Combining Foresight

Methods for Impacts

Dr. Michael Keenan
Manchester Institute of Innovation Research
(and Directorate for Science, Technology & Innovation, OECD)
(with inputs from Rafael Popper, MIoIR)

NISTEP 3rd International Conference on Foresight

Tokyo, November 2007
 Outline
• Framing the issue

• How are methods used? Evidence from foresight mapping

• Using and selecting methods

• Further exploring the relation between methods and objectives

(and expected impacts)

• Ordering methods according to foresight ‘principles’

• Ordering methods according to foresight ‘stages’

• Summary conclusions
 What’s the issue?

• There are already many different methods used in foresight

exercises, with more methods coming online all the time

• Methods can be combined in many different ways to create

an overall methodology (foresight process)

• It would seem obvious that any methodological approach

should be sensitive to the sorts of impacts sought from
foresight

• Ideally, it should be possible to select and combine methods

to achieve certain impacts. But how well do we understand
this relationship? And can we use it to guide methodological
design?
Which are the most
popular methods?

Results of an analysis
of the European
Foresight Monitoring
Network (EFMN)
database
(~800 exercises)
 Common Foresight Methods
• The most widely used methods are without doubt
literature review (437), expert panels (397) and scenarios (324).
Despite these high numbers, we still believe that literature review
and other generic methods are being under-reported in the
database; it is hard to imagine a study without some review of
relevant literature, in particular.

• Other commonly used methods are futures workshops (195),

brainstorming (157), trend extrapolation (133), interviews (127),
questionnaires / surveys (121), Delphi (120), key technologies
(120), megatrend analysis (110) and SWOT analysis (107).

• Some less frequently used methods are technology roadmapping

(76), environmental scanning (69), modelling and simulation (52),
essays (50) and backcasting (42). More than half of the cases
using technology roadmapping are from North America. We are
surprised by the low frequency with which scanning is reported.

• Rarely used methods include stakeholder mapping (30), citizen

panels (28), structural analysis (13), cross-impact analysis (12),
multi-criteria analysis (11), bibliometrics (7), gaming (4),
morphological analysis (4) and relevance trees (2). The numbers
here may not do justice to the application of some of these tools in
sub-national exercises in France and Spain, where methods such
as structural analysis, morphological analysis and relevance trees
are known to have been applied. We hope that further mapping
efforts will improve the sub-national data and give a more accurate
picture here.

Methods Frequency (Popper et al, 2007)

 Top 10 foresight methods per region (Popper et al,
al, 2007)

EU27+ Trans-Europe North America Latin America Asia Africa Oceania

Top 10 (485 cases and (61 cases and (109 cases and (24 cases and (51 cases and (10 cases and (15 cases and
1835 methods) 192 methods) 328 methods) 188 methods) 280 methods) 47 methods) 35 methods)
Average 4 Average 3 Average 3 Average 8 Average 6 Average 5 Average 2
Literature Literature
Expert Panels Other methods Expert Panels Scenarios Backcasting
1 Review Review
(57%) (71%) (80%) (60%) (33%)
(63%) (48%)

Futures Megatrend
Expert Panels Scenarios Expert Panels Scenarios Interviews
2 Workshops Analysis
(52%) (41%) (67%) (57%) (33%)
(46%) (50%)

Literature Literature Literature Literature

Scenarios Expert Panels Citizen Panels
3 Review Review Review Review
(47%) (30%) (33%)
(45%) (67%) (55%) (50%)
Technology Environmental Futures Questionnaire
Other methods Futures Workshops Interviews
4 Roadmapping Scanning Workshops / Survey
(24%) (23%) (45%)
(39%) (63%) (40%) (27%)

Futures Questionnaire Megatrend

Brainstorming Key Technologies Brainstorming Expert Panels
5 Workshops / Survey Analysis
(21%) (28%) (63%) (40%)
(22%) (39%) (20%)
Megatrend Questionnaire Trend
Brainstorming Scenarios Brainstorming Essays
6 Analysis / Survey Extrapolation
(20%) (17%) (37%) (30%)
(19%) (58%) (20%)
Trend Megatrend Questionnaire
Trend Extrapolation Interviews Delphi Delphi
7 Extrapolation Analysis / Survey
(19%) (50%) (35%) (20%)
(19%) (16%) (30%)
SWOT Trend Modelling &
Delphi Other methods Interviews Scenarios
8 Analysis Extrapolation simulation
(17%) (19%) (10%) (13%)
(50%) (27%) (30%)
SWOT Modelling & Megatrend Trend
Essays Scenarios Brainstorming
9 Analysis simulation Analysis Extrapolation
(6%) (42%) (13%)
(15%) (13%) (25%) (30%)
Questionnaire Trend Structural Modelling &
Interviews Other methods Expert Panels
10 / Survey Extrapolation analysis simulation
(15%) (30%) (13%)
(13%) (6%) (38%) (25%)

