
scale) as for to the known vulnerabilities. In this way, information coming from
different sources can be easily integrated into the same assessment process.
5.2 Cumulative and Propagated Risks
When looking at the vulnerabilities identified in the beginning of Step 3 (cf. Section
4.3), we have to be aware that the exploitation of one vulnerability may just be the
entry point of an adversary into a business partner’s infrastructure. For example,
using the enhanced access rights gained by the exploiting a specific vulnerability,
an adversary might be able to further navigate through the organization’s asset
network towards another (and maybe more profitable) target. In particular, this
is the case for Advanced Persistent Threats (APT)s. Therefore, the following two
views also need to be considered in the analysis of a specific vulnerability: on the
one hand, what are the possible ways (paths in the asset network) to reach that
vulnerability instead of attacking it directly (if that is possible at all). On the other
hand, after exploiting one vulnerability, what are the other possible vulnerabilities
417
an adversary is able to reach (e.g., due to additional privileges or access to other

assets).
The MITIGATE methodology accounts for both ideas by introducing the concepts
of Cumulative Vulnerability Level (CVL) and Propagated Vulnerability Level (PVL).
The goal of the CVL is to accurately reflect the exploitation level of the vulnera-
bilities by taking into consideration the IVL and the context within which these
vulnerabilities appear (i.e., the assets’ interdependencies). In other words, the
CVL measures the likelihood that an attacker can successfully reach and exploit a
vulnerability, given a specific path in the asset network. Such a path describes the
list of sequential vulnerabilities on different assets that arise from consequential
multi-steps attacks.
Whereas the CVL focuses on all possible attack chains concluding into the same
target point, the PVL inspects the likelihood that an attacker can penetrate a
network up to some specific depth. In other words, the PVL takes all possible
paths of sequential vulnerabilities of a specific length into account, starting from
one particular vulnerability.
Analogously to the CVL and PVL, the Cumulative Impact Level (CIL) and Propagated
Impact Level (PIL) are defined. As indicated by the naming, the only difference is
that in this case the potential impact of exploiting a specific vulnerability is as-
sessed. Carrying on as already mentioned in Step 5 of the MITIGATE methodology
(cf. Section 4.5), both concepts of vulnerability and impact are combined to result
in the respective notions of risk. Hence, the MITIGATE methodology outputs a
Cumulative Risk Level (CRL) and a Propagated Risk Level (PRL) together with the
IRL already mentioned in Section 4.5.
5.3 Attack Paths Discovery
Essential element of risk management, and of the MITIGATE methodology, is the

mitigation of risks through the identification of appropriate security controls. To
this end, attack paths are a valuable tool to business partners, illustrating paths
an attacker can use to reach a particular cyber asset. It can support the analysis
of risks to a specific cyber asset that may not be the entry point of an attack
and support the examination of possible consequences of a successful attack.
Moreover, the attack paths support the identification of appropriate security
controls by providing knowledge about attributes that make an attack possible.
The generated attack paths can answer ‘what-if’ questions regarding the security
418
implications of configuration changes to assets, such as patching a specific asset.

Furthermore, they can reveal which attacks can be performed by highly skilled
attackers and well-funded attackers and which attacks can be performed by low
skilled attackers.
The MITIGATE methodology includes an algorithm to discover attack paths. In

particular, it examines how an attacker can exploit identified cyber asset vulnera-
bilities in order to perform undesired actions. For every attack, a set of related
weaknesses (CWE) and vulnerability types are defined. It is assumed that to per-
form this kind of attack the attacker must have access to an asset that has one
or more vulnerabilities that are compatible with either the weaknesses or the
type defined. Attack paths are then modelled by employing attack graphs. Each
node in the graph represents a combination of asset and vulnerabilities that an
attacker can exploit. Each edge represents the transition of an attacker from one
asset to another.
The algorithm requires as input a physical network topology, an asset configura-

tion, a set of entry points and target points, and an attacker’s profile. In particular,
the network topology includes a list of cyber assets and their relationships. For
example, an asset may be installed on another asset or it just communicates with
another asset. The asset configuration includes information about a particular
asset. For example, the name of the asset, an id, the business partner to which
this asset belongs, its vulnerabilities, and attributes from the CVE repository, such
as access complexity and access vector. The entry point and the target points
are specific cyber assets on which a business partner wants to focus on. The
attacker’s profile includes information about the assumed attacker, such as the
attacker capability, which is the counterpart to a vulnerability’s access complexity
and the attacker location, which is the counterpart to a vulnerability’s access
location. The attackers profile is used to induce whether a particular attack can
exploit an asset’s vulnerability.
The output of the algorithm is a list of attacks paths. Each attack path contains
an ordered list of cyber assets that an attacker with a particular attacker’s profile
can successfully compromise by exploiting their vulnerabilities. Each cyber asset
in the attack path can be used as a stepping stone to an attack to the next cyber
asset. A business partner must be able to locate all potential attack paths into the
network and prevent attackers from using it. Business partners can hypothesize
new ‘zero-day” vulnerabilities of cyber assets, evaluate the impact of changing
configuration settings, and determining the security effectiveness of adding new
419
security controls. The identification of an optimal set of security controls, which

receives as input the generated attack paths, is described in the next section.
5.4 Game-Theoretic Risk Minimization
Besides identifying and assessing the vulnerabilities of assets and thus obtaining
a risk estimation based on latest threat information, mitigating these risks is an
essential part of the MITIGATE methodology. Whereas other approaches only
offer guidance on which mitigation actions to choose, the MITIGATE methodology
applies a game-theoretic framework to identify the optimal set of mitigation
actions.
The game is setup as a two-player zero-sum game, applying a minimax-approach

(Maschler, Solan and Zamir, 2013). To be more specific, the game describes the
combating situation between two players (in our case an adversary and the de-
fender, i.e., security officer) where each player tries to optimize his payoff. In a
zero-sum game, the gain of one player represents, at the same time, the loss of
the other player, which describes the real-life situation between an adversary
and the defender quite well. Both players have a set of strategies they can follow
and each strategy results in a specific profit for each player. These profits are
collected in the payoff matrix and the goal is to minimize the maximum profit
(i.e., minimax-approach) of the adversary. Thus, the strategies for both the ad-
versary and the defender are the central parts in the MITIGATE methodology.
The adversary’s strategies are defined by the paths through the asset network,
which the adversary is able to take to reach a specific vulnerability. These paths
have been defined in Step 3 (Section 4.3). The defender’s strategies are given
by the respective security measures a business partner is able to implement.
These countermeasures may come from the business partner’s experience or
can be deduced from the information stored in the NVD. Such a defense strategy
could be to do spot checking or patching of a specific asset (i.e., closing a specific
vulnerability).
Each combination of an attack and a defense strategy defines a scenario with a

specific payoff for both the adversary and the defender. In our context, this payoff
is the potential damage caused by the attack (represented by the IIL, CIL and PIL).
The adversary wants to maximize this damage; the defender wants to minimize it.
Since both the CIL and PIL representing the damage are based on the potential
paths an adversary can take in the asset network, the effect of a defense strategy
420
Figure 4: Example of a payoff matrix for the game with attack strategies a1 to a3
and defense strategies d1 to d3 (cf. Schauer et al., 2016).
is modeled by closing some vulnerabilities and thus eliminating some of these

paths. In general, every scenario will consist of multiple paths, each one causing
a specific damage. The best way to represent the collection of all these damages
without losing any information is to use a histogram (Rass, König and Schauer,
2015).
The payoffs for all scenarios are collected in the payoff matrix, which is used to
evaluate the game (cf. Figure 4 for an example). Since we are using histograms as
payoffs, we are going beyond standard game theory and have to apply a novel
framework (Rass, 2015; Rass, König and Schauer, 2015) to solve the game. The
game yields the three main outputs of the risk minimization step: the first is an
optimal attack strategy, i.e., a selection of the identified attack strategies, together
with a probability for each strategy to be played. Following these strategies causes
the maximum amount of damage to the infrastructure (worst case). The second
result is an optimal defense strategy, i.e., a subset of all possible security measures
together with a probability for each strategy. Implementing this strategy protects
421
the infrastructure against the optimal attack strategy. The third result is maximum
damage (characterized by the maximum risk level) an adversary can cause, if the
optimal attack strategy and the optimal defense strategy are implemented.
6 Discussion
The MITIGATE risk management methodology, as presented in the previous sec-

tions, provides a structured approach for maritime information infrastructures
to be prepared for today’s rapidly changing threat landscape and the associated
challenges. Due to the automated integration of publicly accessible information
on threats and vulnerabilities, the estimation of potential risks within the infras-
tructure is updated on a daily basis. Hence, the methodology is able to adapt
quickly to novel threats or incidents and deliver an accurate risk assessment.
The collection and processing of alternative information sources (e.g., social
media), as sketched in Section 4.3, allows to include also possible future (i.e.,
currently unknown) vulnerabilities into the risk assessment. Although this is an
integrated feature of the MITIGATE methodology, the expert knowledge of a risk
officer is still required to evaluate the gathered data. Nevertheless, this marks
an additional step towards an adaptive risk management framework suitable for
todays’ complex and highly dynamic threats.
The application of a game-theoretic approach to identify the optimal mitigation
actions represents an additional benefit of the MITIGATE methodology over other
methodologies and frameworks in this field. Whereas generally the question
which mitigation actions to implement in the end is often left to the risk office, our
methodology outputs an optimal set of security measures to be implemented.
Moreover, the methodology indicates, how often (i.e., at which frequency) the
respective actions have to be carried out.
Nevertheless, the MITIGATE methodology strongly relies on existing information
about the infrastructure of an organization. In particular, the information about
the setup of the supply chain service and about the involved assets gathered in
the first two steps (cf. Section 4.1 and Section 4.2) needs to be as exhaustive and
as complete as possible. This information can be taken from network scanning
tools, existing documentation or expert knowledge, but is created outside of the
methodology. Additionally, the set of available mitigation actions also needs to
be as accurate and complete as possible. Only in that case, all scenarios possible
422
7 Conclusion
in real life are evaluated in the game and the resulting defense strategy will reflect
a realistic setting. In general, the quality of the results heavily depends on the
quality of this information serving as input to the methodology.
Over all, the MITIGATE methodology has the ability to increase the security and
risk awareness not only within ports or other maritime information infrastructures
but also among the various business partners involved in maritime supply chains.
In the end, this is a first - and maybe the most important - step to effectively and
persistently raise the security level in these organizations.
7 Conclusion
Risk management is a core duty of maritime information infrastructures, in partic-

ular when considering the rising number of security incidents all over the world.
The MITIGATE methodology represents a supply chain risk management frame-
work going beyond state-of-the-art standards and guidelines. To this end, it inte-
grates an effective, collaborative, standards-based risk management approach,
which considers up-to-date information on all threats and vulnerability arising
from the supply chain, including threats associated with ports’ interdependencies
and their potential cascading effects.
Although the MITIGATE methodology integrates several open intelligence sources
and thus can quickly adapt to upcoming threats, the results are only as good as the
input data. Especially when it comes to the interdependencies between the cyber
assets within and among business partner’s organizations, expert knowledge
is required to model these relations correctly. Additionally, the experts need
to identify a level of abstraction when analyzing the cyber assets within their
organization since not all cyber assets within the organization will be relevant for
the SCS and thus not all of them need to go into the analysis.
The MITIGATE methodology is currently being implemented in a collaborative sys-
tem (http://mitigate.euprojects.net/) to enable all business partners within a SCS
to perform their cyber risk assessment in context of the entire supply chain. By
the end of the MITIGATE project, a large number of port operators, maritime stake-
holders and security experts will have been engaged in the process of evaluating
the capacity of the MITIGATE methodology and system.
Based on this evaluation, future research in this field will examine the question,
how collaborative aspects of the methodology can be facilitated and strengthened
423
as well as how small companies without specific IT related knowledge can be

further supported in the security management of their cyber assets.
Acknowledgements
We would like to thank our partners in the MITIGATE project, who also con-
tributed to the development of the MITIGATE methodology, in particular Nineta
Polemi,Thanos Karantzias, Christos Douligeris and Evangelos Rekleitis.
This work is supported by the European Commission’s Project No. 653212, MITI-
GATE (Multidimensional, IntegraTed, rIsk assessment framework and dynamic,
collaborative Risk ManaGement tools for critical information infrAstrucTurEs)
under the Horizon 2020 Framework Programme (H2020-DS-2014-1).
References
Allianz Global Corporate & Specialty SE (2016). Safety and Shipping Review 2016.
Allianz Global Corporate & Specialty SE (2017). Allianz Global Risk Barometer Top Business Risks
2017.
Baltic, T. and I. M. C. (BIMCO) (2017). The guidelines on cyber security onboard ships. Version 2.0.
Bateman, T. (2013). Police warning after drug traffickers’ cyber-attack.
Cimpanu, C. (2017). Petya Ransomware Outbreak Originated in Ukraine via Tainted Accounting
Software.
Classification Societies (IACS), I. A. of (2015). IACS Council 72 Press Release 14 Dec 2015 - IACS.
(EU), E. U. (2016). Directive (EU) 2016/1148 of the European Parliament and of the Council of 6
July 2016 concerning measures for a high common level of security of network and information
systems across the Union.
Fox-Brewster, T. (2017). Petya Or NotPetya: Why The Latest Ransomware Is Deadlier Than WannaCry.
Hayes, G. (2016). GPS can be jammed and ’spoofed’–just how vulnerable is it?
Holland, M. (2017). Rückkehr von Petya – Kryptotrojaner legt weltweit Firmen und Behörden lahm.
(IMO), I. M. O. (2016). Interim guidelines on maritime cyber risk management.
International Maritime Organization, ed. (2003). ISPS Code: International Ship and Port Facility
Security Code and SOLAS amendments adopted 12 December 2002. 2003 ed. OCLC: ocm51823054.
London: International Maritime Organization.
424
References
International Standardization Organization (2007). ISO 28001: Security management systems for
the supply chain - Best practices for implementing supply chain security, assessments and plans -
Requirements and guidance. English Version. Geneva, Switzerland.
International Standardization Organization (2009). ISO 31000: Risk Management – Principles and
Guidelines. English Version. Geneva, Switzerland.
International Standardization Organization (2011). ISO/IEC 27005: Information technology - Security
techniques - Information security risk management. English Version. Geneva, Switzerland.
International Standardization Organization (2013). ISO/IEC 27001: Information technology - Security
techniques - Information security management systems - Requirements. English Version. Geneva,
Switzerland.
Kotchetkova, K. (2015). Maritime industry is easy meat for cyber criminals.
Maschler, M., E. Solan, and S. Zamir (2013). Game Theory. Cambridge University Press.
Mell, P. and K. Scarfone (2007). A Complete Guide to the Common Vulnerability Scoring System.
MPS-ISAO (2017). Maritime & Port Security ISAO | Operationalizing Cyber Resilience. 2016.
Oppliger, R. (2015). “Quantitative Risk Analysis in Information Security Management: A Modern
Fairy Tale”. In: IEEE Security Privacy 13.6, pp. 18–21.
Port of Rotterdam (2016). How the Port of Rotterdam is investing in cybersecurity.
PTI (2017). New malware hits JNPT operations as APM Terminals hacked globally | The Indian Express.
Rass, S. (2015). “On Game-Theoretic Risk Management (Part One) – Towards a Theory of Games
with Payoffs that are Probability-Distributions”. In: ArXiv e-prints.
Rass, S., S. König, and S. Schauer (2015). “Uncertainty in Games: Using Probability-Distributions
as Payoffs”. In: Decision and Game Theory for Security. Lecture Notes inComputer Science 9406.
London, UK: Springer, pp. 346–357.
Schauer, S., S. König, S. Rass, and M. Latzenhofer (2016). “Spieltheoretische Risikominimierung in
IKT Infrastrukturen”. In: DACH Security 2016. Klagenfurt, Austria: syssec, pp. 174–187.
Scott, L. (2015). Protecting Position in Critical Operations.
Standards, N. I. of and T. (NIST) (2017). National Vulnerability Database (NVD).
(US-CERT), U. S. C. E. R. T. (2017a). Alert (TA17-132A) Indicators Associated With WannaCry Ran-
somware.
(US-CERT), U. S. C. E. R. T. (2017b). Alert (TA17-181A) Petya Ransomware.
Verizon (2017). 2017 Data Breach Investigations Report.
Wagstaff, J. (2014). All at sea: global shipping fleet exposed to hacking threat.
Wingrove, M. (2017). Ships are already under cyber attack.
425
Simulation-Framework for Illicit-Goods
Detection in Large Volume Freight
Lutz Kretschmann1 , Torsten Münsterberg1
1 – Center für Maritime Logistik und Dienstleistungen (Fraunhofer CML)
Innovative non-intrusive inspection technologies can help customs prevent illicit

trade in large volume freight. Validation whether technologies fulfill their intended
purpose ideally takes place under real conditions. However, constraints limit
the number and type of experiments performed during such field trials. Against
this background simulation offers the opportunity to evaluate improvements in
detection of illicit-goods without interrupting activities on site. A discrete event
simulation framework in the context of large volume freight is introduced in this
paper. It provides the means to compare alternative detection architectures -
combinations of different detection technologies – regarding their effectiveness
in identifying illicit goods in containers while at the same time the flow of goods
through security checkpoints can be analyzed. The framework is applied to an
exemplary case study comparing a single device detection architecture with a two
device system. Results highlight the somewhat counter intuitive logic that adding
a second device to the detection architecture either reduces the overall false clear
probability at the cost of a higher false alarm rate or vice versa. Further the impact
that adding another layer of detection has on the flow of containers through the
detection architecture and in particular on process time is discussed. Findings
described here are only a first step towards building a comprehensive simulation-
framework for illicit-goods detection in large volume freight. Nonetheless, they
illustrate how modelling and simulation can help customs identify the optimal use
of innovative detection technologies to increase overall security at EU-borders.
Keywords: maritime security; illicit trade; detection architecture; discrete event

simulation;
427
Simulation-Framework for Illicit-Goods Detection in Large Volume Freight
1 Introduction
Container transport has become an essential part of global supply chains and
merchandise trade connecting Europe with markets around the world. Since
ports play a key role within the European transport network, maintaining a high
level of port efficiency is of great importance for continued economic prosperity
throughout the EU.
Besides its integral role in the movement of legitimate cargo between countries
intermodal container transport is also exploited in illegal trafficking activities
(WCO, 2016). Table 1 shows the main categories of illicit trade as defined by Inter-
pol (2014). The unlawful transport of illicit goods poses a threat to security, the
environment and the economy. Container shipments are potentially misused in
various ways including the movement of security sensitive goods like weapons,
explosives or even radioactive materials as well as illegal trade in drugs, coun-
terfeit products and environmentally sensitive goods which negatively affects
citizens directly or indirectly. Moreover illegal trade is responsible for a loss of tax
revenues, causes market distortion and can be associated with damages to the
environment. For a discussion of negative socio-economic impacts of illicit trade
see e.g. Hintsa and Mohanty (2014). Customs, as government organizations which
control and administer the movement of goods across borders, play an important
role in securing international supply chains by preventing illicit trade in large
volume freight. Excluding intra-European trade the total number of containers
imported to the EU in 2014 was 17.5 million TEU of which the vast majority is of
legitimate nature (World Shipping Council, 2017). However, the sheer number
of containers moved across EU borders highlights the challenge customs face in
effectively preventing illicit trafficking without interfering unduly in the legitimate
flow of goods. Contributing further to the difficulty of this task is the wide variety
of risk materials that need to be detected both effectively and efficiently.
It is generally agreed that physically inspecting 100 percent of containers for illicit
content is not practical. Instead detection technologies play an essential role
for customs administrations to meet their responsibilities. Such non-intrusive
inspection (NII) technologies enable a detection of possible anomalies within the
container without having to open it.
At this point image based detection technology (X-ray) plays the central role in NII
for most threats. However, several research projects currently underway will result
in important advancements in NII technologies for checking maritime containers
and other large volume freight for anomalies in port and at land borders (see e.g.
428
1 Introduction
Table 1: The main categories of illicit trade
Main sectors of illicit trade