Source: Popper et al (2007)

 The A to Z of Foresight Methods Combination (Popper et al,
al, 2007) 1/3
METHODS A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
A Backcasting 37% 5% 26% 7% 47% 23% 5% 23% 47% 21% 9% 14% 2% 47% 12% 5% 12% 7% 2% 16% 5% 9% 0% 19% 28%

B Brainstorming 11% 7% 26% 9% 69% 43% 1% 26% 70% 14% 3% 19% 1% 45% 31% 6% 30% 31% 2% 13% 8% 7% 1% 9% 18%

C Citizen Panels 12% 59% 41% 18% 76% 71% 0% 47% 47% 6% 6% 35% 6% 59% 41% 6% 18% 0% 0% 24% 24% 0% 0% 0% 18%

D Environmental Scanning 18% 60% 12% 13% 62% 40% 3% 25% 80% 13% 3% 28% 2% 47% 33% 10% 23% 25% 3% 27% 13% 10% 5% 15% 35%

E Essays 5% 19% 5% 13% 32% 29% 2% 17% 49% 32% 5% 14% 2% 33% 14% 5% 5% 10% 3% 8% 6% 5% 2% 6% 22%

F Expert Panels 6% 27% 4% 10% 6% 34% 1% 20% 65% 16% 1% 17% 0% 34% 15% 2% 17% 22% 1% 7% 3% 16% 1% 5% 15%

G Futures Workshops 5% 32% 6% 13% 9% 64% 2% 13% 61% 21% 2% 13% 1% 41% 14% 2% 13% 23% 1% 7% 3% 18% 0% 5% 14%

H Gaming 50% 50% 0% 50% 25% 75% 75% 0% 50% 25% 0% 0% 0% 100% 0% 0% 25% 0% 0% 50% 0% 0% 0% 75% 50%

I Interviews 9% 32% 7% 13% 10% 63% 21% 0% 65% 15% 3% 42% 1% 35% 17% 4% 8% 9% 4% 9% 5% 7% 2% 6% 19%

J Literature Review 5% 24% 2% 12% 7% 57% 28% 0% 18% 16% 1% 15% 0% 41% 14% 2% 15% 20% 2% 5% 2% 12% 1% 8% 21%

K Megatrend Analysis 8% 16% 1% 7% 17% 50% 33% 1% 14% 55% 2% 24% 1% 49% 9% 3% 13% 21% 3% 4% 2% 6% 1% 16% 24%

L Morphological Analysis 80% 80% 20% 40% 60% 60% 80% 0% 60% 80% 40% 40% 20% 100% 40% 40% 20% 20% 0% 60% 60% 20% 0% 40% 40%

M Questionnaire / Survey 5% 23% 5% 15% 8% 56% 22% 0% 42% 57% 25% 2% 1% 38% 20% 7% 19% 8% 3% 6% 8% 4% 1% 7% 20%

N Relevance Trees 50% 50% 50% 50% 50% 50% 50% 0% 50% 100% 50% 50% 50% 50% 50% 100% 0% 0% 50% 50% 50% 0% 0% 50% 50%

O Scenarios 6% 20% 3% 9% 7% 40% 25% 1% 13% 54% 19% 2% 14% 0% 11% 2% 12% 9% 0% 5% 3% 5% 0% 12% 24%

P SWOT Analysis 6% 52% 8% 24% 11% 66% 33% 0% 23% 70% 13% 2% 28% 1% 42% 8% 20% 23% 2% 14% 11% 6% 2% 2% 14%

Q Cross-Impact Analysis 15% 62% 8% 46% 23% 62% 23% 0% 38% 54% 31% 15% 62% 15% 46% 54% 15% 23% 15% 15% 38% 8% 0% 23% 15%

R Delphi 5% 42% 3% 14% 3% 61% 25% 1% 9% 61% 16% 1% 22% 0% 38% 17% 2% 28% 2% 4% 0% 6% 1% 2% 11%

S Key Technologies 3% 39% 0% 14% 5% 71% 39% 0% 9% 75% 23% 1% 8% 0% 25% 17% 3% 25% 3% 2% 0% 35% 2% 7% 10%