Illicit trade in chemical, biological, radiological, nuclear and explosive
material
Illicit trade in arms and weapons
Illicit trade in narcotic drugs and psychotropic substances
Illicit trade in environmentally sensitive goods
Intellectual property crime
Pharmaceutical crime
Illicit trade in excisable products
Illicit trade in cultural property
ACXIS project, 2016; C-BORD project, 2017; CRIM-TRACK project, 2014). Through
these initiatives a number of new and improved NII technologies will become
available which customs can deploy at their respective border crossing sites in
future.
As the number of principally suitable NII systems increases, identifying the opti-
mal set of technologies for a particular border crossing gets ever more complex
and thus challenging. More choices have to be evaluated under a specific set of
requirements and constraints to identify which combination of technologies has
the best cost to benefit ratio. In order to support this decision process, this paper
introduces a discrete event simulation framework in the context of large volume
freight. It enables customs to compare combinations of different NII technologies
regarding their effectiveness and efficiency to detect illicit goods. At the same
time logistical constraints at a particular border crossing point can be taken into
consideration and the impact of new NII systems on the flow of container through
security checkpoints can be analyzed.
The remainder of this paper is structured as follows. Section 2 introduces the
term detection architecture and describes how the performance of a detection
architecture can be described in terms of probabilities. Subsequently Section 3
gives an overview of different approaches in literature which analyze detection ar-
chitectures in order to support their design. The developed simulation-framework
for illicit-goods detection is introduced in Section 4 and applied to an exemplary
case study in Section 5. Concluding remarks and possibilities for future research
are given in Section 6.
429
2 Detection Architecture
Since customs face a wide variety of possible threats and NII technologies vary in
their effectiveness of detecting different threats, ideally they are integrated in a
detection architecture which makes sure that each technology is used according
to its particular strength. The Customs Detection Technology Expert Group defines
a detection architecture as “a construction of individual detection processes into
a defined structure in order to determine whether a consignment is ‘legitimate’
or ‘illicit”’ (CDTEG, 2014).
In accordance with the WCO SAFE Framework of Standards detection architectures
implemented by EU customs follow a risk based approach (WCO, 2012). This
means that available information and intelligence relating to a cargo shipment is
used to identify potentially high-risk containers in a first screening process. These
high-risk containers are selected for one or more subsequent scanning processes
where further information about the content of the container is collected e.g. by
means of NII equipment. In case an anomaly is sustained based on scanning
results the container is unpacked in a last step and the presence of illicit goods
is checked by physical inspection. If a non-conformity is found the goods are
stopped or otherwise released for further transport (see figure 1).
2.1 Objectives of Detection Architecture Design
When identifying a suitable detection architecture for a specific border crossing

customs have to do justice to conflicting objectives. In the case a shipment is
falsely cleared and thus a threat enters the country, society has to absorb the
negative socio-economic effects associated herewith. Accordingly customs have
to make sure certain acceptable detection thresholds for relevant threats are
met.
In order to increase the likelihood of detection, one option would be to implement
a more extensive screening, scanning and physical inspection regime (in effect
inspecting a larger share of containers). Alternatively thresholds could be lowered
beyond which the shipment is escalated to the next detection process instead of
being released (in effect already a smaller suspicion leads to further inspection).
However both will result in more disruptions to the flow of legitimate shipments
and increase the number of cases where scanning indicates a threat that does
not exist (false alarms).
430
2 Detection Architecture
Informa!on
Goods stopped
___
________
________
________
___
Anomaly
confirmed
Anomaly Anomaly
perceived sustained
Physical
Goods Screening Scanning
Inspec!on
Anomaly relieved
No anomaly Goods released Anomaly

found relieved
Figure 1: Detection architecture (CDTEG, 2014)
431
Any (unnecessary) inspection of legitimate cargo and in particular false alarms will
be perceived as a nuisance by the parties involved in the transport chain. More-
over, ports with overlapping hinterlands stand in fierce competition for cargo with
each other. If customs procedures in one port are extensive resulting in possible
delays, generalized transport cost of moving good through this particular port will
increase compared to other ports. This negatively affects the competitive position
of the port and cargo is potentially shifted towards ports with lower standards (De
Langen and Nijdam, 2008). Thus, customs have to maintain an appropriate level
of security for one thing but also ensure that the impact of checking containers
for illicit goods on the flow of freight is kept as low as possible.
2.2 Detection Capability of Detection Architectures
In order to describe the detection capability of individual scanning technologies

as well as that of single or multiple-device detection architectures probability
models can be used. In the simplest case an individual detection process can
lead to two different outcomes: the presence of an anomaly in the container is
sustained (alarm) or the presence of an anomaly in the container is relieved (clear).
Following Kobza and Jacobson (1997) four possible cases have to be distinguished
in this context (see figure 2).
To describe a devices capability of detecting a particular threat, probabilities of
Type I and Type II errors can be used. The Type I error probability is the probability
that a device raises an alarm under the condition that no threat is present in the
container or P(Alarm|NoThreat). The Type II error probability on the other hand is
the probability that a device does not raise an alarm under the condition
that a threat is present in the container or P (N oAlarm|T hreat). The lower
both conditional probabilities are the better is the performance of the device.
Expected values of Type I and Type II error probabilities can be determined exper-
imentally or, where test data is not sufficiently available, have to be estimated
based on expert knowledge and information from system insiders (e.g. customs,
technology providers).
The probability of a false alarm as well as the probability of a false clear is further
influenced by the prevalence of non-conformities (probability that a threat is
present in a container or P (T hreat)) in the population that is being examined.
Accordingly:
432
3 Approaches for Analyzing Detection Architectures
Figure 2: Possible results of detection process
P (F alseAlarm) = P (Alarm|N oT hreat) ∗ P (N oT hreat)

P (F alseClear) = P (N oAlarm|T hreat) ∗ P (T hreat)
Additionally a distinction between a device alarm and a system alarm is necessary

for multiple device detection architectures:
— Case A – a system alarm is triggered if any device in the detection archi-

tecture raises an alarm
— Case B – a system alarm is triggered if all devices in the detection archi-
tecture raise an alarm
Probabilities of a false alarm respectively a false clear for a multiple device de-
tection architecture under Case A and Case B are the combined conditional error
probabilities of the individual devices.
Modelling and simulation methods increasingly gain acceptance in the overall

security research landscape and particularly as effective tools for finding optimal
detection architectures.
433
The primary use of discrete event simulation models, such as the simulation-
framework for illicit-goods detection in large volume freight described in this
paper, in this context is twofold:
— Simulate the detection performance of detection architectures

— Simulate the flow of goods through security checkpoints
Discrete event simulation can support the design process of adequate detection
architectures by offering the opportunity to evaluate and assess improvements
in detecting illicit goods without interrupting activities on site. Operational ef-
fectiveness, efficiency and detection performance of different system designs
can be compared and analyzed according to the priorities and requirements at a
specific border crossing. Moreover, as pointed out by Wilson (2005), the approach
also allows determining the impact of different technology set ups on operation,
logistics and cargo flow, identify bottlenecks and serves as a basis for specifying
resources needed.
Several fields of research are relevant for the work presented in this paper which
are briefly discussed in the following. Overall the interest in studying the de-
sign of detection architectures intensified significantly after the terrorist attacks
of September 11, 2001 in New York (9/11). Understandably a large part of pre-
vious work deals with aviation and airport security as well as the detection of
nuclear material smuggling as these are scenarios directly associated with terror-
ist threats.
Due to changes in aviation security after 9/11 explosive detection systems for 100
percent scanning of checked baggage were deployed at all US airports. Jacobson
et al. (2006) apply probability modeling to evaluate the cost effectiveness of sev-
eral single- and two-device explosive detection architectures at airports. While
the work in this paper follows a similar approach to model the detection perfor-
mance of detection architectures it focuses on a different transport system: sea
ports and maritime container transport. Further it makes use of an event based
simulation method which allows gaining insights into the dynamic performance
of a detection architecture and thus reach a better understanding of bottlenecks
and spikes in the workload
Comparable demands for a 100 percent scanning regime have also been enacted
for containers entering US ports with the goal to prevent potential terrorist attacks
on a port which would have devastating economic impacts (U.S. Congress, 2007).
Several studies dedicated to 100 percent container scanning can be found in
434
literature. Martonosi et al. (2005) apply a cost-benefit analysis methodology in

this context to compare different scanning regimes. They conclude that switching
to 100 percent scanning is not cost-effective for most scenarios unless an attack is
very likely to occur. Another example can be found in the work done by Bakshi et al.
(2011) who develop a discrete event simulation model. They calibrate their model
using historical data from two container terminals and assess the operational
impact of different inspection policies. While their work focuses on 100 percent
scanning and threats associated with terrorism there are some similarities with
the simulation framework described here with regards to methodology and the
modeling of detection processes. Further Bakshi et al. (2011) conclude that a
“one-size-fits-all” approach does not work for all terminals. This underlines the
need to develop a flexible simulation framework, as it is the case in this paper,
which can be applied to the workflow at different border crossings.
Other scholars including e.g. Gaukler et al. (2012) and McLay and Dreiding (2012)
apply operations research methods to the problem of detecting nuclear material
in cargo containers in port. Gaukler et al. (2012) compare the existing inspec-
tion system for nuclear materials detection in containers with two alternatives
which make use of radiography information to decide on the routing of containers
through the detection architecture. They find systems which utilize additional
information gained through X-ray to outperform the current approach for a wide
range of scenarios. The work by McLay and Dreiding (2012) deals with identifying
optimal strategies for escalating containers to a secondary scanning stage given
multiple devices in the first stage such that the overall detection probability is
maximized within certain budget constraints. In contrast to the research on linear
programming in the context of container scanning this paper focusses on event
simulation to analyze security issues and cargo screening problems.
Research comparable to this paper regarding the methodology applied – discrete

event simulation - has been done by Siebers et al (2009). They describe a first
approach to develop a cargo screening process simulator which enables customs
to identify optimal technology set ups given certain commodity-threat combi-
nations in order to maximize the likelihood to detect illicit cargo. In subsequent
work the outlined concept is applied to a case study for cargo screening facilities
in the Port of Calais (Siebers et al, 2011; Sherman et al., 2012). For a very specific
threat - clandestines trying to cross the UK border hidden in lorries – they compare
different methods for conducting a cost-benefit analysis one of which is discrete
event simulation.
435
4 Simulation-framework
EU border crossing sites are characterized by unique conditions and face individ-
ual challenges. Logistical settings such as available space and the specific flow of
containers at each border crossing is different. Further, illicit trade of goods is a
dynamic field with constantly changing routes and adjustments in modality, quan-
tities smuggled per load as well as concealment methods. Accordingly, customs
deal with different types of threat scenarios in terms of a combination of legiti-
mate cargo and illicit goods. Against this background, customs have to identify a
detection architecture which best fulfills their specific needs and requirements
while taking into consideration conflicting objectives (balancing security and the
impairment of legitimate cargo flow) within given constraints regarding budget,
capacity and available resources.
Considering the above, a systemic European solution how to make best use of
innovative NII technologies to prevent illicit trade and smuggling must be adapt-
able to different types of borders and be able to take into account the local risk
profile representing the flow of illicit goods and respective threats in one location.
A one-size-fits-all solution or methodology cannot cope with this challenge.
The simulation-framework for illicit-goods detection in large volume freight pre-

sented in this section is supposed to be a flexible tool which can help customs
configure the detection architecture and workflow concept for a specific border
crossing point. Discrete event simulation provides a practical way to compare
different technology set ups and combinations and analyze their respective de-
tection performance. Further the approach makes it possible to determine how
innovative NII technologies can be integrated best into the overall flow of contain-
ers through the detection architecture and evaluate important performance mea-
sures such as total lead times, waiting times experienced in the inspection process
or NII system utilization. Thus, the simulation framework supports identifying
economically and practically effective technology combinations and scanning
sequences for a given threat profile at a certain border crossing and within specific
local logistics and workflow requirements.
Any model is a simplified representation of a real world system. Accordingly

certain assumptions are necessary to transfer the actual system into a simulation
environment. In order to represent a detection architecture within the proposed
simulation framework the following three elements are defined and transferred
into the simulation model:
436
4 Simulation-framework
— Individual detection performance values for all detection processes

— Decision making logic linking individual detection processes
— Logistical process between individual detection processes
The detection performance value represents the probability that a particular

technology will correctly identify different commodity-threat combinations or
produce a false alarm. Within the simulation framework detection processes are
modeled as “servers” with dedicated properties. Servers are predefined modules
in the simulation software, which represent a generic process step. With regards
to the detection performance of individual technologies the logic of probability
models as introduced in Section 2 is applied. Accordingly an expected value
of the probability that an alarm is raised although no threat is present and an
expected value of the probability that no alarm is raised although a threat is
present is defined. If several threats or commodity-threat combinations are to be
modeled at the same time, respective detection performance values for different
technologies can be integrated in a detection rate matrix as proposes by Siebers
et al (2009).
The decision making logic between individual detection processes within the
simulation framework determines the routing of objects (containers) through the
model. In the simplest case a detection process has two possible outcomes: an
alarm is raised or the object is cleared. The decision making logic defines the next
step within the overall detection architecture for both outcomes. A more complex
decision making logic is required in case a detection process has more than two
possible outcomes (e.g. a high, medium or low detection sensor reading) or in
case additional information (e.g. from manifest or previous scanning process)
is taken into consideration to make a decision about the routing of the object
through the detection architecture.
The logistical process between individual detection steps describes the flow of
containers within the model. It includes all main logistical drivers of system
performance. The following parameters are defined:
— The rate at which containers arrive over time

— The time it takes for an object to move from one server to the next
— The capacity of individual servers (number of objects which can be han-
dled simultaneously)
— The time required for a server to complete a process
437
Probability distributions are used to characterize process and arrival times re-
spectively. Additionally site specific logistical constraints and characteristics such
as available space and the flow of containers at a particular border crossing can
be considered accordingly.
5 Case Study
This section gives results for an exemplary case study comparing a single de-
vice detection architecture with a two device system by applying the proposed
simulation-framework for large volume freight introduced in the previous Sec-
tion 4. Different performance measures are calculated for each scenario in order
to illustrate the tradeoff between reducing the overall false clear probability or
reducing the false alarm rate in case a second device is added to the detection
architecture. Additionally the logistical impact is highlighted by means of lead
time (total inspection time), waiting times experienced and system utilization. In
case the simulation framework is applied to an actual detection architecture both
can support customs management in making an informed decision with regards
to performance and cost factors.
The simulation framework is implemented in the discrete event simulation soft-

ware Enterprise Dynamics by the company INCONTROL. This particular software
was selected due to its high flexibility regarding nonstandard processes on the one
hand and the comprehensive library of modules for standard processes on the
other hand. Different simulation modules, like sources, queues, servers and
sinks, are used to build up the detection architecture and, where necessary,
amended with specific program code to match the desired process logic and
behavior. Sources are the origin of containers entering the detection process.
Queues represent waiting or storage areas, if for example containers line up in
front of a scanning facility, because only one container can be scanned at a time.
Servers take on the role of any time consuming process, like scanning, analyzing
or transportation. Sinks are the final destination of containers after completing
the detection process. Based on these modules detection architectures can be
created and modified quickly. Results of simulation runs are exported to Microsoft
Excel for a comprehensive analysis.
438
5 Case Study
5.1 Design
X-ray imaging detection technology is the central component in almost all cus-
toms detection architectures (CDTEG, 2014). In the case study it is represented
as “Scanning A”. Scenario A represents the single device detection architecture
where only Scanning A is used. For the two device detection architecture in Sce-
nario B and Scenario C, X-ray is complemented with another detection device
“Scanning B”. Due to logistical reasons (lower process time) Scanning B takes
place previous to Scanning A. Naturally other arrangements are possible. The
corresponding decision making logic between individual detection processes for
all three detection architectures is shown in figure 3.
All three scenarios are analyzed for a hypothetical border crossing point with a
simulation run having a time horizon of five years. The process of screening is not
modeled explicitly. It is assumed that containers entering the system had previ-
ously been selected as high-risk and thus forwarded to further scanning. Only one
type of threat is considered. Scanning results of each detection process are based
on the true classification (threat/no threat) of the container. Alarm respectively
clear probabilities are calculated as introduced in Section 2. Individual device
responses are assumed to be independent.
The actual share of containers arriving in a given port that contain a threat is
highly sensitive information. The same applies to detection performance values
of individual NII technologies. Conditional probabilities used in this calculation
are partially estimated based on discussions with customs organizations. Because
of the sensitive nature of some of the data used in the simulation the values itself
are not reported. For a validation results and behavior of the simulated system
were compared with available customs data. A further refinement of the model is
foreseen once customs feedback on simulation results becomes available.
It is not the intention of the case study to calculate the exact detection perfor-
mance. Rather results of the case study are supposed to illustrate the general
behavior of single compared to multiple device detection architectures. For this
purpose Scanning A is assumed to produce a higher rate of false alarms compared
to Scanning B. Scanning B on the other hand is assumed to produce a higher rate
of false clears compared to Scanning A. Physical inspection is considered to be
100 percent successful in detecting illicit goods.
About 250 000 container enter the detection architecture during each simulation
run. The arrival time is characterized by a negative-exponential probability distri-
439
Figure 3: Detection architectures considered in case study
bution. During nighttime the arrival rate is lower by the factor of ten. Process-time
distributions of individual servers in the model are approximated based on system
insider judgements about average, maximum and minimum process time and as
well as observations made at border crossing points. The capacity of individual
servers (number of objects which can be handled simultaneously) assigned to
Scanning A and physical inspection is defined in a way that waiting times during
the day are the exception. For Scenario B and Scenario C the capacity assigned to
Scanning A and physical inspection is not changed. Scanning B capacity, again,
is defined in a way that waiting times at associated servers are the exception.
Several different performance measures are calculated for each scenario:
— Number of false clears
— Share of containers sent to physical inspection
— Time containers spend in the detection architecture
— Utilization rate of physical inspection resources
5.2 Results
Table 2 shows important results of the simulation experiments for the three con-
sidered scenarios. The results are given as relative values compared to Scenario
A. Accordingly, Scenario A has a value of 1 for all three key indicators.
440
5 Case Study
Table 2: Results of case study
Scenario
A B C
Number of false clears1 1.00 0.62 12.20
Number of false alarms1 1.00 1.47 0.02
Physical inspections1 1.00 1.39 0.09
1 - relative to Scenario A
The results underline the previously described effect that adding a second device
to the detection architecture either reduces the overall false clear probability at
the cost of a higher false alarm rate or vice versa. In Scenario B, where any device
alarm results in a physical inspection, the number of false clears is reduced by
38 percent on the one hand, but on the other hand the number of false alarms
increases by 47 percent. Further the number of physical inspections increases as
well compared to Scenario A (plus 39 percent).
The reverse effect can be observed for Scenario C where any device clear results
in goods being released. The number of false alarms is reduced significantly
compared to Scenario A and the demand for physical inspections declines in the
same order of magnitude. However, at the same time the number of false clears
increases by a factor of 12 compared to the level calculated for Scenario A.
Accordingly, adding a second device to the detection architecture is associated

with a benefit in terms of reduced false clears or reduced false alarms. However,
it always comes with the price of increasing the other figure respectively.
The lead time of containers passing through the system for all considered sce-
narios is illustrated in figure 4. The boxplot shows the distribution of lead times
per container. The lower whisker indicates the 2.5 percent quantile, the upper
whiskers marks the 97.5 percent quantile. Values given in the figure represent the
time a freight forwarder can expect the overall inspection process to take, once
the container has arrived at the location where inspections take place. Adding a
second scanning device to the detection architecture increases the median lead
time in Scenario B (53 min) compared to Scenario A (31 min) because a larger
share of containers undergoes physical inspection. On the other hand median
lead times in Scenario C (28 min) are less than in Scenario A since Scanning B has
441
500
400
Lead me in min
300
200
100
0
Scenario A Scenario B Scenario C
Figure 4: Lead time for all containers
a shorter process time than Scanning A and a more containers are released after
the first scanning process.
The distribution of lead times shows a positive skew in all three scenarios. The
reason for particular long lead times is the physical inspection process. The
dispersion of lead times is highest in Scenario B where a large share of containers
undergoes physical inspection and, in addition, containers experience waiting
times where the physical inspection capacity is fully utilized. In contrast, Scenario
C is characterized by smaller deviations in lead time compared to Scenario A and
Scenario B.
Subsequently lead times for containers which did respectively did not go through
physical inspection are discussed in detail.
Figure 5 contains a boxplot only for those containers which were subject to physi-
cal inspection. The median lead time for these containers is 275 min in Scenario A,
334 min in Scenario B and 288 min in Scenario C. The difference between Scenario
A and Scenario C is approximately the time required for Scanning B. In Scenario B
the median of 334 min is already affected by the increased number of cases that
experience waiting times due to full utilization of physical inspection capacity.
Of course this effect could be reduced by an extension of the physical inspection
442
5 Case Study
Figure 5: Lead time for containers that went through physical inspection
443
120
100
80
Lead !me in min
60
40
20
0
Figure 6: Lead time for containers that did not go through physical inspection
facilities. The high values of the 97.5 quantile in Scenario A and Scenario B (900
min and 1140 min) are a combination of waiting times for physical inspection and
a long physical inspection process itself which occurs for some containers. The
latter effect also explains the position of the 97.5 percent quantile in Scenario
C. It is lower than in the other scenarios though since fewer containers undergo
physical inspection in Scenario C and accordingly it is less likely for a container to
experience waiting time prior to physical inspection.
If only the lead time of containers that did not go through physical inspections
are analyzed, the results look somewhat different (see figure 6). The median lead
time per container is shorter with 30 min in Scenario A, 51 min in Scenario B and
28 min in Scenario C. The difference of approximately 20 min between Scenario A
and Scenario B corresponds to the process time of Scanning B plus waiting time
experienced by some of the containers. Lead times on the upper bound (97.5
percent quantile) in Scenario A and Scenario C are around 70 min and 95 min.
The different detection system architectures designs also have a large effect on
the utilization of individual system components and in particular on the utilization
of the physical inspection process. In the base case, Scenario A, the median of
444
5 Case Study
100%
80%
U!liza!on rate
60%
40%
20%
0%
Figure 7: Average utilization rate of physical inspection per month
the average utilization rate per month at the physical inspection facilities is 48
percent (see figure 7). This is relatively high considering that the majority of the
containers arrive during daytime and only few during night. As discussed before,
the facilities for physical inspection are in little use in Scenario C where the median
utilization rate is 4 percent. In Scenario B, on the other hand, the overall number
of physical inspections is high. Consequently the utilization rate of the physical
inspection facilities goes up and containers frequently experience waiting times.
This increases costs either in terms of higher generalized transport costs or due to
an extension of the physical inspection capacity to reduce waiting times. On the
other hand the number of false negatives is lower in Scenario B which represents
a comparative benefit. Fewer shipments are falsely cleared, thus the cost which
society has to absorb in terms of negative socio-economic effects associated with
illicit trade is reduced. This underlines that comparing the cost to benefit ratio of
different detection architectures can be a suitable basis for decisions by customs
management and policy makers.
445
6 Conclusion
Overall the role of maritime container trade in the transportation of illegal goods
across borders is characterized by a comparatively low number of seizures but
generally larger consignments of illicit goods than in other transport modes (see
e.g. WCO, 2016). At the same time scholars have highlighted the potentially
devastating impact that terrorist attacks could have in the maritime context overall
and associated with maritime container transport in particular (see e.g. Greenberg
et al., 2006; Schneider, 2011). Both emphasizes the importance of constantly
improving detection architectures and thus effectively and efficiently preventing
illicit trade while facilitating free flow of legitimate cargo across EU-borders.
This paper introduces a simulation-framework for illicit-goods detection in large
volume freight. The framework represents an effective and flexible tool for cus-
toms to analyze different NII technology set ups - detection architectures – against
their requirements without interrupting activities at a border crossing point. At
the same time the applied method, discrete event simulation, provides the means
to take logistical settings such as available space and the specific flow of contain-
ers into account and analyze important performance indicators such as average
inspection times but also its variation e.g. expressed by inspection times in the
upper quartile.
An illustrative case study demonstrated how a detection architecture can be
analyzed with the simulation-framework. It compared a single device detection
architecture with two variants of a two device architecture regarding the overall
detection performance. Results show that by adding a second device to the
detection architecture it is not possible to increase the relative chance to detect a
threat (lower probability of false clears) and reduce the share of containers which
unnecessarily undergo physical inspection (lower probability of false alarms) at
the same time. It was demonstrated that a multi device detection architecture
where any device alarm results in physical inspection will reduce the probability
of false clears while a detection architecture where any device clear results in
goods being released will reduce the probability of false alarms compared to the
single device architecture.
Furthermore the impact of adding a second device on lead times and capacity
utilization was shown in the case study. A multi device detection architecture does
not inevitably mean that lead times increase. Whether they do or not depends
on the logical structure of the detection architecture. The same applies for the
utilization rate of e.g. the physical inspection resources which can either increase
446
References
or decrease compared to the single device architecture depending on the logic

linking individual detection processes.
Findings described here are only a first step towards building a comprehensive
simulation-framework for illicit-goods detection in large volume freight. There
are several possible extensions to this work. A first extension is to implement
different types of threats and specific capabilities of technologies to detect these
threats in the model in order to analyze the full range of illicit trade customs
deal with. A second extension is to consider information gained in one scanning
process in the decision how the container is routed through subsequent scanning
devices. This implicates a more complex decision making logic linking individual
detection processes which might also take into account the current utilization of
NII scanning equipment and physical inspection stations to reduce waiting times
during peaks. A third extension is to quantify direct and indirect cost associated
with different detection outcomes (true alarm, false alarm, false clear, true clear),
the detection process itself as well as increases of generalized transport cost due
to potential delays. This way changes in different cost factors could be weighed
against each other, adopting the concept of cost-benefit analysis, to provide a
more comprehensive basis for customs management and policy makers to take
informed decisions. Work is in progress to address these extensions.
This work has received funding from the European Union’s Horizon 2020 research
and innovation programme, under grant agreement No 653323. This text reflects
only the author’s views and the Commission is not liable for any use that may be
made of the information contained therein.
References
Acxis project (2016). ACXIS Automated Comparison of X-ray Images for cargo Scanning.
Bakshi, N., S. E. Flynn, and N. Gans (2011). “Estimating the Operational Impact of Container
Inspections at International Ports”. In: Management Science 57.1, pp. 1–20.
C-BORD project (2017). Effective Container Inspection at BORDer Control Points. www.cbord-h2020.eu.
CDTEG (2014). Detection Architecture. Brussels.
CRIM-TRACK project (2014). CRIM-TRACK. www.crimtrack.eu.
447
Gaukler, G. M., C. Li, Y. Ding, and S. S. Chirayath (2012). “Detecting Nuclear Materials Smuggling:
Performance Evaluation of Container Inspection Policies”. In: Risk Analysis 32.3, pp. 531–554.
Greenberg, M. D., P. Chalk, H. H. Willis, I. Khilko, and D. S. Ortiz (2006). Maritime Terrorism: Risk and
Liability. Santa Monica, CA: RAND Corporation.
Hintsa, J. and S. Mohanty (2014). “A Literature-Based Qualitative Framework for Assessment of
Socio-Economic Negative Impacts of Common Illicit Cross-border Freight Logistics Flows”. In:
Innovative Methods in Logistics and Supply Chain Management: Current Issues and Emerging
Practices. Ed. by T. Becker, W. Kersten, and C. M. Ringle. Berlin: epubli, pp. 319–340.
INTERPOL (2014). Countering Illicit Trade in Goods: A Guide for Policy-Makers. Lyon.
Jacobson, S. H., T. Karnani, J. E. Kobza, and L. Ritchie (2006). “A Cost-Benefit Analysis of Alternative
Device Configurations for Aviation-Checked Baggage Security Screening”. In: Risk Analysis 26.2,
pp. 297–310.
Kobza, J. E. and S. H. Jacobson (1997). “Probability models for access security system architec-
tures”. In: Journal of the Operational Research Society 48.3, pp. 255–263.
Langen, P. W. d. and M. N. Nijdam (2007). “Charging Systems for Waste Reception Facilities in Ports
and the Level Playing Field: A Case from North-West Europe”. In: Coastal Management 36.1,
pp. 109–124.
Martonosi, S. E., D. S. Ortiz, and H. H. Willis (2005). “Evaluating the viability of 100 per cent con-
tainer inspection at America’s ports”. In: The Economic Impacts of Terrorist Attacks. Ed. by H. W.
Richardson, P. Gordon, and J. E. Moore II. Cheltenham: Edward Elgar Publishing, pp. 218–241.
McLay, L. A. and R. Dreiding (2012). “Multilevel, threshold-based policies for cargo container
security screening systems”. In: European Journal of Operational Research 220.2, pp. 522–529.
Schneider, P. (2011). Maritimer Terrorismus: Tätergruppen und Anschlagstypen 1968 – 2010. PiraT
Arbeitspapier zur Maritimen Sicherheit Nr. 13. Hamburg.
Sherman, G., P.-O. Siebers, D. Menachof, and U. Aickelin (2012). “Evaluating Different Cost-Benefit
Analysis Methods for Port Security Operations”. In: Decision Making in Service Industries: A
Practical Approach. Ed. by J. Faulin, A. A. Juan, S. E. Grasman, and M. J. Fry. Boca Raton, FL:
CRC Press, pp. 279–303.
Siebers, P.-O., U. Aickelin, and G. Sherman (2009). “Development of a Cargo Screening Process
Simulator: A First Approach”. In: 21st European Modeling and Simulation Symposium (EMSS
2009), Volume I. Ed. by R. M. Aguilar, A. G. Bruzzone, and M. A. Piera.
Siebers, P.-O., G. Sherman, U. Aickelin, and D. Menachof (2011). “Comparing Decision Support
Tools for Cargo Screening Processes”. In: The 10th International Conference on Modeling and
Applied Simulation (MAS 2011). Ed. by A. Bruzzone, C. Frydman, M. Massei, M. McGinnis, M. A.
Piera, and G. Zacharewicz.
U.S. Congress (n.d.). H.R.1: Implementing recommendations of the 9/11 commission act of 2007:
Pub. L. No. 110-53, 110th Congress.
Wilson, D. (2005). “Use of modeling and simulation to support airport security”. In: IEEE Aerospace
and Electronic Systems Magazine 20.8, pp. 3–6.
World Customs Organization (2012). SAFE Framework of Standards to Secure and Facilitate Global
Trade. Brussels.
World Customs Organization (2016). Illicit Trade Report 2015. Brussels.
World Shipping Council (2017). Trade Statistics.
448
Relational View on Collaborative
Supply Chain Disruption Recoveries
Marie Brüning1 , Julia Bendul1
1 – Jacobs University Bremen
Due to their growing global and complex nature, supply chains are increasingly
vulnerable to natural and man-made disasters that disrupt the flow of goods.
Today, recovering from supply chain disruptions represents a major challenge
for supply chain professionals. In research, most recovery methods suggested
are based on redundancy and flexibility. In practice, a different approach gains
momentum: companies recover by collaborating with their supply chain partners,
especially by temporarily sharing resources. Based on the relational view theory,
this paper uses multiple case studies and expert interview to develop a framework
which describes the factors promoting collaborative resource sharing as well as
the effect on supply chain resilience.
Keywords: Supply Chain Risk Management; Supply Chain Resilience;

Supply Chain Collaboration; Shared Resources
449
Relational View on Collaborative Supply Chain Disruption Recoveries
1 Introduction
Supply chain (SC) disruptions are “unplanned and unanticipated events that dis-
rupt the normal flow of goods and materials within a supply chain” (Craighead et
al., 2007). In 2016, two of Volkswagen’s suppliers halted delivery of transmission
and seat parts, following previous unresolved disputes between the parties. This
led to stops of series production in seven German Volkswagen production sites.
Additionally, several other suppliers were affected as well (Reuters, 2016). This
example shows immediate consequences of SC disruptions, such as increasing
costs as well as decreasing profitability and net sales. Long-term effects of disrup-
tions are negative stock market reactions, damaged brand image and decreased
customer service levels (Bello and Bovell, 2012; Craighead et al., 2007; Hendricks
and Singhal, 2005).
The topic of managing SC disruptions becomes increasingly important due to

three reasons. First, quantities as well as severities of natural and man-made dis-
asters are expected to continue rising (Munich Re, 2016; SwissRe, 2016). Second,
current business trends in the area of supply chain management (SCM), such as
reduction of supplier base, just-in-time inventory system, outsourced manufactur-
ing, and global sourcing, create highly interconnected global SC with fewer buffers
and more risk exposure points. This increase the potential for and the impact of a
SC disruption (Bello and Bovell, 2012; Craighead et al., 2007; Stecke and Kumar,
2006; Thun and Hoenig, 2011; Tukamuhabwa et al., 2015). Third, the recovery
from SC disruptions is one of the main concerns of SC managers. However, it
seems that companies are not well prepared for a fast recovery (Deloitte, 2013).
Thus, there is a need to understand and implement risk management methods
that facilitate the recovery of disruptions (Bovell, 2012).
Collaborative recovery as a reactive risk management method is rather new to the

field of supply chain risk management (SCRM) and not yet holistically explained
in literature. SC partners gain collaborative advantage by sharing their resources
during disruption recoveries. This perspective is based on the relational view
theory which states that collaborative advantage can be gained by leveraging
inter-organizational resources available in the network (Dyer and Singh, 1998).
In this research specifically one element of supply chain collaboration (SCC) is
analysed, namely the sharing of resources. This paper aims to develop a collabo-
rative recovery framework that encompasses promoting factors of collaborative
resource sharing and its effect on supply chain resilience (SCRES). In addition,
managerial implications are derived.
450
2 Literature Review
The paper is structured as follows. Chapter two covers a literature review of the
concepts SCC and SCRES as well as the connection between both. In chapter
three the methodology is outlined. In chapter four the developed framework
is described in detail. Chapter five covers limitations, suggestions for further
research and managerial implications.
2 Literature Review
This research project builds on the relational view as a theoretical foundation.

The relational view theory, initially developed by Dyer and Singh (1998), is an
extension of the resource-based view. The trend that networks of collaborating
companies, instead of individual companies, compete against each other was the
basis for its development. Dyer and Singh (1998) state that inter-firm linkages and
inter-organizational resources may be a source of relational rents and collabora-
tive advantage. Relational rent refers to “supernormal profit jointly generated in
an exchange relationship that […] can only be created through the joint idiosyn-
cratic contribution of the specific alliance partners” (Dyer and Singh, 1998). One
of the determinants of relational rent, and thus of collaborative advantage, is
complementary resource endowment. It is defined as “distinctive resources of
alliance partners that collectively generate greater rents than the sum of those ob-
tained from the individual endowments of each partner” (Dyer and Singh, 1998).
Collaborative advantage refers to joint value creation and benefits gained over
competitors in the marketplace through supply chain partnering (Cao and Zhang,
2013).
Within SCRM literature, there are two categories of SC disruption management

methods: Preventive methods are used in advance of a disruption to reduce its
likelihood. Reactive methods are used if a disruption occurred, thus during the
recovery. The focus is to limit the severity of disruptions (Simchi-Levi, Schmidt
and Wei, 2014; Stecke and Kumar, 2006). In SCRM literature, reactive measures
receive considerably less attention than preventive measures (Tukamuhabwa et
al., 2015). Furthermore, most of the existing reactive methods, such as excess
inventory, multiple sourcing and rerouting of transports, focus on redundancy
and flexibility (Christopher and Peck, 2004; Stecke and Kumar, 2006). These
are not in line with the described current business trends as they involve high
inventory and coordination costs (Chopra and Sodhi, 2014).
451
In this research the term SCRES is used to describe reactive methods only. SCRES
is defined as the SC’s ability to recover from a disruption and restore to normal
operations in a timely manner more favourable than competing SC (Christopher
and Peck, 2004; Ponis and Koronis, 2012; Ponomarov, 2012; Wieland and Wallen-
burg, 2013). This definition implies the core idea of the relational view theory by
arguing that SC recover better, i.e. faster, than competing SC and can, therefore,
gain a collaborative advantage.
SCC is defined by Cao et al. (2010) as a “long-term partnership process where sup-
ply chain partners with common goals work closely together to achieve mutual
advantages that are greater than the firms would achieve individually”. The rela-
tional view theory is also reflected in this definition. Long-term SC relationships
create value that neither partner would have been able to create independently
(Nyaga, Whipple and Lynch, 2010). Companies seek to build collaborative relation-
ships because in this way they can access complementary resources and improve
collaborative advantage (Cao and Zhang, 2013; Dyer and Singh, 1998; Mohr and
Spekman, 1994).
According to Cao et al. (2010), resource sharing is an element of SCC and defined
as “the process of leveraging capabilities and assets and investing in capabilities
and assets with supply chain partners”. They do not discuss resource sharing in
the context of a time-limited event, such as a SC disruption recovery. Rather they
discuss resources which are continuously shared. Resources are a key component
of the relational view theory (Dyer and Singh, 1998). Lavie (2006) argues that
companies can obtain value from resources which are not controlled or fully
owned by their internal organization.
The topic of collaborative recovery can be located at the intersection of SCC and
SCRES because it is a reactive disruption management method which is based
on the effort of multiple companies within a SC. There are a few similar concepts
described in literature. The research of Bello and Bovell (2012) as well as Wieland
and Wallenburg (2013) focus on relational factors relevant to gain SCRES in collab-
orative relations. According to Bello and Bovell (2012), a gap exists in disruption
recovery literature regarding the attention paid to relational resources that enable
collaboration. Bovell (2012) highlight the value of collaborative resource sharing
with respect to SCRES. Bode et al. (2011) presents the reactive methods ‘bridging’
which focuses on collaborative actions, like establishing relationships with influ-
ential individuals in partner companies. Furthermore, they find that the levels
of trust and dependency are decisive in choosing a reactive method. Whitney et
al. (2014) discuss and analyze the method ‘temporary sourcing diversification’
452
3 Methodology
which captures the temporary nature of the recovery process. Overall, there is
a research gap with regard to reactive risk management methods that focus on
relational resources as well as on resources temporarily shared during disruption
recoveries. This research attempts to give some insights into this area.
3 Methodology
Developing a theoretical framework is understood as an iterative process. Kubicek

(1976) describes the attempt to derive practical knowledge by interacting with
practitioners, theoretical processing and reflection of the obtained knowledge
in order to derive new questions for practitioners as iterative heuristic. The the-
oretical framework was continually modified with experience and information
from a comprehensive literature review, case studies and expert interviews (for
an overview of the executed interviews and case studies see appendix).
Six case studies were conducted based on secondary data (Brüning, Hartono and
Bendul, 2014). The method is appropriate for this research because case studies
are valuable for rather unexplored topics (Eisenhardt, 1989; Voss, Tsikriktsis and
Frohlich, 2002). To ensure validity and reliability, established structured proce-
dures were employed, such as case study protocols and case selection criteria.
The case subjects had to have experienced a disruption that had consequences
for multiple SC actors and there had to be some kind of SCC during the recovery.
For the data collection different types of archival sources were consulted, such
as company reports, reports from recognized organizations, newspaper articles,
and academic literature.
Eight semi-structured expert interviews were executed. The dialog between in-
terviewer and interviewee reflects the iterative framework development process
(Kubicek, 1976). The interviewed experts are practitioners working in manufac-
turing industry, insurance or consulting, who are either dealing with or are/were
affected by SC disruptions. An interview guideline was developed (Flick, 2006)
based on the findings from literature review and case studies. The questions
mainly focus on the company’s experiences with collaboration during disruption
recoveries and the general collaboration in their SC. There was sufficient time
between the interviews to refine the guideline in between sessions. Eight expert
interviews approximately 30 minutes were carried out over a time period of six
months. The interviews were held in person or via telephone, depending on
the preference of the interviewee. Seven of the interviewees gave permission
453
for the recording of the interview via digital voice recorder. Each interview was
summarized in a protocol. The interviewees reviewed their protocols which, in
turn, were improved based on their feedback (Voss et al., 2002).
4 Framework Development
Based on the methods described, a framework was developed that encompasses

relevant factors to describe the risk management method collaborative recovery
(figure 1).
4.1 Collaborative Resource Sharing
Starting point for the framework development is the central construct ‘collabo-
rative resource sharing’. It describes the actual temporary sharing of resources
during disruption recoveries. The interviews and case studies were analysed
with regard to three topics, namely the types of resources shared, the resources’
abilities, and the actors sharing the resources (Brüning, Hartono and Bendul,
2015).
First, human resources (HR) and production resources (PR) are the two identi-
fied categories of resources shared during collaborative recoveries. HR include
employees, managers and engineers as well as their know-how and skills. In all
analysed cases HR were shared. For instance, to support Aisin’s recovery, Toyota
Figure 1: Theoretical Framework
454
sent around 400 employees from various departments (Nishiguchi and Beaudet,
1998). In the automotive and aerospace industry, original equipment manufac-
turers (OEM) have task forces which consist of skilled engineers and/or managers
(interviews AeroOEM1, AutoOEM1, AutoOEM2). They can be sent on site in case
of a SC disruption. PR cover equipment, tools, facilities, machine capacities and
warehouse capacities as well as information and know-how, for instance about
production processes. There are examples of companies in the food industry that
used machine and warehousing capacities of competitors during SC disruption
recoveries (interview Funk RMCE).
Second, adaptability, mobility and availability of capacity are the identified abili-
ties which enable resources to be shared during SC disruption recoveries. HR have
to be able to adapt to new tasks. For instance, in the case of Aisin, employees from
their suppliers and customers could quickly adapt to produce the needed parts
(Nishiguchi and Beaudet, 1998). Adaptability of PR refers to the ability of changing
tools, machines or warehouses according to new requirements. The ability ‘mo-
bility’ refers to HR’s and PR’s ability to move. In the case of Renesas, more than
2,500 engineers from other companies were mobilized and came to the disrupted
plant. In order to exchange information, recovery activity team leaders had daily
meetings to discuss the progress of the recovery (Renesas, 2011). Resources need
to have capacity available (Whitney et al., 2014). With regard to HR, employees
need to have capacity to take over new tasks. This may be accomplished with
additional shifts or with established task forces (interviews AeroOEM1, AutoOEM1,
AutoOEM2, AutoOEM3). For PR, it is required that machines or warehouses have
capacity available to take over production processes on short notice (interviews
AeroOEM1, Funk RMCE).
Third, the identified involved actors cover intra-organizational, horizontal as well
as vertical collaboration. Intra-organizational collaboration exists, for instance,
between different subsidiaries of one company (Barratt, 2004). During Nissan’s re-
covery, hundreds of employees from other Nissan plants worldwide were involved
in the repair work at the damaged facilities (Nissan, 2012). Vertical collaboration
with customers and suppliers (Barratt, 2004; Simatupang and Sridharan, 2002)
is most frequently used (interviews AeroOEM1, AutoOEM3, Controllit OHB, Funk
RMCE). Customers are the actors that are involved in all analyzed cases. Second
tier customers and second tier suppliers may also be involved in collaborative
recoveries, like in Aisin’s recovery (Whitney et al., 2014). These indirect relation-
ships are activated only for the time of the recovery (interviews AutoOEM1, Funk
RMCE). Horizontal collaboration refers to collaboration with competitors or other
unrelated organizations (Barratt, 2004; Simatupang and Sridharan, 2002). Collab-
455
oration with competitors during recoveries happens rarely (interviews AutoOEM1,