T Multi-criteria Analysis 13% 38% 0% 25% 25% 38% 25% 0% 50% 88% 50% 0% 38% 13% 13% 25% 25% 25% 38% 25% 0% 13% 13% 38% 38%

U Stakeholder Mapping 24% 62% 14% 55% 17% 83% 45% 7% 34% 66% 17% 10% 24% 3% 55% 41% 7% 14% 7% 7% 21% 0% 10% 17% 41%

V Structural Analysis (MICMAC) 15% 85% 31% 62% 31% 77% 46% 0% 46% 62% 15% 23% 69% 8% 62% 69% 38% 0% 0% 0% 46% 0% 0% 8% 15%

W Technology Roadmapping 6% 14% 0% 8% 4% 80% 48% 0% 11% 70% 10% 1% 6% 0% 23% 7% 1% 8% 55% 1% 0% 0% 0% 7% 11%

X Bibliometrics 0% 40% 0% 60% 20% 60% 0% 0% 40% 80% 20% 0% 20% 0% 20% 40% 0% 20% 40% 20% 60% 0% 0% 20% 60%

Y Modelling and simulation 14% 23% 0% 16% 7% 30% 18% 5% 13% 59% 34% 4% 14% 2% 66% 4% 5% 4% 14% 5% 9% 2% 9% 2% 45%

Z Trend Extrapolation 8% 17% 2% 15% 10% 38% 18% 1% 15% 62% 20% 1% 16% 1% 51% 8% 1% 8% 8% 2% 8% 1% 6% 2% 17%

13,088 combinations 229 898 132 454 288 1652 961 32 588 1860 573 62 581 25 1289 518 110 504 588 62 237 128 338 38 289 652

2,584 applications 43 140 17 60 63 361 190 4 113 414 119 5 113 2 309 83 13 100 110 8 29 13 71 5 56 143
3 categories Qualitative Semi-quantitative Quantitative
Source: Popper et al, 2007 (Global Foresight Outlook 2007)
 The A to Z of Foresight Methods Combination (Popper et al,
al, 2007) 2/3

• As expected, most methods are highly combined with expert panels, literature review and
scenarios. However, in order to avoid repetitions we do not refer to these in subsequent
highlights but we hope the reader will keep this in mind.
• Backcasting is often combined with brainstorming (37%), trend extrapolation (28%) and
environmental scanning (26%).
• Brainstorming is often combined with futures workshops (43%), SWOT (31%), key
technologies (31%), Delphi (30%), environmental scanning (26%) and interviews (26%).
• Citizen panels are very often combined with futures workshops (71%), brainstorming (59%),
interviews (47%), environmental scanning (41%), SWOT (41%), and questionnaire /survey
(35%).
• Environmental scanning is often combined with brainstorming (60%), futures workshops
(40%), trend extrapolation (35%), SWOT analysis (33%), questionnaires / surveys (28%),
stakeholder mapping (27%), interviews (25%) and key technologies (25%).
• Essays are often combined with megatrend analysis (33%) and futures workshops (29%).
• Expert Panels are often combined with futures workshops (34%) and brainstorming (27).
• Futures Workshops are often combined with brainstorming (32%).
• Gaming was only applied in 4 cases of the sample and it was mainly combined with futures
workshops, modelling and simulation.
• Interviews are often combined with questionnaires / surveys (42%) and brainstorming (32%).
• Literature Review is commonly combined with futures workshops (28%).
• Megatrend Analysis is commonly combined with futures workshops (33%).
• Morphological Analysis was used in 5 cases. It was combined with backcasting,
brainstorming, stakeholder mapping and structural analysis.
 The A to Z of Foresight Methods Combination (Popper et al,
al, 2007) 3/3

• Questionnaires / surveys are often combined with interviews (42%) and megatrend analysis
(25%).
• Relevance Trees was used only in 2 cases. In both cases it was combined with cross-impact
analysis.
• Scenarios are commonly combined with futures workshops (25%).
• SWOT Analysis is commonly combined with brainstorming (52%), futures workshops (33%) and
questionnaires / surveys (28%).
• Cross-Impact Analysis is often combined with brainstorming (62%) and questionnaires /
surveys (62%).
• Delphi is commonly combined with brainstorming (42%), key technologies (28%) and futures
workshops (25%).
• Key Technologies is commonly combined with brainstorming (39%), futures workshops (39%),
technology roadmapping (35%) and Delphi (25%).
• Multi-criteria Analysis has been used in 8 cases only with half of those combining it with
interviews and megatrend analysis.
• Stakeholder Mapping is often combined with brainstorming (62%), environmental scanning
(55%), futures workshops (45%), SWOT (41%) and trend extrapolation (41%).
• Structural Analysis is often combined with brainstorming (85%), questionnaires / surveys
(69%), SWOT (69%) and environmental scanning (62%) and stakeholder mapping (46%)
• Technology Roadmapping is often combined with key technologies (55%) and futures
workshops (48%).
• Bibliometrics was used in 5 cases and was mainly combined with environmental scanning,
stakeholder mapping, and trend extrapolation.
• Modelling and simulation is often combined with trend extrapolation (45%) and megatrend
analysis (34%).
• Trend Extrapolation is mainly combined with the three most common methods (expert panels,
literature review and scenarios).
Methods vs. Time Horizon (Keenan et al, 2006)
Brainstorm ing 2010
2015
E xpert P anels 2020