AutoOEM2) due to antitrust and privacy issues (interview OHB). Also governmen-
tal organizations may share their resources. For instance, soldiers and equipment
from the Thai Army and Thai Navy supported the recovery of Western Digital (Wai
and Wongsurawat, 2013). The relevant roles of logistics service providers and
insurance companies were highlighted by some interviewed experts (interviews
AutoOEM1, Lampe and Schwartze). They seem to be valuable due to their neutral
supervising position within the network.
4.2 Promoting Factors
Three relational resources that promote collaborative recoveries are identified. A

collaborative culture among SC partners seems to be beneficial for collaborative
recoveries and is characterized by the relational resources ‘trust’ and ‘commit-
ment’. Literature in the field of SCC frequently emphasizes a close connection
between the two concepts (Kwon and Suh, 2005; Morgan and Hunt, 1994; Nyaga
et al., 2010; Ryu, So and Koo, 2009). Both will be discussed in the following
paragraphs. ‘Dependency’ seems to be the main motivation for companies to
collaborate during SC disruption recoveries and will be discussed afterwards.
4.2.1 Trust and Commitment
Trust is defined by Rousseau et al. (1998) as “the intention to accept vulnerability

based upon positive expectations of the intentions or behavior of another”. The
concept of trust is composed of two main dimensions (Bode et al., 2011; Cao
and Zhang, 2013; Kumar, Scheer and Steenkamp, 1995; Nyaga et al., 2010; Sheu,
Yen and Chae, 2006). According to Doney and Cannon (1997), credibility is the
“expectancy that the partner’s word or written statement can be relied on” and
benevolence is the “extent to which one partner is genuinely interested in the
other partner’s welfare and motivated to seek joint gain”.
Several studies found strong associations between SC partners’ trust and rela-
tionship success. For instance, Sheu et al. (2006) point out the criticality of trust
for supplier-retailer collaboration. According to Mohr and Spekman (1994), trust
is a primary characteristic of partnership success. Relationships with a high level
of trust are willing to take risks between partners (Kwon and Suh, 2005; Li et al.,
2015). According to McEvily et al. (2003), trust can be a mobilizer that motivates
456
companies to contribute and combine their resources in joint activities to achieve

their goals and to resolve problems.
Morgan and Hunt (1994) define commitment as “an exchange partner believing
that an ongoing relationship with another is so important as to warrant maximum
efforts at maintaining it; that is, the committed party believes the relationship is
worth working on to ensure that it endures indefinitely”. Commitment generally
encompass three dimensions, namely affective commitment (“the desire to con-
tinue a relationship because of positive affect toward the partner”), expectation
of continuity (“the firm’s perceptions of both its own and its partner’s intent to re-
main in the relationship, which, thereby, reflect the relationship’s stability”), and
willingness to invest in the relationship (“the intention to become more deeply
involved in the relationship through investments of capital and effort”) (Kumar et
al., 1995).
A large amount of studies stress the relevance of commitment in collaborative

relationships. The results of Mohr and Spekman’s study (1994) indicate that one
primary characteristics of partnership success is commitment. This is because
commitment creates a context in which parties can achieve their long-term individ-
ual and joint goals without being concerned about opportunistic behaviour. Ryu
et al. (2009) argue that commitment foster the integration of partnerships among
SC partners. The studies of Nyaga et al. (2010) show that trust and commitment
leads to improved satisfaction and performance in collaborative relationships
between SC partners.
The analysed cases show a high degree of trust and commitment between part-
ners that collaborated during SC disruption recoveries. For example, Aisin was
one of the most trusted of Toyota’s suppliers and main Japanese supplier of P-
valves, the critical part of which the SC was disrupted (Nishiguchi and Beaudet,
1998). During Western Digital’s recovery, the customers expressed their trust and
commitment in agreeing upon special provisions that deviate from normal con-
tractual agreements (Wai and Wongsurawat, 2013). Many interviewees confirmed
that high levels of trust and commitment lead to higher willingness to collaborate
during disruption recoveries (interviews Lampe and Schwartze, OHB). For exam-
ple, OHB treat their suppliers as partners that share risk and solve problems jointly
as they work towards a common goal (interview OHB). Based on the described
literature that emphasizes the key role of trust and commitment in collaborative
interorganizational relationships as well as the results of the case study analysis
and interviews, the following propositions are derived:
457
Proposition 1: A high level of trust between SC partners leads to a high level of

collaborative resource sharing during disruption recoveries.
Proposition 2: A high level of commitment between SC partners leads to a high
level of collaborative resource sharing during disruption recoveries.
4.2.2 Dependency
“Dependence exists when one party does not entirely control all of the condi-
tions necessary for achievement of an action or a desired outcome” (Handfield
and Bechtel, 2002; Monczka et al., 1998). In literature, several authors suggest a
positive effect of dependency on SCC. Sheu et al. (2006) show that interdepen-
dence between parties is one essential factor for interorganizational relationships
because it motivates to share key information and participate in joint efforts.
According to Monczka et al. (1998), interdependence as an attribute of supplier
alliances was found to be significantly related to partnership success. As stated
by Hudnurkar et al. (2014), the structure of dependency relationships is highly
relevant for successful collaborations as it has essential implications for joint
efforts, including problem solving.
The dependency on a partner is often connected to the relevance of a product
which is characterized by high levels of customization, supply scarcity, technical
complexity and/or novelty of technology (De Leeuw and Francoo, 2009; Whitney
et al., 2014). According to De Leeuw and Fransoo (2009), the more the poducts
are relevant, the more close SCC is expected. Nokia depends on Philips’ produc-
tion because only they can supply specific components needed for producing a
type of cell phone chip (Sheffi, 2005). Similarly, Renesas‘ microcontrollers are
custom-made which makes switching to another supplier difficult (Pollack, 2011).
However, the interviews and case studies show that also other forms of depen-
dency exist. The disruption of one supplier can lead to strategic problems for
other suppliers. In the case of Aisin, hundreds of suppliers, local electricity, gas
and transportation companies had to wait for Aisin to reopen the plant in order to
resume deliveries (Nishiguchi and Beaudet, 1998). In addition, the current SCM
business trends, especially single sourcing, leads to a high level of dependency.
The importance of dependency with regard to collaborative recovery was stressed
by all interviewed experts. The interviewees of Lampe and Schwartze said that
during disruption recoveries, they and their partners are “all in the same boat”.
Dependency seems to be the main trigger that motivates actors to get involved in
collaborative recoveries, which is reflected in the following proposition:
458
Proposition 3: A high level of dependency between SC partners leads to a high

level of collaborative resource sharing during disruption recoveries.
4.3 Supply Chain Resilience
Following the relational view line of argumentation, SCRES is achieved if a SC

recovers better than its competing SC, thus if a collaborative advantage is derived.
The success of a recovery is derived from its so-called time to recover (Simchi-Levi
et al., 2014). This is the time period between the occurrence of the disruption
and the complete restoring of the SC. Several authors stress the competitive
positioning role of resilience. For example, based on an extensive literature review,
Tukamuhabwa et al. (2015) conclude that companies that respond to a disruption
better than its competitors can improve their market position. Rice and Caniato
(2003) state that in case of disruptions, companies can compete on their resilience
capabilities.
In the analysed cases, companies’ time to recover was shorter compared to their
competitors. Unfortunately, the cases do not always provide sufficient infor-
mation about the time to recover of competing SC or companies. Instead, the
expected and the real time to recover were analysed. For instance, Western Digital
could successfully recover within two months. In comparison, four months after
the flood that caused the disruption, only about half of the 90 affected factories
in the area resumed production (Wai and Wongsurawat, 2013). The restart of
Renesas’ limited production was three months faster than initially anticipated
by it’s engineers (Olcott and Oliver, 2014; Renesas, 2011). With regard to Riken’s
recovery, multiple sources state that the support of its customers enabled the
company to restart production much quicker than it would have without their
help (Global Risk Miyamoto, 2007; Whitney et al., 2014).
As outlined by existing research and supported by the relational view theory,

there is a positive relation between the concepts SCC and SCRES (Bello and Bovell,
2012; Bovell, 2012; Brüning and Bendul, 2015; Wieland and Wallenburg, 2013). For
example, Christopher and Peck (2004) state that “building resilience to SC risks
requires a high degree of collaboration”. Based on the analyses and literature
results, the following proposition can be derived:
Proposition 4: A high level of collaborative resource sharing during disruption

recoveries leads to a high level of supply chain resilience.
459
5 Conclusion
Several companies applied the sharing of resources during disruption recoveries.

However, the method of collaborative recovery was not yet holistically explained
in literature. Collaborative recovery is based on the idea of the relational view
theory and which states that collaborative advantage can be derived from inter-
organizational resources (Dyer and Singh, 1998). It is suggested that by collab-
orating during SC disruption recoveries, a SC can be more resilient than their
competing SC. A framework to describe the promoting factors for collaborative
resource sharing and its effect on SCRES was developed. Based on a literature
review, case studies and expert interviews, three promoting factors were defined,
namely trust, commitment and dependency. The developed propositions indi-
cate that collaborative SC recovery has the potential to be a promising reactive
risk management method.
Based on the results presented in this paper, some managerial implications can be
derived. First, adaptability, mobility and availability of capacity are the identified
abilities which enable resources to be shared during SC disruption recoveries.
SC managers can prepare for fast recoveries by taking care that large amounts
of resources (HR and PR) have these abilities. For example, interorganizational
training can improve the adaptability and mobility of HR. Second, collaboration
took place between different SC actors. When planning for collaborative recovery
actions, SC managers may consider also collaborating with their indirect partners,
for instance second-tier suppliers. In addition, companies can take advantage
of logistics service providers or insurance companies and their neutral coordi-
nating roles in SC. Third, commitment and trust were identified as the relational
resources that promote collaborative recoveries. Thus, investing in strong, long-
term partnerships seems to pay off during SC disruption recoveries.
The methods employed in this research have several limitations. The amount of
cases and interviews conducted was limited. In addition, the cases were based
solely on secondary data. Further research could aim at triangulating the informa-
tion gathered. In general, case studies have limitations regarding the attainable
level of generalizability (Eisenhardt, 1989).
This paper is part of a larger research project. The next steps are the operational-
ization of the identified constructs, the execution of a large-scale survey, and the
statistical analysis of the data gathered to test the developed propositions.
460
5 Conclusion
Appendix: Overview of Interviews
Table 1: Overview of Interviews
Date/
Name of
Position Company Industry Type of
Interviewee
Interview
Manager 04.06.2015/
Anonymous AutoOEM 1 Automotive
Production Logistics face-to-face
30.07.2015/
Anonymous Quality Sales AutoOEM 2 Automotive
phone
Manager 06.08.2015/
Anonymous AutoOEM 3 Automotive
Logistics Planning phone
Senior Manager, 07.05.2015/

Anonymous AeroOEM 1 Aerospace
Procurement face-to-face
Director 18.02.2015/
Mr. Heinrich OHB Aerospace
Procurement face-to-face
Managing Funk 02.06.2015/

Mr. Löffler Insurance
Director RMCE phone
Mr. Viethen; Management Lampe and 18.02.2015/

Insurance
Mr. Bening Assistant Schwartze face-to-face
Managing 29.05.2015/
Mr. Rosenberg Controllit AG Consulting
Director phone
461
Appendix: Overview of Case Studie
Table 2: Overview of Case Studies

Company description Disruption description Recovery description SC
actors (resources shared
during SC disruption
recoveries)
Case: Western Digital October 2011: heavy flood-
Leading electronics ing in Bang Pa-In Industrial Es- – Customers (monetary
company; tate, Thailand; precision equip- resources (MR))
headquarter in the US; ment and supplies of mate- – Navy, army (human resources
several subsidiaries in rial were destroyed (Wai and (HR), production resources
Asia Wongsurawat, 2013); affected (PR))
customers worldwide, as the – Intra-organizational actors
price of external hard drives in- (employees of Western
creased at least by 10% (Fuller, Digitial’s national and
2011) international plants) (HR)
Case: Renesas March 2011: the Great East
Leading electronics Japan Earthquake severely – Customers (HR, PR)
and automotive damaged Renesas’, Naka man- – Competitors (MR)
company, world ufacturing plant; cleanrooms – Governmental agencies (MR)
market, share of 44% were destroyed; just-in- – Intra-organizational actors
(Olcott and Oliver, time, production resulted (HR)
2014); headquarter in in production shortages at
Japan major automotive,customers
(Pollack, 2011)
Case: Nissan One of March 2011: the Great East – Customers/competitors (HR)
Japan’s largest Japan Earthquake hit two as- – Suppliers (HR)
automotive sembly plants; loss of produc- – Competitors (HR)
manufacturers tion capacity of about 270,000 – Intra-organizational actors
cars; about 50 suppliers of Nis- (HR, PR)
san were directly affected by
the disaster as well (Greenway,
2014)
Continued on next page
462
References
Table 2: Overview of Case Studies (continued)

Company description Disruption description Recovery description SC
actors (resources shared
during SC disruption
recoveries)
Case: Riken Japan’s July 2007: Niigata Chuetsu

largest supplier of off-shore earthquake affected – Customers (automotive
piston rings in the one main plant and nine satel- manufacturers, equipment
automotive industry lite companies (Whitney et manufacturers) (HR, PR)
(50% market share) al., 2014); operations at main – Intra-organizational actors
(Whitney et al., 2014) plant were suspended for two (HR)
weeks (Global Risk Miyamoto,
2007; Whitney et al., 2014);
several automobile manufac-
turers (e.g. Toyota and Mit-
subishi) interrupted their oper-
ations (Global Risk Miyamoto,
2007)
Case: Philips March 2000: plant in Albu-
Worldwide operating querque, New Mexico, caught a – Customer (Nokia) (HR)
electronics company; fire because of lightning; clean- – Nokia’s suppliers (PR)
headquarter in the rooms were ruined and mil- – Intra-organizational actors
Netherlands lions of chips were contam- (PR)
inated by smoke and water
(Mukherjee, 2008)
Case: Aisin Supplies February 1997: fire in the fac-
mostly automotive tory in Aichi destroyed spe- – Customers (HR, PR)
parts, headquarter in cialized equipment; it’s cus- – Suppliers (HR, PR)
Aichi, Japan tomer Toyota announced a – Toyota’s suppliers (HR, PR)
shut-down of its production
lines on the following day
(Nishiguchi and Beaudet, 1998;
Whitney et al., 2014)
References
Barratt, M. (2004). “Understanding the meaning of collaboration in the supply chain”. In: Supply
Chain Management: An International Journal 1.9, pp. 30–42.
463
Bello, D. and L. Bovell (2012). “Collaboration analysis: joint resolution of problems in global supply
networks”. In: Information Knowledge Systems Management 1/2.11, pp. 77–99.
Bode, C., S. M. Wagner, K. J. Petersen, and L. M. Ellram (2011). “Understanding responses to
supply chain disruptions: Insights from information processing and resource dependence
perspectives”. In: Academy of Management Journal 4.54, pp. 833–856.
Bovell, L. J. (2012). “Joint resolution of supply chain risks: the role of risk characteristics and
problem solving approach”. In: Department of Marketing, Georgia State University.
Brüning, M. and J. Bendul (2015). “Self-healing supply networks: Collaborative Recovery from Sup-
ply Network Disruptions by Sharing Network Resources”. In: Proceedings of 24th International
Purchasing and Supply Education and Research Association.
Brüning, M., N. T. Hartono, and J. C. Bendul (2014). “Collaborative Recovery from Supply Network
Disruption: A Multiple Case Study”. In: Proceedings of the 12th International Logistics and Supply
Chain Congress.
Brüning, M., N. Hartono, and J. C. Bendul (2015). “Collaborative recovery from supply chain
disruptions: characteristics and enablers”. In: Research in Logistics and Production 3.5, pp. 225–
237.
Cao, M. and Q. Zhang (2013). Supply chain collaboration. Hamburg: Springer.
Christopher, M. and H. Peck (2004). “Building the resilient supply chain”. In: International Journal
of Logistics Management 2.15, pp. 1–14.
Craighead, C. W., J. Blackhurst, M. J. Rungtusanatham, and R. B. Handfield (2007). “The severity
of supply chain disruptions: design characteristics and mitigation capabilities”. In: Decision
Sciences 1.38, pp. 131–156.
Craighead, C. W., J. Blackhurst, M. J. Rungtusanatham, and R. B. Handfield (2013). “Supply chain
resilience: a risk intelligent approach to managing global supply chains”. In:
De Leeuw, S. and J. Fransoo (2009). “Drivers of close supply chain collaboration: one size fits all?”
In: International Journal of Operations and Production Management 7.29, pp. 720–739.
Doney, P. M. and J. P. Cannon (1997). “An examination of the nature of trust in buyer-seller rela-
tionships”. In: Journal of Marketing 2.61, pp. 35–51.
Dyer, J. H. and H. Singh (1998). “The relational view: cooperative strategy and sources of interor-
ganizational competitive advantage”. In: Academy of Management Review 4.23, pp. 660–679.
Eisenhardt, K. M. (1989). “Building theories from case study research”. In: Academy of Management
Review 4.14, pp. 532–550.
Flick, U. (2006). An introduction to qualitative research. London: Sage Publications.
Fuller, T. (2011). “Thailand Flooding Cripples Hard-Drive Suppliers”. In: NYTimes.com.
Ghadge, A., S. Dani, and R. Kalawsky (2012). “Supply chain risk management: present and future
scope”. In: The International Journal of Logistics Management 3.23, pp. 131–339.
Greenway, T. (2014). “How safe is your Supply Chain?” In: Delivered DHL.
Handfield, R. B. and C. Bechtel (2002). “The role of trust and relationship structure in improving
supply chain responsiveness”. In: Industrial marketing management 4.31, pp. 367–382.
Hendricks, K. and V. Singhal (2005). “An empirical analysis of the effect of supply chain disruptions
on long-run stock price performance and equity risk of the firm”. In: Production and Operations
Hudnurkar, M., S. Jakhar, and U. Rathod (2014). “Factors affecting collaboration in supply chain: a
literature review”. In: Procedia-Social and Behavioral Sciences 133, pp. 189–202.
464
References
Kubicek, H. (1976). Heuristische Bezugsrahmen und heuristisch angelegte Forschungsdesigns als

Elemente einer Konstruktionsstrategie empirischer Forschung. Stuttgart: J. B. Metzlersche Ver-
lagsbuchhandlung und Carl Ernst Poeschel Verlag GmbH.
Kumar, N., L. K. Scheer, and J.-B. E. M. Steenkamp (1995). “The effects of perceived interdepen-
dence on dealer attitudes”. In: Journal of Marketing Research 3.32, pp. 348–358.
Kwon, I. W. G. and T. Suh (2005). “Trust, commitment and relationships in supply chain manage-
ment: a path analysis”. In: Supply chain management: an international journal 1.10, pp. 26–
33.
Lavie, D. (2006). “The competitive advantage of interconnected firms: an extension of the resource-
based view”. In: Academy of Management Review 3.31, pp. 638–658.
Li, G., H. Fan, P. K. Lee, and T. C. E. Cheng (2015). “Joint supply chain risk management: an agency
and collaboration perspective”. In: International Journal of Production Economics 164, pp. 83–
94.
M., C., M. A. Vonderembse, Q. Zhang, and T. S. Ragu-Nathan (2010). “Supply chain collabora-
tion: conceptualisation and instrument development”. In: International Journal of Production
Research 22.48, pp. 6613–6635.
McEvily, B., V. Perrone, and A. Zaheer (2003). “Trust as an organizing principle”. In: Organization
Science 1.14, pp. 91–103.
Miyamoto, G. R. (2007). “2007 Niigata Chuetsu-Oki Japan Earthquake Reconnaissance Report”. In:
Mohr, J. and R. Spekman (1994). “Characteristics of partnership success: partnership attributes,
communication behavior, and conflict resolution techniques”. In: Strategic Management Jour-
nal 2.15, pp. 135–152.
Monczka, R. M., K. J. Petersen, R. B. Handfield, and G. L. Ragatz (1998). “Success factors in strategic
supplier alliances: the buying company perspective”. In: Decision Sciences 3.29, pp. 553–577.
Morgan, R. M. and S. D. Hunt (1994). “The commitment-trust theory of relationship marketing”. In:
The Journal of Marketing 3.58, pp. 20–38.
Mukherjee, A. (2008). “The fire that changed an industry: a case study on thriving in a networked
world”. In: Pearson education.
MunichRe (2016). “Overview natural catastrophes 2015”. In: Geo Risks Research.
Nishiguchi, T. and A. Beaudet (1998). “The Toyota group and the Aisin fire”. In: MIT Sloan Manage-
ment Review 1.40, pp. 49–59.
Nissan (2012). “Disaster response at Nissan”. In: Sustainability Report 2012, pp. 8–9.
Nyaga, G. N., J. M. Whipple, and D. F. Lynch (2010). “Examining supply chain relationships: do
buyer and supplier perspectives on collaborative relationships differ?” In: Journal of Operations
Olcott, G. and N. Oliver (2014). “Social capital, sense-making and recovery from disaster: Japanese
companies and the March 2011 earthquake”. In: California Management Review 2.56, pp. 5–22.
Pollack, A. (2011). “The assembly line is rolling again, tenuously, at Honda in Japan”. In: New York
Times.
Ponis, S. T. and E. Koronis (2012). “Supply chain resilience: definition of concept and its formative
elements”. In: Journal of Applied Business Research 5.28, pp. 921–930.
Ponomarov, S. (2012). “Antecedents and consequences of supply chain resilience: a dynamic
capabilities perspective”. In: Doctoral Dissertation, University of Tennessee.
Renesas (2011). “CSR and Environmental Report”. In: Renesas Electronics Corporation.
Reuters (2016). “VW’s key plant hit by parts shortage in supplier row”. In:
465
Rice, J. B. and F. Caniato (2003). “Building a secure and resilient supply network”. In: Supply Chain
Management Review 5.7, pp. 22–30.
Rousseau, D. M., S. B. Sitkin, R. S. Burt, and C. Camerer (1998). “Not so different after all: a cross-
discipline view of trust”. In: Academy of Management Review 3.23, pp. 393–404.
Ryu, I., S. So, and C. Koo (2009). “The role of partnership in supply chain performance”. In: Industrial
Management and Data Systems 4.109, pp. 496–514.
Sheffi, Y. (2005). The Resilient Enterprise: Overcoming Vulnerability for Competitive Advantage.
Cambridge, Massachusetts: MIT Press.
Sheu, C., H. Yen, and D. Chae (2006). “Determinants of supplier-retailer collaboration: evidence
from an international study”. In: International Journal of Operations and Production Manage-
ment 1.26, pp. 24–49.
Simatupang, T. M. and R. Sridharan (2002). “The collaborative supply chain”. In: The International
Journal of Logistics Management 1.13, pp. 15–30.
Simchi-Levi, D., W. Schmidt, and Y. Wei (2014). “From superstorms to factory fires”. In: Harvard
Business Review 1/2.92, pp. 97–101.
Stecke, K. E. and S. Kumar (2006). “Sources of supply chain disruptions, factors that breed vulner-
ability, and mitigating strategies”. In: Journal of Marketing Channels 3.16, pp. 193–226.
SwissRe (2015). “Natural catastrophes and man-made disasters in 2014: convective and winter
storms generate most losses”. In:
Thun, J. H. and D. Hoenig (2011). “An empirical analysis of supply chain risk management in
the German automotive industry”. In: International Journal of Production Economics 1.131,
pp. 242–249.
Tukamuhabwa, B. R., M. Stevenson, J. Busby, and M. Zorzini (2015). “Supply chain resilience:
definition, review and theoretical foundations for further study”. In: International Journal of
Production Research, pp. 1–32.
Voss, C., N. Tsikriktsis, and M. Frohlich (2002). “Case research in operations management”. In:
International Journal of Operations and Production Management 2.22, pp. 195–219.
Wai, L. C. and W. Wongsurawat (2013). “Crisis management: Western Digital’s 46-day recovery
from the 2011 flood disaster in Thailand”. In: Strategy and Leadership 2.41, pp. 34–38.
Whitney, D. E., L. J., and H. D. A. (2014). “The benefits and constraints of temporary sourcing
diversification in supply chain disruption and recovery”. In: Journal of Purchasing and Supply
Wieland, A. and C. M. Wallenburg (2013). “The influence of relational competencies on supply
chain resilience: a relational view”. In: International Journal of Physical Distribution and Logistics
466
Automotive Supply-Chain Requirements
for a Time-Critical Knowledge Management
Ann-Carina Tietze1 , Jan Cirullies2 , Boris Otto2
1 – Volkswagen Commercial Vehicles; TU Dortmund
2 – Fraunhofer Institut für Software- und Systemtechnik ISST
Transforming increasingly growing data volumes into knowledge and improving