Questionnaire / S urvey 2025

2030
Literature Review 2050

Interviews
Trend E xtrapolation
S WOT Analysis
M odelling & sim ulation
M egatrend Analysis
S cenarios
Key Technologies
Futures Workshops
E ssays
Delphi
Technology Roadm apping
0% 20% 40% 60% 80% 100%
No. of Methods reportedly used in Foresight exercises (Keenan et al, 2006)

160

140

120

100

80

60

40

20

1 2 3 4 5 6 7 8 9 10 11
 Typical selection criteria
• Proof of concept – learning from other sites of application

• Available resources (time, money . . . )

• Nature of desired participation

• Suitability for combination with other methods

• Desired outputs of a foresight exercise (e.g. product vs.

process)

• Quantitative / Qualitative data requirements of methods

• Methodological competence often a key factor

 Let’s start with objectives (a proxy for impacts)
(Keenan and Popper (eds.), 2007)

Informing decision-making processes

• Formulate funding and investment priorities for public policies • Evaluate existing strategies against
potential futures, and devise future-proof strategies • Develop reference material for policy-makers
and other actors to use, broadening the knowledge base around which decisions are made, thereby
resulting in better informed public policies or organisational strategies • Provide anticipatory strategic
intelligence to innovation system actors • Detect and analyse weak signals to ‘foresee’ likely future
changes and to gain insights into complex interactions and emerging drivers of change • Identify new
S&T, business, societal, policy and political opportunities • Increase awareness of possible risks, and
hence the basis for more effective contingency planning, and the design and development of appropriate
forms of resilience

Assisting the implementation of decisions

• Improve implementation by enabling buy-in to decision-making processes • Increase understanding
and build trust between participants, thereby contributing to the building of shared agendas • Develop
widely shared visions of the future with which actors can identify and thereby better co-ordinate their
activities, be they individuals or organisations • Disrupt ‘lock-in’ thinking and challenge fixed mindsets
• Build hybrid networks and strengthen communities • Aid communication, understanding and
collaboration across boundaries, be they geographical, organisational or disciplinary in nature •
Deepen dialogue with society and improve governance

Creating new capabilities

• Enhance strategic capabilities of organisations by helping to develop a language and practice for thinking
about the future – something that is often termed a ‘foresight culture’ • Enhance the standing and
image of organisations using foresight, showing them to be future-oriented and open, and attractive
places for investment
 Objectives Methods?

• Each objective could be addressed through any number

of methodological approaches

• Moreover, most exercises will have multiple objectives

• Result: Complexity!

• Here, we present two alternative starting points for

ordering and thereby selecting methods:
– Principles of foresight
– Stages in the foresight process
 (Methodological?) Principles

Future-oriented
Participative
Evidence-based
Multidisciplinary
Coordinating
Action-oriented

There have been some attempts to map methods against principles such as these
 Creativity
Foresight
Methods Wild Cards

Diamond
Science Fiction
Simulation Gaming
Essays / Scenario writing
Genius forecasting Role Play/Acting
Backcasting SWOT Brainstorming
Relevance trees / Logic chart Scenario workshop
Roadmapping Delphi Survey Citizen Panel
Expertise Expert Panel Morphological analysis Conferences / Workshops Interaction
Key/Critical Technologies Multi-criteria Voting / Polling
Quantitative Scenarios/SMIC Stakeholders Analysis
Interviews Cross-impact / Structural analysis
Indicators / TSA Patent analysis
Bibliometrics Benchmarking
Extrapolation Scanning
Literature review
Qualitative (19)
Modelling
Semi-quantitative (8)
Quantitative (6)
R. Popper (2008) Evidence
Tracing a methodology – hypothetical case featured in the Guide to RI Foresight
(Keenan and Popper (eds.), 2007)

Large-Scale RIs Research Process (RP) Diamond

Stage 1: (deskwork) to map current RI

capacities and limitations (based upon
expert interviews), extrapolation of
important trends, and international
benchmarking with the US and Japan.
Scenario writing

Stage 2: International workshop to

identify and scope possible RI options.