its usage requires knowledge management models (KMM). KMM structures the
workflow for decision taking based on knowledge. Industry-suitable requirements
for a KMM, in particular for automotive supply chains (SC) and supply-critical
bottlenecks, are not raised, especially concerning the crucial parameter of time-
criticality. As none of the investigated models suits time-related specifications,
requirements for time-critical knowledge management (KM) are derived from
former case studies (CS) in the manufacturing automotive industry by literature
research. These requirements will be used to evaluate existing KMM proposed in lit-
erature. Requirements for a KMM, which supports the manufacturing automotive
industry (AI) in time-critical cases, are collected from practice by means of group
discussions, generalised, abstracted and verified such as real-time capability,
availability and accessibility, incentives for knowledge-sharing or intuitive han-
dling. In particular, it addresses the application case of a supply-critical bottleneck
in the inbound logistics. This results in rethinking of knowledge as a fundamental,
time-critical resource for the reduction of supply risks. Currently, there are neither
KMMs that involve time-criticality supporting industry to deal with increasing
data and knowledge volumes nor precise requirements for time-critical KM in
case of a supply-bottleneck in the AI. The importance of time-critical knowledge in
contrast to mere data is shown. Finally, time-criticality is highlighted by showing
its value to minimise production-breakdown-risks. The aim is to raise awareness
about the need for changes in existing processes in the AI and to define the scope
of scientific research needs.
Keywords: time-critical knowledge management; bottleneck management; automo-

tive industry requirements; case study research
467
Automotive Supply-Chain Requirements for Knowledge Management
1 Introduction
Knowledge and fundamental data have, as a key success factor to the creation
of competitive advantages, placed their importance as significant and precious
resources in the management ideas of the AI. Fashionable concepts such as Big
Data and Industry4.0 are an integral part of the company environment. Data is
collected and stored which makes the data volume grow exponentially. This poses
new challenges for the AI, as the collected data has to be useable and applied to
generate and to disseminate knowledge. It is becoming increasingly clear that the
collection and storage of large amounts of data only helps if it is not understood
as a self-purpose, but the use of this data is paramount.
What is the role of time? Time as a success factor for the competitive advantages of
producing companies in the AI is fundamental: the production times of individual
products, as well as the delivery times of suppliers and to customers must always
be reduced. Innovations must be brought rapidly to market; the products are
subject to constant change. The consumer adapts to these shortened cycles
and expects a constant stream of innovations. True to the motto of Rudolf von
Bennigsen-Foerderer (1923-89) ”Standstill is a step back” (Buchenau, 2016, p.191),
the industries do their utmost to promote change. The bulk of the information
used in the AI is not static, but rather time-critical, flexible and changeable. A KMM,
which takes knowledge as static, fixed quantities, may not meet the requirements
of the industry and, rather than help, delay, unsettle and thus create risks for the
supply of the production.
The use of collected data results in the necessity of integrating time-criticality into
KMMs. In the course of this, the requirements for such a time-critical KMM have
to be determined and presented in the manufacturing AI. These requirements
thus form the basis of a KMM which is adapted to current needs of the AI for the
optimal use of collected data in the form of knowledge.
The target of this paper is to define precise requirements for a KMM from the AI
that are not yet fulfilled by existing KMM. Their suitability for the industrial envi-
ronment of the AI, especially in time-critical fields, is the focus of the investigation.
The central aspect is the application of a supply-critical bottleneck. The structure
of the work provides for an examination of the theoretical and practical problem
background in chapter 2, followed by a description of the methodological proce-
dure in chapter 3. Empirical analyses in the form of the inclusion of investigative
CSs, whose findings are cumulatively evaluated, are used to meet requirements
468
2 Theoretical Perspectives on Knowledge Management
with the objective of providing comprehensive knowledge of time-critical informa-

tion. Chapter 4 deals with the implementation and the findings, followed by the
evaluation (chapter 5). The special requirements of the subjects are abstracted
and generalised. The work concludes with a summary in chapter 6.
2.1 Difference between Knowledge and Data in Literature
Despite the frequently erroneously synonymous use of the terms ”data”, ”infor-
mation” and ”knowledge” in everyday usage, there are serious differences in the
actual meaning, which makes a delimitation of the terms necessary.
Signs form the basis. Data includes arbitrary strings and value is assigned to
the character by syntax. Through collusion on specific characters and strings,
the user is able to recognise the contents of data. As soon as the data is linked
together and a causal link is assigned to them in form of semantics, they become
information (Bodendorf, 2006, pp.1-2; North, 2011, pp.36-37). Use of information
in this context already means the generation of knowledge which is understood
to mean the networking of information with the objective of using and apply-
ing it purposefully. In order to connect and interpret information in the correct,
meaningful and effective manner, it is often necessary to have prior knowledge
that the user either already possesses or has to acquire (Bodendorf, 2006, pp.1-2;
North, 2011, pp.36-37; Nikodemus, 2017, pp.83-86).This leads to the conclusion
that knowledge includes not only the networking of information but also experi-
ences and knowledge, abilities and skills, values, know-how and qualifications
(Nikodemus, 2017, pp.85-86).
This distinction also is highly relevant today as it corresponds to the Industry
4.0 Maturity Index, a recent study from catech, which describes the digital trans-
formation of companies. The acatech industry 4.0 Maturity Index serves as a
guide for companies interested in advancing their degree of digitalisation and
is build up as a six stepped model: ”computerisation” and ”connectivity”, which
form together the group ”digitalisation”, ”visibility”, ”transparancy”, ”predictive
power” and ”adaptability”, which form the category ”Industry 4.0” (Schuh et al.,
2017, pp.15-18). These model levels are in line with the knowledge pyramid of
Bodendorf (2006) as they represent a detailing and extension of it. ”Signs” have to
469
be reconciled with the ”digitalisation”. This is followed by the ”data” level, which
is equivalent to ”visibility”: ”What happens?” A causal link follows in the next step,
the ”transparency”. A forecast for the future, which is possible only by means of
knowledge, completes the model for the Industry4.0 Maturity Index.
2.2 Knowledge Management in Automotive Industrial Environment
In the AI, data is collected and used in large quantities. However, this develop-
ment is often still at an early stage, so that not all potentials are fully exploited:
knowledge is collected, but not processed; processed, but not linked; linked but
not communicated. In a company, knowledge is often only present within certain
fields and is not shared by either the system or the employees of different depart-
ments. The transition from implicit to explicit knowledge is often interrupted (cf.
section 2.3.1). This makes the integration of a new KM in companies necessary.
2.2.1 Special Case: Time-Criticality
KM as a necessary concept for a modern company organisation is not a new

approach (Born, 2009; Lakshman and Parente, 2008; Jane Zhao and Anand, 2009).
However, one important factor is missing: the time. It plays an important role
in the manufacturing industry. True to the principle of ”time is money”, time is
a great influence on competitiveness, securing the market position as well as
the size of the output and productivity. A KM concept for the industry has to
consider and integrate this factor. The time-criticality includes on the one hand
time-critical processes and a time-critical information provision, on the other
hand also time-critical information.
2.2.2 Bottleneck-Management in the Automotive Industry
This chapter is an excerpt from practice in the form of an experience report.

In the AI a distinction is made between bottleneck-management, which aims to
prevent bottlenecks on the production proactively, and the bottleneck-control,
which takes place after the occurrence of a supply-critical bottleneck and deals
with the regulation and distribution over affected plants (reactive).
470
Bottleneck-management is a particularly illustrative application for a time-critical

KM. Within the scope of capacity management, the supply assurance of the pro-
duction is checked at regular intervals. For this purpose, the production capacity
provided by the supplier is compared with the parts requirements determined by
the Original Equipment Manufacturer (OEM). Since these needs and capacities
vary according to the order, to the machine utilisation, shift models and holidays,
a system determines ”secured” and ”unsecured” weeks by comparison. Identified
”unsecured” weeks will be investigated in the daily hedging process, so that risks
can be ruled out at an early stage.
However, the supplier often suffers short-term supply disruptions, for example due
to political unrest, natural catastrophes or financial disasters. In these cases, rapid
action is crucial in order not to endanger production. Manual data is collected
along the SC. Contact persons are identified and get informed. An internal team
consisting of procurement, logistics and partial quality assurance is compiled and
contacts the supplier in regular telephone conferences. For these conferences,
data is collected in advance. This proves to be difficult because of the person-
dependent limitation with regard to data’s availability: information on logistics,
procurement and quality assurance is only disclosed to the other involved parties
in order to deal with the time-critical situation in the best possible way. The
manual effort to receive the message with regard to a supplier-critical situation
at the supplier up to the time the SC and the contact persons are known and all
relevant data for the management of a supply-bottleneck, is enormous and can
take an average of up to a week. In a time-critical supply-bottleneck, one week
could be the cause for a production breakdown and profit loss. Therefore a time-
critical KMM is needed, which helps to deal with acute knowledge in time-critical
situations which are of paramount importance.
2.3 Knowledge Management Models
A literature review for KMM by using the keyword ”knowledge management”

yields numerous hits. Adding the keyword ”time-critical” has no useful result.
For this reason, a further literature review for time-critical KM is not effective.
Nevertheless, this chapter will provide an overview of the core idea of KM but
will not give a summary of all KMM. The models generally deal with the process
of generating and disseminating knowledge. Some of these models address
theoretical approaches to deal with knowledge, for example knowledge-based
theory or competence-based theory (Martensson, 2000, p.205). Others do not
471
integrate industrial but individual requirements (Reinmann-Rothmeier and Eppler,

2008). The third category identifies requirements of industry and abstracts these
in the form of business driven KMM.
From the third group, three models will be presented, which are covering a
wide spectrum and are the basis for many other models, just like North (2011),
Gronau (2009) or Nikodemus (2017). The SECI-model from Nonaka and Takeuchi
(1997) is one of the fundamental models for the understanding of KM today.
Probst, Raub and Romhardt (2012) take up further industry requirements and
expand the SECI-model. Their approach to anchoring KM in the industry becomes
apparent. The ”Magical Triangle of KM” is presented as the third model. This
model takes on other requirements of industry and attempts to abstract and
generalise them (Gehle, 2006; Rücker, 2002).
2.3.1 SECI-Model by Nonaka and Takeuchi
Nonaka and Takeuchi distinguish between explicit and implicit knowledge. Ex-
plicit knowledge means written knowledge, which can be easily transmitted. In
contrast, implicit knowledge is only available to the knowledge carrier, often
unconsciously. It is hard or impossible to be transferred (Nonaka and Takeuchi,
1997, p.75).
According to Nonaka and Takeuchi (1997, p.75) these two types of knowledge can
be transformed in the form of a continuous knowledge spiral, thus generating
knowledge. Four successive phases are performed regularly:
— Transformation from implicit to implicit knowledge is called Socialisation.

It means the acquisition of individual, implicit knowledge.
— Transformation from implicit to explicit knowledge is called Externalisa-
tion. It describes the documentation of implicit knowledge to become
explicit.
— Transformation from explicit to explicit knowledge is called Combination
and means the arrangement of the new explicit knowledge with already
existing explicit knowledge.
— Transformation from explicit to implicit knowledge is called Internalisa-
tion and refers about the assimilation of new knowledge and becoming
implicit knowledge.
472
This means that individual knowledge can be transferred to an ever-increasing

organisational level within an organisation, making it accessible. Nonaka and
Takeuchi (1997, p.75) have thus created a general model that helps for example
in a company, to develop and multiply knowledge, and thereby to manage the
generation of knowledge.
This model is therefore a basis for the understanding of the generation of knowl-
edge, which is also applied in companies, but the time factor is not integrated.
Time-critical knowledge is treated according to this model as well as time-uncritical.
The model must therefore be adapted to the industry.
2.3.2 Knowledge Circulation by Probst, Raub and Romhardt
Probst, Raub and Romhardt (2012) proceed inductively from the requirements of
the industry and abstract their results in a model. There are six problem categories
that influence each other and, according to Probst, Raub and Romhardt (2012,
pp.30-33), are the core processes of KM. The process of knowledge identifica-
tion is intended to create transparency about internally and externally available
knowledge, the process of knowledge acquisition addresses the question of the
acquisition of new knowledge outside the organisation. Complementary to this
is the process of knowledge development, which focuses on building new inter-
nally and externally knowledge. The knowledge sharing involves the transfer
of knowledge to the right place, while the knowledge exploitation ensures the
productive use of organisational knowledge. The knowledge preservation takes
care of the protection against loss of knowledge. In addition, they introduce two
further categories that are intended to anchor the importance of knowledge in
the company: the knowledge objectives which define the direction of learning, as
well as the knowledge validation for the measurement of learning success (Probst,
Raub and Romhardt, 2012, pp.30-33).
New to this model is the focus on the requirements of the industry, which are
necessary to implement the model in practice. However, the time factor is also
not integrated in this model.
2.3.3 ”Magical Triangle of Knowledge Management”
The magic triangle of KM is widely discussed in the literature (Gehle, 2006, p.11;
Rücker, 2002, p.32). All however address the approach of adapting KM to the
473
needs of the industry by integrating the fields of ”employee/culture/personnel

management”, ”organisation/processes” and ”information and communication
technology” (Gehle, 2006, p.11; Rücker, 2002, p.32). This is intended to create
an interdisciplinarity of KM, which is necessary for the establishment in practice
(Probst, Raub and Romhardt, 2012, p.259).
In detail, this means that transparent business processes and decision-making

structures, as well as training and incentives to increase knowledge are neces-
sary. Appropriate technical preconditions must also be created (Rücker, 2002,
p.32). Nevertheless, the time-criticality of knowledge is not mentioned in these
models.
2.4 Need for Action
Data as a resource of the present time implies the construction of large data
volumes and also requires KM, which helps to make the collected data usable.
There are numerous models that deal with the generation of knowledge. Their
suitability for the AI is limited, because they do not take the factor of time-criticality
of knowledge and information into account.
The volume-adequate supply of the production with parts in a supply-bottleneck

depends decisively on the handling of time-critical knowledge. The result is
a time-critical KMM, which is adapted to industrial needs. As a basis for this,
requirements for a time-critical KM have to be collected, compared and assessed
from industry and science. In order to produce a representative image of reality,
the requirements are gathered from several group discussions conducted in loops,
which are based on the application of bottleneck-management. These practical
requirements represent the starting point for further deepening with scientific
research focus.
474
3 Case Studies
3 Methodological Approaches for the Development of

Time-Critical Knowledge Management: Case Studies
3.1 Research Method Selection and Types of Designs
CSs from literature are compared with a newly recorded CS and requirements for
a time-critical KM in the manufacturing AI are derived. The inclusion of CSs shall
ensure a high degree of practical relevance and timeliness (Gehle, 2006, p.16). It
is a qualitative research approach that is applied when complex, little explored
fields of interest are considered with contextual relevance. It has the advantage
that hypotheses can be derived from the practice which are verified and validated
with the help of literature. This leads to new perspectives and results in a new
insight and the development of new knowledge (Borchardt and Göthlich, 2009,
p.46).
Following YIN (2014, pp.49-50), an embedded single-case study is presented.
The bottleneck-management is used as a relevant context in the time-critical
supply-bottleneck, and the different fields involved in the internal and external
management are used as analysis objects.
3.2 Data Collection Method
CSs as a qualitative research method allow a variety of data collection methods

(Riesenhuber, 2009, pp.6-7). According to Borchardt and Göthlich (2009, p.38),
these can be divided into three categories: observation, questioning and content
analysis, which differ in their approach.
The observation shows a rather passive, non-participating basic idea. Information
is collected, not determined (Borchardt and Göthlich, 2009, p.38).
The survey is of a reactive nature as participants react to the directly respondent
(Weber, 2009, p.147; Riesenhuber, 2009, p.12). Differentiation will be made in
individual interviews, which may be personal, by telephone or in writing, and
group interviews. Depending on the application, the construction is structured,
partially structured or unstructured. The choice of the method depends on the
application, the research question and the result to be achieved. Complex, poorly
researched fields often require semi-structured interviews, which only slightly
limit the possible answers. In contrast to this, a structuring is advantageous in
475
a specific questioning, which leads to focused responses (Lamnek, 2002, p.173;

Meuser and Nagel, 1991, pp.448-449).
A special form of the group interview is the group discussion. It is particularly

concerned with the interaction of the group members (Fay, Garrod and Carletta,
2000, p.481; Vogel, 2014, p.581). The framework of the discussion is provided
by a guideline, which supports to focus on the research topic. It includes the
framework theme as well as a few individual aspects (Dammer and Szymkowiak,
1998, pp.94-95; Vogel, 2014, p.583). This method has a rather unstructured char-
acter, but it also allows a great gain in knowledge. The participants inspire each
other with opinions, insights and new ideas. The interplay in a group on a partic-
ular research subject enables knowledge that is not possible through individual
queries of prefabricated questionnaires (Müller, 2008, p.5; Vogel, 2014, p.582).
This enables existing hidden problems and potentials to be uncovered, which the
individual is not aware of or with which he has arranged himself. An advantage
of this method is, on the one hand, the breaking open of fixed views and ideas;
on the other hand the collective orientation can be worked out. The influencing
of the participants can be disadvantageous, just as the hindrance of individual
expressions of opinion due to adaptation mechanisms (Frey, 2007, p.3; Vogel,
2014, p.582).
The content analysis as the third category according to Borchardt and Göthlich
is based on observations and interviews, and supplements them with informa-
tion based on documents and data sets. It is thus a summary and explanation
(Borchardt and Göthlich, 2009, p.38).
3.3 Further Requirements of the Method
The selection of the participants is crucial for the quality of the results. Rack
and Christophersen (2009, p.28) point out that there may be differences in the
survey results, depending on whether the participants participate voluntarily or
involuntarily in the survey.
In group interviews or group discussions, the group’s composition is also of in-

terest. A distinction exists between real or artificial groups and homogeneous
or heterogeneous groups. The former are more realistic in comparison to the
latter and have a homogenous background, but are often also restricted in the
openness of the discussions. In group discussions, the optimal group strength
476
4 Case Study from Industrial Practice and Requirements in Literature
is between six and ten participants (Liebig and Nentwig-Gesemann, 2009, p.105;
Vogel, 2014, p.581-584).
The optimal quantity of discussions depends on the theoretical saturation. As
soon as a new survey or discussion does not bring any added value to the find-
ings, this saturation is achieved and thus the optimal quantity of interviews or
discussions (Vogel, 2014, p.584).
4 Case Study from Industrial Practice and Requirements

in Literature
The method of the CS is chosen to create an image of reality and to deduce induc-
tively from the requirements of practice to the theory behind it. Group discussions
as survey methods should promote the creativity of respondents. In order to break
down established behavioral patterns and perspectives in the first step, which can
only be identified by the interaction in the group, this method is very suitable.
4.1 Object of the Survey
The object of the survey is a multinational large group of the AI, headquartered in
Germany. A practical CS is taken as a use case. The application of a supply-critical
bottleneck by the manufacturing AI provides the framework conditions for the
extent of the CSs taken.
All priority fields involved in a time-critical lack of material are surveyed as shown
in figure 2. Therefore, their requirements for a time-critical KMM are recorded
and then compared with each other. The requirements of the participants are
used to derive abstracted, generally valid requirements for time-critical KM for
volume-critical supply-bottlenecks in the AI.
4.2 Process, Participants and Timeframe
The method of group discussion is applied on the basis of suitability for the
present application. The target is to uncover hidden potentials in bottleneck-
management, which are unaware of the individual. Further advantages in the
477
benefit of group dynamics lie in the integration of all involved parties to increase
the motivation as well as the promotion of creativity. A guide is limited to the
theme of bottleneck-management and defines the calls to discuss the process,
to highlight obstacles, to uncover information gaps, and to present personal
requirements to an ”optimal” process.
The participants are selected on the basis of the need for motivated and com-
mitted participants on a voluntary basis. All those involved in the bottleneck-
management application are invited to participate in the group discussion.
Due to the assignment to specific subject areas, they are homogeneous real
groups with a range of between six and ten participants. Group-wide employees
in the fields of procurement and logistics are surveyed in three large companies
in the AI (OEM) as well as in an international medium-sized supplier. In total, 10
discussions are held with a total of 35 participants and, in order to ensure the
anonymity of the participants, the results of the evaluation are not broken down
into the fields involved.
4.3 Procedure of the Group Discussions
The discussions lasted a period of five months in the first half of 2017. The initial
question, which can be seen in figure 1, is set independently for each group. To
answer this question, group discussions take place. It is only in this discussion that
disclosure of existing potentials is possible since the group participants are partly
challenging each other by their different perceptions, partly confirming each other
by their same perceptions. The contents of the discussion topics as well as the
results of the discussions are documented in short protocols. Subsequently, the
results are clustered and submitted to the group participants again. The group
itself analyses and evaluates the results. It prioritises certain topics and places
the importance of others in the background, as shown in figure 1. The clusters
of results are prioritised in high potential, coloured dark grey, and low potential
(light grey).
4.4 Findings from the Bottleneck-Survey
The ”lived” bottleneck-process is shown in figure. In most cases, the supplier

provides the first information regarding a supply risk. Depending on the risk
478
4 Case Study from Industrial Practice and Requirements in Literature
Figure 1: Clusters of potentials for the bottleneck-process in the AI
location within the SC, this information can take several days to reach the right
contact person at the OEM, the responsible buyer in the normal case.
The latter then provides information on logistic and procurement, as shown
in figure 2, as well as on the development of the individual brands and plants
concerned. Together, they set up a team, which exchanges itself by telephone with
the responsible supplier. The necessary information is determined on the basis
of the system’s access to the individual areas. Further information is collected
manually.
A range-wide monitor correcting acute requirements with current stock allows
the available parts to be allocated to the plants. If the SC is filled again and the
delivery call can be operated again, the control process is passed on.
The obstacles in the current bottleneck-process are identified. The necessary
manual recording of information is a great deal of time and effort for the editor.
In addition, the distribution of information via e-mail by missing contact persons
in the mailing list and misdirected or non-sent mail represents a major obstacle
in the processing of bottlenecks. Another barrier is missing and limited access
rights to information and systems. In addition, trust in the systems is limited due
to faulty or obsolete data. It would be helpful to know the SC. However, these
data are not available and cannot be determined due to legal restrictions in the
479
Figure 2: Bottleneck-process including the integrated departments
individual case. As a result, contacts are also lacking along the SC, which in the
event of a bottleneck might have the potential to ease the situation.
480
5 Abstraction and Evaluation of Requirements
5 Abstraction and Evaluation of Requirements for

Time-Critical Knowledge Management
5.1 Derived, General Requirements
The described obstacles lead to requirements for bottleneck-management which

can be divided into three priority categories: Requirements for data, for handlings
and for time, as shown in figure 1. Requirements for the category ”time” are not
broken down in detail since this paper generally deals with the requirements for
time-critical KM. As a result, only two priorities remain.
5.1.1 Requirements for Data
The results of the group discussions can be summarised as follows with regard to
the requirements for a time-critical KM:
Actuality of data
Time-critical knowledge requires real-time, up-to-date information. An outdated

state of data can lead to incorrect decisions in specific situations, since a time-
critical process requires quick decisions.
Probst, Raub and Romhardt (2012, p.217) define the actuality of data as a basis
for data quality. They justify this approach, for example, with high costs for in-
vestment decisions based on outdated information (Probst, Raub and Romhardt,
2012, p.217). North (2011, p.52) also refers to the actuality of information bases.
Weber and Berendt (2017, p.28) even note that a system cannot be used without
a suitable data update.
Timely availability of data
In time-critical situations requiring rapid action, requested data must be immedi-

ately available. Any additional time required to find the information means that
there are temporal and thus monetary losses.
481
North (2011, p.52) points out that time and place are the availability of knowledge-
determining factors. Time differences due to global companies may not affect
the availability of information.
Local availability and accessibility of data
Information must be accessible to everyone involved. In addition, it is necessary

that all involved parties, regardless of their location, are informed.
According to North (2011, p.52), a local availability of data is to be taken into
account especially in globally operating companies.
Quality of data
In the time-critical case, the quality of data is important, in particular, the com-
pleteness, authenticity and resilience of the information. In the case of poor
information quality, optimal decisions cannot be made. Trust in the use of avail-
able data would decrease.
Probst, Raub and Romhardt (2012, p.205) declare the quality of information as
crucial for databases, since databases with irrelevant data lead to mistrust and
consequently to a reduced or lack of use.
5.1.2 Requirements for Handling
Trust and silo mentality
The exchange of knowledge depends to a great extent on the willingness to coop-

erate between the parties involved. A ”silo mentality” creates boundaries. Trust
among each other and a breakup of silo thinking is therefore indispensable.
Probst, Raub and Romhardt (2012, p.127) mention the problem of the formation
of inefficient knowledge bases in organisations with high communication barriers.
The keys to collective knowledge creation are, according to Probst, Raub and
Romhardt (2012, p.127), interaction, communication, transparency and integra-
tion.
482
Means of communication
A network of all participants creates the conditions for a targeted exchange of

information. The means of communication must be accessible to everyone in-
volved. They must also provide a framework appropriate to the application and
generation.
According to Probst, Raub and Romhardt (2012, p.127) the success of organi-
sations the individual knowledge of the individual can be less helpful than the
knowledge between the individual, the dependencies and relationships. These
relationships can only be established and maintained through interaction and
communication. North (2011, p.52) puts the factor of communication in the gen-
eration of knowledge in the foreground by focusing on a personal learning of the
participants.
Motivation
Incentives are needed that motivate the participants to share their knowledge
and experience. This can take the form of a change in corporate culture, in which
knowledge is rewarded and not the restraint of knowledge with the target of
becoming irreplaceable. Another possibility consists in the approach of the Gami-
fication of the work (Heilbrunn and Sammet, 2017, p.83).
An increased commitment by all parties involved in the implementation of com-

mon goals can be achieved through motivation (Probst, Raub and Romhardt,
2012, p.60). In the case of a lack of motivation, a transfer of knowledge is made
more difficult (Probst, Raub and Romhardt, 2012, p.170).
5.2 Evaluation of Generalised Requirements
Even if the identified requirements are justified by the expert’s knowledge of the
group discussions, it is necessary to analyse the hurdles that the requirements
have currently failed in order to allow a reduction.
483
The actuality of data
Data must be up-to-date and available in real-time. For this purpose, a networking
of all source systems according to a suitable structure is necessary. Constant
updating of certain records is required.
Timely availability of data
The information supply must be supported on the IT side in order to avoid tempo-
ral losses due to analogue information procurement.
Local availability and accessibility of data
Unrestricted access of all parties to necessary, existing data is essential. For this, a
suitable tool must be available to each participant. Unrestricted communication
is necessary. Access permissions for specific data headers must be extended.
Quality of data
In order to ensure high data quality, a validation of the information takes place in
advance. This is possible, for example, by evaluating the reliability of the source
system.
Trust and silo mentality
There has to be no holding on to old structures but openness for new solutions.
An integration of all stakeholders is indispensable.
Means of communication
The means of communication must be accessible to everyone involved. They

must also provide a framework appropriate to the application and generation.
484
Motivation
One possibility is the approach of gamification at work (Heilbrunn and Sammet,

2017, pp.85-87). It is an incentive from within. The work and especially the use
of tools should be fun and therefore promote motivation. Another approach is
a change in corporate culture. This is an incentive from outside, by rewarding
knowledge and not by restraining knowledge with the aim of becoming irreplace-
able.
5.3 Evaluation of Methods
The models of KM discussed in chapter 2.3 do not cover the requirements of indus-
trial practice in the AI in a time-critical situation. In order to encounter this, it is
necessary to deduce, outgoing from the requirements of the practice, knowledge
and to transfer it into the theory. The CS is the appropriate methodology to illus-
trate the complexity of the practice and make it tangible. Through the inclusion of
expert knowledge by group discussions, an enormous return of know-how from
practice to theory is possible.
Nevertheless, it should be noted that the extent of the CS can be expanded by
extending a questionnaire regarding the number of participants as well as the
number and structure of questions.
A full evaluation of the added value of the methodology and the results is only
possible if the identified requirements are transferred to a KMM and this model is
applied in practice.
5.4 Limitations
This paper describes the basis for an adapted KMM for the structuring of knowl-
edge. These are the items that provide a guide. The requirements are particularly
applicable in the AI and cannot be fully applicable to other industries and com-
panies. The requirements determined by surveys of a wide range of different
automotive manufacturers are intended to become a general invalidity which,
however, has gaps for limiting the selection.
The described case of handling time-critical knowledge in the application of the
supply-critical bottleneck-management of the AI places special emphasis on the
485
requirements in the time-critical case. These requirements may differ in the

absence of time-criticality.
In addition, it will not be sufficient to rely solely on a KMM in the time-critical
supply-risk. Such a model may support, but cannot replace the experience of the
employees.
6 Conclusions
The present work shows that there are numerous, partly unused sources of knowl-
edge in the AI that must be prepared and used. This is conceivable in the form of
a KMM. However, literature currently lacks KMM that integrate the requirements
of the time-criticality of data and processes. They are not optimally adapted to
the needs of the industry.
The subdivision of the requirements (cf. figure 3) consists of two classes and
depicts two fields of action: On the one hand, an optimisation of data structures is
necessary in order to provide an up-to-date, complete and high-quality providing
of the user’s information to the users. This includes the adapted networking of
data sources and the restructuring of the current network, as well as making the
information accessible to all participants in the form of timely, adapted tools. On
the other hand, a change in the corporate culture away from silo mentality and to
the idea of a team is also a matter of cross-company activity.
However, there are further research needs. The requirements are the basis for
a time-critical KMM. It is therefore necessary to analyse further models and to
reconcile them with the identified needs of the industry. Hence, an adaptation
of existing models is required. It should be noted that an adapted model is not
sufficient for applicability in industry alone. Structuring the generated knowledge
is necessary, for example in the form of an information object model.
In addition, a comprehensive analysis of the motivation factors of the users is
essential, which is fundamental for the implementation and integration of a time-
critical KM.
486
6 Conclusions
Figure 3: Conclusion of the determined requirements of the AI for time-critical

KMM
487
References
Albers, S., D. Klapper, U. Konradt, A. Walter, and J. Wolf, eds. (2009). Methodik der empirischen
Forschung. 3., überarbeitete und erweiterte Auflage. Wiesbaden and s.l.: Gabler Verlag.
Baur, N. and J. Blasius, eds. (2014). Handbuch Methoden der empirischen Sozialforschung. Wies-
baden: Springer VS.
Bodendorf, F. (2006). Daten- und Wissensmanagement. 2., aktualisierte und erweiterte Auflage.
Springer-Lehrbuch. Berlin, Heidelberg: Springer-Verlag Berlin Heidelberg.
Borchardt, A. and S. Göthlich (2009). “Erkenntnisgewinnung durch Fallstudien”. In: Methodik der
empirischen Forschung. Ed. by S. Albers, D. Klapper, U. Konradt, A. Walter, and J. Wolf. Wiesbaden
and s.l.: Gabler Verlag, pp. 33–48.
Born, V. (2009). Problemorientiertes Wissensmanagement in der Automobilindustrie: Ein interdiszi-
plinärer Ansatz. Wiesbaden: Gabler Verlag / GWV Fachverlage GmbH Wiesbaden.
Buchenau, P., ed. (2016). Chefsache Gesundheit II: Der Führungsratgeber fürs 21. Jahrhundert. 1.
Aufl. 2016. Wiesbaden: Springer Gabler.
Dammer, I. and F. Szymkowiak (1998). Die Gruppendiskussion in der Marktforschung: Grundlagen -
Moderation - Auswertung Ein Praxisleitfaden. Wiesbaden: VS Verlag für Sozialwissenschaften.
Fay, N., S. Garrod, and J. Carletta (2000). “Group discussion as interactive dialogue or as serial
monologue: the influence of group size”. In: Psychological science 11.6, pp. 481–486.
Frey, F. (2007). Die Gruppendiskussion als Instrument der Datenerhebung in der Marktforschung.
Grin Verlag.
Garz, D. and K. Kraimer, eds. (1991). Qualitativ-empirische Sozialforschung: Konzepte, Methoden,
Analysen. Wiesbaden: VS Verlag für Sozialwissenschaften.
Gehle, M. (2006). Internationales Wissensmanagement: Zur Steigerung der Flexibilität und Schlagkraft.
Deutscher Universitts Ver.
Gronau, N. (2009). Wissen prozessorientiert managen: Methode und Werkzeuge für die Nutzung des
Wettbewerbsfaktors Wissen in Unternehmen. München: Oldenbourg.
Heilbrunn, B. and I. Sammet (2017). “G-Learning- Gamification im Kontext von betrieblichem
eLearning”. In: Gamification und Serious Games. Ed. by S. Strahringer and C. Leyh. Edition HMD,
pp. 83–94.
Jane Zhao, Z. and J. Anand (2009). “A multilevel perspective on knowledge transfer: Evidence
from the Chinese automotive industry”. In: Strategic Management Journal 30.9, pp. 959–983.
König, E., ed. (2002). Qualitative Forschung: Grundlagen und Methoden. 2., völlig überarb. Aufl., Dr.
nach Typoskript. Vol. 8218. UTB. Weinheim u.a.: Beltz.
Kühl, S., ed. (2009). Handbuch Methoden der Organisationsforschung: Quantitative und qualitative
Methoden. 1. Aufl. Wiesbaden: Verl. für Sozialwiss. / GWV Fachverl.
Lakshman, C. and R. C. Parente (2008). “Supplier-Focused Knowledge Management in the Au-
tomobile Industry and Its Implications for Product Performance”. In: Journal of Management
Studies 42.2, pp. 317–342.
Lamnek, S. (2002). “Qualitative Interviews”. In: Qualitative Forschung. Ed. by E. König. Vol. 8218.
UTB. Weinheim u.a.: Beltz, pp. 157–193.
Liebig, B. and I. Nentwig-Gesemann (2009). “Gruppendiskussion”. In: Handbuch Methoden der
Organisationsforschung. Ed. by S. Kühl. Wiesbaden: Verl. für Sozialwiss. / GWV Fachverl., pp. 102–
123.
488
References
Mårtensson, M. (2000). “A critical review of knowledge management as a management tool”. In:

Journal of Knowledge Management 4.3, pp. 204–216.
Meuser, M. and U. Nagel (1991). “ExpertInneninterviews: vielfach erprobt, wenig bedacht : ein
Beitrag zur qualitativen Methodendiskussion”. In: Qualitativ-empirische Sozialforschung. Ed. by
D. Garz and K. Kraimer. Wiesbaden: VS Verlag für Sozialwissenschaften, pp. 441–471.
Müller, F. (2008). Die Gruppendiskussion als Instrument der qualitativen Marktforschung und der-
enAnwendung im Rahmen einer vergleichenden Analys zwischen physischem und digitalem
Musikgeschäft. Grin Verlag.
Nikodemus, P. (2017). Lernprozessorientiertes Wissensmanagement und kooperatives Lernen: Kon-
figuration und Koordination der Prozesse. AKAD University Edition.
Nonaka, I. and H. Takeuchi (1997). Die Organisation des Wissens: Wie japanische Unternehmen eine
brachliegende Ressource nutzbar machen. Frankfurt Main u.a.: Campus-Verl.
North, K. (2011). Wissensorientierte Unternehmensführung: Wertschöpfung durch Wissen. 5., aktu-
alisierte und erw. Aufl. Lehrbuch. Wiesbaden: Gabler.
Probst, G., S. Raub, and K. Romhardt (2012). Wissen managen: Wie Unternehmen ihre wertvollste
Ressource optimal nutzen. 7., überarb. und erw. Aufl. Wiesbaden: Springer Gabler.
Rack, O. and T. Christophersen (2009). “Experimente”. In: Methodik der empirischen Forschung.
Ed. by S. Albers, D. Klapper, U. Konradt, A. Walter, and J. Wolf. Wiesbaden and s.l.: Gabler Verlag,
pp. 17–32.
Radermacher, F. J., ed. (2002). Wissensmanagement: Überlebensfrage für Großunternehmen und
mittelständische Wirtschaft - das magische Dreieck: Humanressourcen, Organisation und IT-
Basis: Ergebnisband Ulmer Forum 2001. Vol. 19. Wissensverarbeitung und Gesellschaft. Ulm:
Universitätsverlag Ulm.
Reinmann-Rothmeier, G. and M. J. Eppler (2008). Wissenswege: Methoden für das persönliche
Wissensmanagement. 1. Aufl. Psychologie Praxis : Lernen mit neuen Medien. Bern: Hans Huber.
Riesenhuber, F. (2009). “Großzahlige empirische Forschung”. In: Methodik der empirischen Forschung.
Ed. by S. Albers, D. Klapper, U. Konradt, A. Walter, and J. Wolf. Wiesbaden and s.l.: Gabler Verlag,
pp. 1–16.
Rücker, B. (2002). “Überlebensfrage für Großunternehmen und mittelständische Wirtschaft: das
neue magische Dreieck, Human Resources, Organisation und IT-Basis”. In: Wissensmanage-
ment: Überlebensfrage für Großunternehmen und mittelständische Wirtschaft - das magische
Dreieck: Humanressourcen, Organisation und IT-Basis. Ed. by F. J. Radermacher. Vol. 19. Wis-
sensverarbeitung und Gesellschaft. Ulm: Universitätsverlag Ulm, pp. 29–45.
Schuh, G., R. Anderl, J. Gausemeier, M. ten Hompel, and W. Wahlster, eds. (2017). Industrie 4.0 Ma-
turity Index: Die digitale Transformation von Unternehmen gestalten. acatech STUDIE. München:
Herbert Utz Verlag.
Strahringer, S. and C. Leyh, eds. (2017). Gamification und Serious Games: Grundlagen, Vorgehen
und Anwendungen. Edition HMD.
Vogel, S. (2014). “Gruppendiskussion”. In: Handbuch Methoden der empirischen Sozialforschung.
Ed. by N. Baur and J. Blasius. Wiesbaden: Springer VS, pp. 581–586.
Weber, F. and J. Berendt (2017). Robuste Unternehmen: Krisenfest in Zeiten des Umbruchs. essentials.
Wiesbaden and s.l.: Springer Fachmedien Wiesbaden.
Weber, S. M. (2009). “Großgruppenverfahren als Methoden transformativer Organisationsforschung”.
In: Handbuch Methoden der Organisationsforschung. Ed. by S. Kühl. Wiesbaden: Verl. für Sozial-
wiss. / GWV Fachverl., pp. 145–179.
Yin, R. K. (2014). Case study research: Design and methods. 5. ed. Los Angeles CA u.a.: SAGE Publ.
489
Bibliometric Analysis of Risk
Management in Seaports
Ayman Nagi1 , Marius Indorf1 , Wolfgang Kersten1
Seaports play an essential role in global supply chains. Incidents and accidents in
such ports may lead to service delays, port closure as well as damage to people,
property and the environment. Therefore, risk management in seaports plays an
important role in mitigating and preventing possible accidents and disruptions.
This paper aims to explore the structure of literature on risk management in sea-
ports using co-citation analysis in order to reveal its current main research areas
and gaps as well as its future trends. Document co-citation analysis is performed
using the organization risk analyzer (ORA) software based on a specific threshold
and employing the CoCit-Score method of calculation. Suggested future research
areas on risk assessment and management methods as well as cooperative risk
management are revealed based on the results of the co-citation analysis.
Keywords: Risk management; Seaports; Co-citation analysis; Supply chain risk

management
491
Bibliometric Analysis of Risk Management in Seaports
1 Motivation and Introduction
More than 80 percent of global trade is forwarded by sea, making maritime trans-
port the most important mode of transport and a key enabler for international
commerce and globalization (UNCTAD, 2015, p. 22). This development is the result
of the continuous growth of seaborne shipments throughout the last centuries
and decades. Various commodities, such as oil, gas, iron ore, or grain are shipped,
totaling 10,047 million tons of cargo in 2015 which is an increase of around 200 mil-
lion tons compared to 2014 (UNCAD, 2016, p. 6). Naturally, all geographic regions,
e.g. the Americas, Asia-Pacific, or Europe, are involved in global seaborne trade
acting as exporter and importer in different extent. In any case, the cargo is han-
dled in large seaport facilities at the point of origin and destination. Depending on
the actual route, one or more transshipment ports are utilized in order to ensure
a time-saving and economic transport. Seaports connect the seaside/foreland
to the hinterland, where the cargo is further transported, handled, and finally
consumed.
In Europe’s top 20 cargo ports 1,723.4 million tons of goods were handled in 2015
based on its gross weight. Of this amount, 1,032.5 million tons were reported
on the direction inwards and 690.9 million tons were reported on the direction
outwards. In total, 3,838.3 million tons of goods (2,277.6 inwards and 1,560.7
outwards) were handled in all ports of the European countries in 2015 (Eurostat,
2017, pp. 2–3). On top of this, 395.4 million seaborne passengers embarked and
disembarked in all of the ports of the European countries in 2015 (Eurostat, 2017,
p. 6).
These figures illustrate the significance of seaports in the area of global trade and
transport. To cope with this challenge, seaports have become highly complex
systems involving various actors, such as operators, authorities, crews, port work-
ers, passengers, and the public (Andritsos and Mosconi, 2010, p. 1). In addition,
seaports comprise highly specialized physical installations and facilities, operated
according to demanding processes and procedures. In many cases, seaports have
a unique geographical location and have developed in immediate proximity to
residential areas. The port of Hamburg for example is a tidal port located at the
river Elbe and is the heart of Germany’s second largest city. In total, the city of
Hamburg and its metropolitan region are home to more than 5.2 million inhab-
itants (Statistical Office of the Free and Hanseatic City of Hamburg, 2017). As a
result, seaports have to fit in their surrounding community and to follow a vast
492
1 Motivation and Introduction
number of standards and regulations in order to become and remain a safe and
secure area for its internal and neighboring stakeholders.
To summarize, seaports play an essential role in modern trade and mobility and
are expected to continue growing also in upcoming decades in terms of volume
and complexity. At the same time, seaports have to follow strict principles and
have to burden enormous efforts in order to guarantee safety and security to its
bordering entities. This means, already today and even more in future seaports
face an extremely difficult situation, supporting global trade on the one hand and
ensuring safe and secure operations for its internal actors and its external envi-
ronment. Seaports face risk challenges from operational, technical and economic
perspectives (Alyami, et al., 2016, p. 10). Examples for these risks include flooding,
oil spills and fire explosions (Becker, et al., 2012; Valdor, Gómez and Puente, 2015;
Zhao, 2016). The only way to cope with these major challenges is to install and
maintain a proper risk management in those seaports. Due to its complex nature,
seaports require special risk management processes and measures that in many
cases span across multiple entities and organizations.
In scientific literature, the field of risk management in seaports has been present
at least since the mid-1980s and has gained increased attention throughout the
last decade. In contrast to most other objects considered by risk management
research, seaports rarely comprise a single organization. Thus, risk management
in the field of seaports needs to address these special circumstances. Reviewing
the available research on risk management in seaports, we realized the broad
range of perspectives and approaches in this field. In fact, they range from very
practical aspects, e.g. ballast water or oil spill treatment, to rather academic risk
management frameworks serving an overarching perspective. For this reason,
this paper aims to explore and to structure the scientific literature on risk man-
agement in seaports. The methodological approach chosen for this purpose is
a bibliometric analysis. Focusing on co-citations in particular, we identify main
areas as well as gaps in the current literature and propose future research areas
in the field of risk management in seaports.
In Chapter 1, we outline the motivation for our research and provide a brief intro-
duction to the topic. Chapter 2 frames the theoretical background on seaports
as well as risk management and shapes the scope of our work. Chapter 3 elabo-
rates the methodological approach chosen which is foremost characterized by
a co-citation analysis. The results of our research are presented in Chapter 4.
Conclusions based on the findings of the work undertaken are presented in Chap-
ter 5.
493
In this section, theoretical background in the area of seaports and risk manage-
ment is provided. The aspects addressed create the basis for understanding and
reflecting the chosen methodology and obtained results.
2.1 Seaports as Logistical Hubs for International Trade
Seaports are very complex systems and are critical elements of the global trans-
portation infrastructure (Christopher, 2015, p. 21). Even though seaports are im-
mobile conglomerations of different facilities and actors with a unique geographi-
cal position, they face severe competition within the international transportation
network. As a result, seaports organizations strive to meet the market require-
ments in the best possible fashion. In this context, it is vital to understand that
a port cannot be considered as an economic unit offering or producing a single
service to its customers. Instead, a broad range of processes and services is car-
ried out in the port area (Tovar, Trujillo and Jara-Díaz, 2004, p. 190). Hence, ports
in general and seaports, in particular, must be considered as a semi-structured
compound of individual organizations serving a wide diversity of customers.
Apart from rather large seaports, also a number of other ports exist. In fact, ports
can range from small-sized marinas servicing sport boat enthusiasts, to a mid-
sized company owned and operated industrial ports to the Port of Singapore,
which handles more than 1,000 vessels at any time of the day. Another dimension
to categorize ports apart from their size is their primary purpose. Ports can be
used for civilian, commercial or military/security purposes (Christopher, 2015,
p. 27). Obviously, ports can also have hybrid forms, serving multiple purposes
at the same time. In addition, the ownership structure of a port is important
to consider. Generally, two main types of ownership exist in view of the port’s
infrastructure, superstructure, and its staff. On the one side of the spectrum, ports
are owned and operated by private organizations and on the other side of the
spectrum, they are part of a complex body of local, state, or national government.
Very commonly, ports are owned and operated as landlord/tenant ports in which a
public authority rents the port areas to private actors, such as terminal operators,
whom in return pay a usage fee (Christopher, 2015, p. 27). A landlord model is
also a hybrid form of a fully public and a fully private port. The paper at hand
494
solely focusses on seaports, which we consider to be commercial ports, located

directly at or in close distance to the open sea.
The processes of transporting, handling, and storing cargo in a seaport involves

numerous actors. During the port entrance and exit of a ship, many shipping com-
panies rely on the assistance of pilots and tugboat companies for safely reaching
the quay wall. At the mooring, the cargo is unloaded by a terminal operator, which
also forwards the cargo to storing facilities and later passes it on to carriers trans-
porting the cargo through the hinterland. At any point in time, the processes are
controlled by the port authority. Apart from the mentioned actors, several others,
such as safety inspectors, customs, and repairers might be involved as well (Mar-
tino, et al., 2013, p. 125; Huber, 2014, p. 98). Apart from the different processes and
services performed in a seaport, the high number of actors involved clearly drive
complexity. This also underlines the importance of a thorough risk management
in seaports that serve as important logistical hubs for international trade. The
next section provides a brief overview of the literature on risk management in
seaports.
2.2 Risk Management in Seaports
There are many definitions for risk management in the literature. For instance,
Coyle, et al. (2010, p. 294) defined risk management as the process of identifying
risks along with its causes and effects in order to mitigate, prevent, transfer, or
eliminate possible threats to the overall supply chain success. Since seaports
play a core role in global supply chains, risk management is defined in this paper
as “the identification and management of risks for the supply chain, through a
coordinated approach amongst supply chain members, to reduce supply chain
vulnerability as a whole” (Jüttner, 2005, p. 124). The risk management process
normally includes risk identification, assessment, handling, and monitoring of
the implemented measures. (Garvey, 2008, p. 5).
Even though risk management has gained increased awareness among the actors
and has become a central point of attention in a seaport, universally applicable
standardized processes and measures universally applicable do not exist. In fact,
up to now, risk management is very port specific in terms of scope and substance.
Presumably, differing geographical and economic conditions in seaports, among
many others, are the main reasons in this context. In seaports, the actors employ
different risk management systems and approaches that define risk handling
495
initiatives in the case of an emergency, such as the explosion of gases or chemicals.

Apart from an intrinsic motivation to rely on suitable safety and security measures,
also various guidelines and standards oblige seaport organizations to conduct
a thorough risk management. These guidelines and standards exist on local,
national, and global levels. A few examples are presented below.
The ISO 31000, for example, provides principles and generic guidelines on risk
management that are applicable not only to seaport organizations but to any
public, private or community enterprise, association, group or individual (In-
ternational Organization for Standardization, 2009). Apart from such generic
standards, numerous additional approaches persist. One important document
is the International Ship and Port Facility Code (ISPS), which was established by
the International Maritime Organization (IMO). The ISPS Code contains minimum
requirements for port facilities and ship security being applied in all countries
that are part of the IMO convention. One aspect of the ISPS Code is that it reg-
ulates how and to which extent a ship is required to report freight and cargo to
the port which it seeks to enter (International Maritime Organization, 2003). In
order to implement the ISPS Code requirements, several legal provision, such as
the Maritime Transportation Security Act of 2002 in the U.S., were established.
This act contains regulations specifying requirements, e.g. for security assess-
ments or annual exercises and/or drills (U.S. Department of Homeland Security,
2003), and defines the implementation of the ISPS on a national level. In the
research literature, a broad range of approaches and perspectives in the area of
risk management in seaports exists. In the following, selected approaches are
presented.
Focusing on container terminal operational systems Alyami, et al. (2016) present

a method to facilitate the application of the Failure Mode and Effects Analysis
(FMEA) in assessing the safety performance. For this purpose, they develop a hier-
archical structure for risk factors during terminal operations and integrate a Fuzzy
Rule-Based Bayesian Network with Evidential Reasoning. This paper strongly
concentrates on the methodological approach in a locally limited environment
(Alyami, et al., 2016).
The paper of Bruzzone, et al. (2000) provides the development of an integrated

interactive environment for risk analysis applied to port and maritime environ-
ments. The presented approach helps to design the port and maritime infras-
tructure including its resources, structures, and services potentially confronted
with an emergency. The architecture of the tool developed is described and is
496
being applied by the integration of a specific simulation module to an oil spill

scenario.
In contrast to the aforementioned articles, the paper of Gharehgozli, et al. (2016)

proposes a conceptual framework on seaport resilience in the event of weather
disruptions. The suggested approach strategically helps ports to prepare for risks
associated with these potential disruptions and to design relevant contingency
plans. The proposed framework consists of four steps that port managers can
apply in order to increase port resilience. This paper understands ports as commu-
nities of many stakeholders and consequently has a more holistic perspective.
As a result of a European Union-funded project, Karlsson, Olsson and Riedel

(2016) present in their report several aspects of risk management. First, they dis-
cuss existing capability assessment approaches in the European Union (EU) and
Baltic Sea Region. Second, empirical data in this field is presented and analyzed.
Third, the EU document “Risk Management Capability Assessment Guidelines”
(2015/C261/03)” and its potential use are discussed. Fourth, a methodology for
risk and capability assessment is proposed as a result of the previous chapters.
This report has a clear geographical focus and features conceptual as well as
empirical elements.
The broad range of scientific approaches and perspectives in the area of risk
management in seaports underlines the necessity for structuring the field. Such
a structured overview would allow researchers to integrate their own activities
much quicker and to discover relevant research streams for future opportuni-
ties. So far, the literature is lacking a quantitative study analyzing bibliometric
information of available research work with the aim of framing the main research
areas and identifying existing research gaps. For instance, a study conducted by
Colicchia and Strozzi (2012) combined a SLR with a citation network analysis to
identify directions for future research in the field of supply chain risk management
with no focus on seaports. We address this fact with the work presented in this
paper. The next chapter presents the methodology applied in the course of this
research.
497
Figure 1: Research Design of the Performed Bibliometric Analysis
3 Methodology
The research design is developed according to specific research questions that ad-
dress the objectives of this paper. The authors identified three research questions
as follows:
1. What are the main and minor research areas of risk management in sea-
ports?
2. What are the current research gaps in the existing literature on risk man-
agement in seaports?
3. What are future research areas in the field of risk management in seaports
based on the current literature?
In order to answer these research questions, this paper employs a quantitative

approach in terms of a bibliometric analysis with a co-citation analysis as the
central methodological element. The overall research design of the performed
analysis is adapted and further developed from Soni and Kodali (2011, p. 241)
incorporating the bibliometric analysis as shown in Figure 1. The period of ex-
498
3 Methodology
tracting, cleansing and analyzing the articles lasted from April to June 2017 and
involved regular discussions and preparations among the authors.
First, a systematic identification of data sources is carried out by selecting scien-
tific databases and adequate keyword combinations to extract relevant articles
creating the data set for further analysis. Second, data cleansing and preparation
follows to check for duplicates and irrelevant articles. Additionally, the data set
obtained in this phase is transformed into a processible format. Third, the data
is analyzed in terms of a descriptive as well as a co-citation analysis. Then the
results of the co-citation analysis is further examined in order to reveal the main
research areas and gaps of current literature on risk management in seaports.
Fourth, future research areas for risk management in seaports are suggested.
The steps performed in each phase mentioned above are further detailed in the
following.
Two scientific databases were chosen in order to prepare the set of articles for
the bibliometric analysis using co-citation analysis. Web of Science (WOS) and
Scopus were selected since they provide the largest set of peer-reviewed papers
including required citation information as well as valuable cited-in references.
Google Scholar was used for extracting the cited-in references for each article. All
information combined constitutes the essential input for the co-citation analy-
sis.
Based on the objective of this paper, a specific set of keywords was defined en-
abling the extraction of suitable articles for the co-citation analysis. The keywords
selected cover risk management, risk analysis, risk assessment, and seaport.
The keywords were selected according to a related notion analysis that presents
the top keywords used for risk management in seaports. These keyword com-
binations were entered in the search field of Scopus and WOS using the logical
operators ”AND” and ”OR” as follow:
1. Risk Management AND (Seaport OR Sea Port),

2. Risk Analysis AND (Seaport OR Sea Port),
3. Risk Assessment AND (Seaport OR Sea Port).
The steps of data cleansing and preparation reflect the filtering mechanism that
is required in order to obtain a set of articles that is relevant to the field under
investigation. Duplications and articles with missing information were removed
from the dataset. Afterward, the abstract for each article was carefully read and
499
Figure 2: Steps of the Document Co-citation Analysis (based on Gmür, 2003)
examined to remove any irrelevant article. This final dataset of articles is used for
the descriptive and co-citation analysis.
The co-citation analysis as a research methodology is based on a co-occurrence

quantitative analysis technique that is used to measure the frequency of co-
occurrence for a pair of authors, titles, or other keywords in a document to map
the dynamics and structure of scientific research (Braam, Moed and van Raan,
1991, p. 24; Gmür, 2003, p. 27). The co-citation starts with defining the selected
object and the choice of the approach that is limited to an author or document
co-citation analysis. The document co-citation approach reflects the proximity in
terms of content within a school of research by analyzing the co-citation between
peer-reviewed selected documents. In contrast to this, author co-citation analysis
is based on the assumption that any co-citation between documents by any two
authors is a suitable measure for the proximity in terms of content in a certain
school of research. As the document co-citation analysis promises better results
in view of exploring the current research areas of risk management in seaports,
this approach is selected for our analysis. The steps of the document co-citation
analysis follow the steps shown in Figure 2.
After the data collection and preparation phase, specific methods are selected
for calculating the co-citation frequencies as well as to generate the co-citation
networks and clusters. Cluster distribution and analysis is followed by analyzing
the research stream of each cluster individually via an associated co-citation
network that shows the relationship between each reference in the given cluster.
Organizational risk analyzer (ORA) is used as the main software to extract and
analyze the clusters. ORA is a dynamic meta-network analysis and assessment
500
4 Results
tool developed by CASOS at Carnegie Mellon in Pittsburgh, U.S. The results of this
paper based on the aforementioned methodology are presented in the following
chapter.
4 Results
The results of this paper based on the elaborated methodology are presented in
this chapter. They comprise the descriptive analysis of the selected articles as
well as the process of clustering using the co-citation analysis.
4.1 Descriptive Analysis
Out of the 395 articles obtained from the database inquiries, 108 relevant papers
were carefully selected in the phase of data cleansing and preparation. Descriptive
figures are used in this subchapter to characterize this final dataset of relevant
articles. This descriptive analysis consists of the chronological trend of articles
since 1984, top 15 keywords of risk management at seaports as well as the top 10
cited authors.
There is an increasing trend of articles published containing aspects related to risk

management, risk assessment, and risk analysis in seaports, particularly starting
in 2008 with some deviations occurring in between. As observed in Figure 3, a
lower number of articles is observed in 2017 since the analysis was carried out
until April 2017.
An extraction of all author keywords from the selected databases was conducted
in order to extract the top 15 frequent keywords related to risk management in
seaports. Risk assessment and ballast water are frequent keywords that appear
in the selected articles. Many articles used ’port’ as a keyword associated with
the maritime environment whereas a set of other articles used ’seaport’ as a core
identifier. Ballast water management as well as studies that discuss the various
implications of climate change on ports and coastal cities are accompanied by
the keywords ’ballast water’ and ’climate change’. Several studies are related to
the concentration of heavy metals in sediments. All other keywords are presented
in Figure 4.
501
Figure 3: Number of Articles Identified per Year
502
4 Results
Figure 4: Top 15 Keywords associated with Risk Management in Seaports
503
The most cited first author based on the citation information from WOS and
Scopus in the selected 108 articles focused on papers related to the impact of
climate change and natural risks on seaports, such as flood losses in port cities is S.
Hallegatte. Similarly, the authors R. L. Wilby, S. Hanson, C. B. Awuor, and H. Sterr,
H presented studies related to climate change and flood risk as well as sea-level
rise and its impacts on port cities. For instance, S. Hanson provided a related
study on the assessment of exposure of large port cities to climate extremes such
as flooding and storm surge.
In contrast to the aforementioned authors, P. Trucco focused on the integration
of Human and Organizational Factors (HOF) into risk analysis by employing Fault
Tree Analysis and a Bayesian Belief Network (BBN). BBNs are used in order to
represent dependencies among a set of variables probabilistically (Cooper, 1990,
p. 393). Decision support framework for risk management and analysis of acci-
dents in seaports are topics that are as well highly cited (R. M. Darbra, K. Mokhtari).
The other set of authors focused on assessment studies related to risk assessment
of pollutant concentrations and collision avoidance (D. Haynes, J. M. Mou). The
top 10 cited authors are revealed in Figure 5.
Following the descriptive analysis of the selected articles, the results of the cluster
analysis based on the document co-citation are presented in the next section.
4.2 Co-citation Analysis
The results of the co-citation analysis are presented in this section based on the
aforementioned described steps (see Figure 2).
An additional filtering mechanism was applied to the original list of 108 articles
sorting out articles with less than two cited-in references. This means that articles
that are not cited or cited only once in other articles in the literature are filtered
out. Consequently, 15 articles are excluded from the final dataset. Additionally,
all cited-in references for each article were stored separately in Excel worksheets
that correspond to each article. These articles were downloaded using the library
of Google Scholar and saved as CSV files. In total, 2,702 cited-in references are
distributed among the 93 articles.
A 93 x 93 raw co-citation matrix was programmed using Visual Basic for Appli-
cations (VBA) in order to calculate the co-citation frequencies. The co-citation
matrix is an essential input for the ORA software. A Microsoft Excel macro was
504
4 Results
Figure 5: Top 10 Cited Authors in the Field of Risk Management in Seaports
505
programmed to generate the raw co-citation matrix by comparing the list of cited-
in references for each article in each worksheet. The macro loops through each
article and records the co-citation frequency in the relevant field in the co-citation
matrix.
The CoCit score was chosen as the main method for generating the co-citation
network and clusters. According to Gmür (2003, p. 40), the CoCit minimizes the
citation relation of both co-citation partners. The method takes a value between
0 and 1. It relates the sum of the co-citation count in relation to the mean and
minimum counts of the two individual citations. The CoCit is calculated using the
following formula (Gmür, 2003, p. 41):
(co-citation)2
CoCitAB =
min(citationA ; citationB ) ∗ mean(citationA ; citationB )
The analysis is done in the software ORA with a threshold value of 0.01. This
threshold value is manually adjusted until a clear pattern is detected. The co-
citation network as shown in Figure 6 has emerged with the associated identified
eight clusters. Additional revision of the abstracts and introduction parts of the
articles is conducted to extract the clusters from the network. Out of the total 67
co-cited articles in the final dataset, 36 are clustered references in the co-citation
network. Therefore, the network has a penetration value of 53.7%, which is used
to assess the coverage of the co-citation network. This is a higher value compared
to the value achieved by (Gmür, 2003, p. 45) who achieved a penetration value of
48% using the CoCit score.
In the following sections, the clusters obtained are further described. The articles
referenced as well as their associated reference number can be found in the
appendix.
Cluster I: Risk Management of Ballast Water (3 articles)
The first cluster is associated with studies related to risk management of ballast
water with a focus on the Baltic Sea. Ballast water management is a complex issue
that requires careful assessment of various species that are transported in ballast
water (Endresen, et al., 2004, p. 615). The cluster includes a risk assessment
study for exemptions related to intra-Baltic shipping using different methods
506
4 Results
Figure 6: Clusters of the CoCit Co-citation Network
507
such as environmental matching (David et al. 2013 - A23). The risk assessment
study of Gollasch and Leoppaskoski (2007 - A63) covers nine ports distributed
in five countries around the Baltic Sea. It presents scenarios of ballast water
management for intra-Baltic shipping as well as voyages outside the Baltic. A
discharge assessment model for ballast water was developed by David, et al.
(2012 - A92) to predict the possibility of ballast water discharge of vessels arriving
at ports. This model is based on vessel dimensions and vessel cargo operations.
Cluster II: Collision Risks (4 articles)
This cluster comprises studies that analyze and evaluate collision risks. A marine
complexity model was introduced by Wen, et al. (2015 - A2) to analyze the degree
of crowding and risk of collision. A similar model was built by Debnath, et al.
(2011 - A36) for collision risks using traffic conflicts. A binomial logistic model
was derived from traffic conflicts and calibrated by the authors for the Singapore
port fairways. As a proactive approach, a study for collision avoidance in busy
waterways was conducted by Mou, et al. (2010 - A48). Concerning navigational
collisions, Debnath and Chin (2009 - A76) used hierarchical modeling to analyze
the relationships between perceived risks, pilot attributes as well as geometric
and traffic characteristics of fairways.
Cluster III: Approaches and Studies for Improvements and Decision-making

(8 articles)
The third cluster comprises articles that analyze improvement solutions to sup-
port decision maker in aspects related to port disruptions, economic losses, miti-
gation strategies, ship safety index, decision support frameworks, and recommen-
dations for maritime safety. Economic losses of port disruptions, such as extreme
wind events, are mentioned in Zhang and Lam (2016 - A1) as well as Zhang and
Lam (2015 - A4). Li, et al. (2014 - A17) developed a ship safety index that can
aid different parties, such as port authorities, in examining areas that should be
inspected, repaired, and maintained based on a relative risk score. A short-term
wind forecast is described by Burlando, et al. (2014 - A13) to improve the safety of
the largest ports in Northern Tyrrhenian Sea. Lam and Su (2015 - A7) elaborated
mitigation strategies based on proactive and reactive measures to minimize the
probability of occurrence and severity of various types of port disruptions. A
508
4 Results
similar approach but related to inventory management for port-of-entry disrup-

tion risks including container seaports is discussed by Lewis, et al. (2013 - A24).
The authors investigated potential economic and operational impacts with the
focus on supply chain inventory as a mitigation strategy. Mokhtari, et al. (2012
- A32) developed a decision support framework using fuzzy set theory (FST) to
analyze risk factors for ports and terminal operations and management (PTOM).
FST is based on a mathematical framework that simplifies the study of vague
conceptual phenomena (Zimmermann, 2011, p. 318). Knapp and Franses (2009 -
A50) discussed the improvement in the risk profiling by combining data sources
on inspections to improve the maritime safety system that comprises different
players such as shipyards.
Cluster IV: Impacts of Natural Hazards and Climate Change on Coastal and Port
Areas (7 articles)
Studies of the impact of natural hazards and climate change on coastal and port
areas are the focus of this cluster. An economic model for the analysis of disaster
prevention investments is developed by Xiao, et al. (2015 - A6). An evaluation
framework for climate change and sea level rise potential impacts is presented by
Messner et al. (2013 - A26). Hallegatte, et al. (2011 - A41) conducted a similar study
by assessing the risks of sea level rise, climate change impacts and storm surge in
port cities. Kron (2013 - A27) elaborated in his paper the threats and associated
risks of natural hazards and recommended solutions and safeguards to provide
an efficient protection. A ranking based on the estimation of the exposure of large
port cities to storm surge and coastal flooding is presented by Hanson et al. (2011
- A42). Hallegatte, et al. (2011 - A83) and Raposeior, et al. (2013 - A88) elucidated in
their papers the impacts of flood losses along with an assessment of the flooding
risk of port and coastal areas.
Cluster V: Risk Assessment of Liquefied Natural Gas (3 articles)
This cluster comprises articles that focus on the risk assessment of Liquefied
Natural Gas (LNG) carriers. Risk assessment of LNG carriers using fuzzy TOPSIS
method for order preference was conducted by Elsayed, et al. (2014 - A21). Tech-
nique for Order of Preference by Similarity to Ideal Solution (TOPSIS), according
to Wang and Elhag (2006, p. 310), is one of the popular approaches for Multiple
509
Criterion Decision Making (MCDM). A software tool and a methodology for car-
riers during loading/offloading were developed by Elsayed, et al. (2009 - A55)
based on utility theory and multi-attribute risk assessment. A similar approach
for multiple attribute risk assessment based on fuzzy inference system (FIS) was
developed by Elsayed (2009 - A56). FIS is expressed in terms of “IF-THEN” rules to
predict uncertain systems (Kazeminezhad, Etemad-Shahidi and Mousavi, 2005,
p. 1710).
Cluster VI: Risk Assessment of Hazardous Spills (4 articles)
This cluster comprises articles that is related to the risk assessment of hazardous
spills. These spills include hazardous and noxious substances (HNS) as well as oil
spills. Harold, et al. (2014 - A15) developed a methodology to prioritize HNS with a
risk prioritization matrix to assess the acute risks of HNS spills. The classification of
risk zones represented in risk maps for oil spills was carried out by Singkran (2013
- A26). This classification considered the number of ports and the frequency of oil
spill incidents as well. Similarly, an oil spill hazard assessment was conducted
by Garcia, et al. (2013 - A82) in Italian ports based on the development of an Oil
Spill Hazard Index (OSHI) for hydrocarbons handled at ports and in transit. An
assessment of hazards from oil ship discharges, based on oil hazard maps, was
the focus of the study conducted by Liubartseva, et al. (2015 - A85).
Cluster VII: Environmental Risk Analysis (4 articles)
This cluster consists of four articles that elaborate the environmental analysis of
harbors. A subjective and objective assessment of environmental risks of a tourist
harbor in southern Italy was conducted by Irene, et al. (2010 - A51). A multistep
indicator-based approach that comprises the development of a tool, suitable
environmental quality indicators, and a user-friendly development scheme was
carried out by Marin, et al. (2008 - A59) to monitor environmental quality and the
associated risks of harbors. Environmental risk analysis of oil handling facilities
is presented by Valdor, et al. (2015 - A11). The main pollutant sources and a
numerical analysis of several pollution incidents are elaborated by Mestres, et al.
(2010 - A47) in order to assess the potential environmental risks.
510
4 Results
Cluster VIII: Risk Assessment Approaches and Frameworks (3 articles)
This cluster comprises three articles related to assessment approaches and frame-
works for risks in seaports. A fuzzy risk assessment approach for seaports was
carried out by John, et al. (2014 - A16) to analyze their complex structure of op-
erations. Pak, et al. (2015 - A78) followed a fuzzy analytical hierarchy process
to analyze risk factors that affect navigational safety with their associated impli-
cations in seaports. A generic bow-tie analysis framework combining Fault Tree
Analysis (FTA) and Event Tree Analysis (ETA) was used by Mokhtari, et al. (2011
- A37) for the risk assessment phase in seaports and offshore terminals. Based
on the results of the co-citation analysis, the main research areas and gaps are
elaborated on in the next section.
4.3 Main Research Areas and Gaps
The main research areas and gaps in the current literature on risk management in
seaports are elaborated in this subchapter based on the results of the co-citation
analysis.
4.3.1 Main Research Areas
Two main research areas are revealed based on the third and fourth cluster which
consist according to the clusters in the CoCit co-citation network of eight and
seven articles respectively. Approaches and studies for analysis and decision-
making are the first main research area which encompasses solutions and detailed
analysis for improvements as an aid for decision makers. The main aspects consid-
ered in this cluster comprise economic, safety and operational aspects. Examples
include studies related to economic losses, improvement in maritime safety and
operational impacts of inventory management due to port disruptions.
The second main research area is represented by the fourth cluster that presents
the impacts of climate change and natural hazards on coastal and port areas.
Based on Becker, et al. (2012), anticipations and assessment with regards to the
impacts of climate change should be taken into considerations to proactively
prepare for different natural hazards such as flooding and sea-level rise.
511
Figure 7: Minor Research Areas
4.3.2 Research Gaps
In order to detect the research gaps, the minor research areas were extracted
from the clusters in the co-citation network. The threshold to distinguish major
from minor research areas was set to be four articles. Figure 7 shows the minor
research areas based on the results of the co-citation analysis.
There is a lack of studies with respect to risk assessment and management meth-
ods that can be used by the diverse actors in a seaport. No studies, from the
dataset, covered risks such as the improper handling of dangerous goods and
the explosions of gases and chemicals. Two research areas focused on the quan-
titative assessment of risks at seaports in general and specifically on the risk
assessment of hazardous spills and LNG. This also includes the integration of
cooperative risk management. According to Mokhtari, et al. (2012, p. 5088), appro-
priate techniques to support the risk management cycle in seaports is required in
order to assess the overall risk level. These techniques should reflect the role of
each stakeholder at the seaport with regards to the risk management process.
512
5 Conclusion
4.4 Suggested Future Research Areas
There is a need to conduct further research on quantitative and qualitative risk

assessment and management methods that could be applied to the different
operations and sources of risks in seaports. For instance, operations related to
the loading, storage, and distribution of cargo. Additional research is suggested
with regards to risk assessment methods for the handling of dangerous goods
and the explosions of gases and chemicals along with case studies. Furthermore,
simulation approaches such as Monte Carlo simulation can support in aggregating
the various risks in the port system. Additionally, the future research work should
consider cooperative risk management for the parties involved in the different
operations of a seaport. A model for cooperative risk management will enable a
better preparedness and visibility for risks that occur within the port system. A
common language and understanding of risks should be developed and integrated
into a database management system in order to increase the risk management
know-how on one hand and to increase the visibility among the different members
of a port system on the other hand. This will allow the different members to
share an understanding of risks facing the seaport and guide them to implement
effective proactive measures as well as contingency plans.
5 Conclusion
This research paper analyzes current literature on risk management in seaports

and proposes a suitable structure. The methodological approach chosen in this
regard is a bibliometric analysis focusing on documents co-citations. From this
analysis, two main research areas were extracted: (1) Approaches and studies for
improvements and decision-making and (2) Impacts of natural hazards and cli-
mate change on coastal and port areas. Besides already existing research streams,
also potential research gaps were considered in our analysis. The first deficiency
identified is a lack of studies about risk assessment and risk management meth-
ods that are suitable for the usage across the different stakeholders in a seaport.
Additionally, a gap was detected concerning other types of risks, such as the im-
proper handling of dangerous goods. Therefore, the future research areas should
be tailored to quantitative and qualitative risk assessment and risk management
methods that could be used by the various stakeholders in the different operations
and sources of risk at seaports. This includes the consideration of a cooperative
513
risk management to understand and better identify, assess, and manage the risks
among the different stakeholders.
The study presented is limited to specific keywords in the data extraction phase
and may hence not cover all scientific papers in the field of risk management in
seaports. In addition, the dataset of the databases consulted may not be fully
comprehensive. The cluster analysis performed relies on the CoCit score as main
method for generating the co-citation network. The clusters created are based
on a specific threshold that is manually adjusted. Future work should consider
performing a co-citation proximity analysis to achieve a high penetration value.
The articles not clustered in in the co-citation analysis presented here could be
analyzed further in order to identify additional research trends and research areas.
Future research could also consider combining a SLR with a co-citation analysis
in order to provide an exhaustive overview of all possible research areas.
This work is supported by the EU project HAZARD (Mitigating the Effects of Emer-
gencies in the Baltic Sea Region Ports). HAZARD is a Flagship Project in the EU
Strategy for the Baltic Sea Region, which is partly funded by the EU’s Interreg
Program.
Appendix Referenzes for the clusters
514
5 Conclusion
515
516
5 Conclusion
517
518
5 Conclusion
519
References
Alyami, H., Z. Yang, R. Riahi, S. Bonsall, and J. Wang (2016). “Advanced uncertainty modelling for
container port risk analysis”. In: Accident; analysis and prevention.
Andritsos, F. and M. Mosconi, eds. (2010). Port Security in EU: a Systemic Approach.
Becker, A., S. Inoue, M. Fischer, and B. Schwegler (2012). “Climate change impacts on international
seaports: Knowledge, perceptions, and planning efforts among port administrators”. In: Climatic
change 110.1, pp. 5–29.
Braam, R. R., H. F. Moed, and A. F. J. van Raan (1991). “Mapping of science by combined co-citation
and word analysis I. Structural aspects”. In: Journal of the American Society for information
science 42.4, p. 233.
Bruzzone, A., R. Mosca, R. Revetria, and S. Rapallo (2000). “Risk analysis in harbor environments
using simulation”. In: Safety Science 35.1-3, pp. 75–86.
Christopher, K. (2015). Port security management. 2. ed., [elektronische Ressource]. Boca Raton,
Fla.: CRC Press/Taylor & Francis.
Colicchia, C. and F. Strozzi (2012). “Supply chain risk management: A new methodology for a
systematic literature review”. In: Supply Chain Management: An International Journal 17.4,
pp. 403–418.
Cooper, G. F. (1990). “The computational complexity of probabilistic inference using Bayesian
belief networks”. In: Artificial intelligence 42.2-3, pp. 393–405.
Coyle, J. J., R. A. Novack, B. Gibson, and E. J. Bardi (2010). Transportation: A supply chain perspective.
Cengage Learning.
Endresen, Ø., H. L. Behrens, S. Brynestad, A. B. Andersen, and R. Skjong (2004). “Challenges in
global ballast water management”. In: Marine Pollution Bulletin 48.7, pp. 615–623.
Eurostat (2017). Maritime Ports Freight and Passenger Statistics.
Garvey, P. R. (2008). Analytical methods for risk management: A systems engineering perspective.
Crc Press.
Gharehgozli, A. H., J. Mileski, A. Adams, and W. von Zharen (2016). “Evaluating a “wicked prob-
lem”: A conceptual framework on seaport resiliency in the event of weather disruptions”. In:
Technological Forecasting and Social Change 121, pp. 65–75.
Gmür, M. (2003). “Co-citation analysis and the search for invisible colleges: A methodological
evaluation”. In: Scientometrics 57.1, pp. 27–57.
Huber, T. (2014). ““Global Ports” in der maritimen Transportwirtschaft: Akteursbasierende Bewer-
tung des weltweiten Netzwerks von Hafenstandorten”. Dissertation. Regensburg.
International Maritime Organization (2003). International Ship and Port Facility Security Code.
International Organization for Standardization (2009). Risk management - Principles and guidelines.
John, A., D. Paraskevadakis, A. Bury, Z. Yang, R. Riahi, and J. Wang (2014). “An integrated fuzzy risk
assessment for seaport operations”. In: Safety Science 68, pp. 180–194.
Jüttner, U. (2005). “Supply chain risk management: Understanding the business requirements
from a practitioner perspective”. In: The International Journal of Logistics Management 16.1,
pp. 120–141.
Karlsson, B., A. J. Olsson, and M. Riedel, eds. (2016). From Gaps to Caps: Report on National
Capability and Risk Assessments and related challenges in the Baltic Sea Region.
Kazeminezhad, M. H., A. Etemad-Shahidi, and S. J. Mousavi (2005). “Application of fuzzy inference
system in the prediction of wave parameters”. In: Ocean Engineering 32.14, pp. 1709–1725.
520
References
La Tovar de Fé, B., L. Trujillo, and S. Jara-Díaz (2004). “Organization and regulation of the port
industry: Europe and Spain”. In: Essays on Microeconomics and Industrial Organisation, 2nd edn,
Physica-Verlag, a Springer-Verlag Company, Germany.
Martino, M. de, L. Errichiello, A. Marasco, and A. Morvillo (2013). “Logistics innovation in Seaports:
An inter-organizational perspective”. In: Research in Transportation Business & Management 8,
pp. 123–133.
Mokhtari, K., J. Ren, C. Roberts, and J. Wang (2012). “Decision support framework for risk manage-
ment on sea ports and terminals using fuzzy set theory and evidential reasoning approach”. In:
Expert systems with applications 39.5, pp. 5087–5103.
Soni, G. and R. Kodali (2011). “A critical analysis of supply chain management content in empirical
research”. In: Business Process Management Journal 17.2, pp. 238–266.
UNCAD (2016). Review of Maritime Transport 2016. [S.l.]: United Nations.
UNCTAD (2015). Review of Maritime Transport 2015. Review of maritime transport. New York and
Geneva: United Nations.
U.S. Department of Homeland Security, ed. (2003). Protecting America’s Ports: Maritime Trans-
portation Security Act of 2002.
Valdor, P. F., A. G. Gómez, and A. Puente (2015). “Environmental risk analysis of oil handling
facilities in port areas. Application to Tarragona harbor (NE Spain)”. In: Marine Pollution Bulletin
90.1, pp. 78–87.
Wang, Y.-M. and T. M. S. Elhag (2006). “Fuzzy TOPSIS method based on alpha level sets with an
application to bridge risk assessment”. In: Expert systems with applications 31.2, pp. 309–319.
Zhao, B. (2016). “Facts and lessons related to the explosion accident in Tianjin Port, China”. In:
Natural Hazards 84.1, pp. 707–713.
Zimmermann, H.-J. (2011). Fuzzy set theory—and its applications. Springer Science & Business
Media.
521
About HICL
Since 2006 the annual conference Hamburg International Conference of Logistics
(HICL) at Hamburg University of Technology (TUHH) is dedicated to facilitate the
exchange of ideas and contribute to the improved understanding and practice of
Logistics and Supply Chain Management. HICL provides a creative environment,
which attracts researchers and practitioners from all over the world.
This volume contains research contributions by an international group of authors

addressing innovative and technology-based approaches for logistics and supply
chains. They present business models and investment options for enhanced
strategic decision making as well as recent approaches for supply chain risk
management.
This volume, edited by Wolfgang Kersten, Thorsten Blecker and Christian Ringle,
provides valuable insights into the digitalization of Supply Chain Management and
Logistics with regard to:
 Innovation and Technology Management
 Advanced Manufacturing and Industry 4.0
 Risk and Security Management
ISBN (Print): 9783745043280

ISSN (Print) : 2635-4430
ISSN (Online) : 2365-5070

Hicl Vol 23 PDF

Hochgeladen von

Copyright:

Verfügbare Formate

Hicl Vol 23 PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Hicl Vol 23 PDF

Hochgeladen von

Copyright:

Verfügbare Formate

econstor

Make Your Publications Visible.

Kersten, Wolfgang (Ed.); Blecker, Thorsten (Ed.); Ringle, Christian M. (Ed.)

Proceedings of the Hamburg International Conference of Logistics (HICL), No. 23

Provided in Cooperation with:

This Version is available at:

Standard-Nutzungsbedingungen: Terms of use:

Wolfgang Kersten, Thorsten Blecker and

Digitalization in Supply Chain

Editors: Kersten, W., Blecker, T. and Ringle, C.M. 2015

Editors: Kersten, W., Blecker, T. and Ringle, C.M. 2014

Editors: Kersten, W., Blecker, T. and Ringle, C.M. 2013

Editors: Kersten, W., Blecker, T. and Ringle, C.M. 2012

…find more proceedings on https://hicl.org/publications

Digitalization in Supply Chain

Prof. Dr. Dr. h. c. Wolfgang Kersten

Edition 1st edition, September 2017

Cover (Photo) Nick Scheerbart – unsplash.com/@nck

I Innovation and Technology Management 1

A Peer-To-Peer Platform for Decentralized Logistics . . . . . . . . . . . . . 19

IoT-coordinated Logistics in Product-Service Systems . . . . . . . . . . . . 35

Towards a Web Based Transportation Infrastructure . . . . . . . . . . . . . 55

Lead-time Optimization Potential of Digitization in Air Cargo . . . . . . . . . 75

Investigating the Factors Influencing the Acceptance of Fully Autonomous Cars 99

Data Source Taxonomy for Supply Network Structure Visibility . . . . . . . . 117

Supply Chain Flexibility and SMEs Internationalization. . . . . . . . . . . . . 195

II Advanced Manufacturing and Industry 4.0 213

Data Mining and Fault Tolerance in Warehousing . . . . . . . . . . . . . . . 215

Influence of Patterns and Data-Analytics on Production logistics . . . . . . . 233

Use Case of Self-Organizing Adaptive Supply Chain . . . . . . . . . . . . . 255

Supply Chain Process Oriented Technology-Framework for Industry 4.0 . . . 275

Cooperation Strategies among SMEs for Implementing Industry 4.0 . . . . . 301

Design of a Cyber-Physical Production System for Semiconductor Manufacturing 319

III Risk and Security Management 341

Cybersecurity in Ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343

Smart Supply Chain Risk Management - A Conceptual Framework . . . . . . 361

Framework for Digitalized Proactive Supply Chain Risk Management . . . . . 381

An Adaptive Supply Chain Cyber Risk Management Methodology . . . . . . 405

Simulation-Framework for Illicit-Goods Detection in Large Volume Freight . . 427

Relational View on Collaborative Supply Chain Disruption Recoveries . . . . 449

Automotive Supply-Chain Requirements for Knowledge Management . . . . 467

Bibliometric Analysis of Risk Management in Seaports . . . . . . . . . . . . 491

Prof. Dr. Dr. h. c. Wolfgang Kersten

Innovation and Technology

Keywords: Blockchain; Logistics; Supply Chain Management; Use Cases

Blockchain is everywhere. Invented by Satoshi Nakamoto, one or more mysterious

The Blockchain is a distributed digital ledger of transactions that cannot be tam-

Decentralized Verified Immutable

through peer-2-peer through signatures through consensus

Figure 1: Basic Properties of Blockchain

3 Blockchain in Logistics and SCM

As introduced, Blockchain is considered to offer large potential for improving

1 Global container shipping still involves a lot of

2 Counterfeit medicine is a growing problem for

3 In the food supply chain, foodborne out-breaks

4 More and more logistics objects are equipped

Figure 2: Overview of Use Case Exemplars

3.1 Ease Paperwork Processing in Ocean Freight

International container transports have a long trail of paperwork associated with