Stage 3: Expert panel to define statements Delphi

for a Delphi, to be used to obtain views on Expert panel Workshops
RI options and the factors that underpin
them. International online Delphi. Interviews

Stage 4: (deskwork) to generate baseline Extrapolation Benchmarking

scenarios that are used to ‘test’ the
spectrum of RI options.

Stage 5: Multiple options drafted that set

out assumptions and priorities. These are
discussed and revised in workshops.
Assessment of foresight methods
Barend van der Meulen, 2007
 Stages Approach: Cycles of ‘Extension’ and
‘Concentration’ (Rémi Barré, 2001)

• Foresight exercises consist fundamentally of a succession of

‘extension’ and ‘concentration’ steps: the participants in the
exercise engage in interactive activities consisting of an
exploration and hypothesis-building stage (extension),
followed by a selection – convergence and synthesis stage
(concentration).

• Foresight methodologies are the ways by which these

extension and concentration steps are carried out. Such
extension - concentration sequences lead to a description of
Foresight as a learning process from tacit to codified
knowledge transformation cycles
Stages using Nonaka’s SECI Model
(Eerola & Joergensen, 2002)
 Z_punkt Corporate Foresight Toolbox
Basic Process

Monitoring Analysis Projection Transformation

Recognising Understanding Anticipating Draw

Relevant Trends. Drivers of Change. the Future. Implications for
Business.

Toolbox

Weak Signals Key Factors Scenarios Strategic

Management

Environmental Emerging
Trends Roadmaps
Issues
New Business
Development
Consumer/ Creative Processes
Industry Trends Wild Cards
/Workshops

Portfolio
Z_trend database Customer
Innovation Innovation
Foresight
Contexts
 Five mental acts (stages) of Foresight
• Understanding
– Gains a shared understanding and mutual appreciation of topics and
influencing factors as systems in their own contexts

• Synthesis and Models of the future

– Explores alternative courses for development and their integration into
designs for a new context

• Analysis and Selection

– Analyses the alternative futures and decisions on the desired future

• Transformation
– Establishes the relationship between the future and the present for the
change programme

• Action
– Creates plans to inform present day decisions concerning immediate
change actions to provide structural and behavioural transformations
Systemic Foresight Methodology,
SFM [Saritas, 2006]
 Mental Acts & Methods
Understanding Synthesis & Analysis & Transformation Actions
Models Selection

SWOT Scenario
Scanning Gaming Priority Lists
Analysis Planning
Creativity / Expertise / Evidence

Multi Criteria Critical/Key

Bibliometrics Scenarios Backcasting
Analysis Technologies

Literature Wild Card Morphological Relevance R&D

Review Analysis Analysis Trees Planning

Syst.Dynamics Cross Impact Road Action

Interviews
Simulation Analysis Mapping Planning

Prioritisation Linear Operational

Benchmarking Weak Signals
(e.g. Delphi) Programming Planning

Syst. Analysis Mega trends Strategic Impact

Voting
Modelling Analysis Planning Assessment
Interaction

Panels Panels Panels Panels Panels

Conferences Conferences Conferences Conferences Conferences
 Summary conclusions

• We have sought to present some of the ways foresight

analysts have attempted to classify methods with a
view to aiding their selection by practitioners

• However, all attempts at ordering methods should be

treated with caution because of
– Complexity of multiple objectives / principles / stages
– Versatility of many methods

• Such classifications should therefore serve only as a

guiding point of departure and should not be strictly
adhered to
 References
• Barré, R (2001) “Synthesis of Technology Foresight” in Tubke et al, Strategic Policy
Intelligence: Current Trends, the State of Play and Perspectives, Seville: IPTS
• Eerola, A & Joergensen, BH (2002) Technology Foresight in the Nordic Countries,
report to the Nordic Industrial Fund, Center for Innovation and Commercial
Development
• Keenan, M et al (2006) Mapping Foresight in Europe and other Regions of the World:
The EFMN Annual Mapping Report 2006, report to the European Commission, Delft:
TNO
• Keenan, M & Popper, R (eds.) (2007) Practical Guide for Integrating Foresight into
Research Infrastructures Policy Formation, ForeIntegra
• Popper, R (2008) ‘Foresight Methodology’, in Georghiou, L, Cassingena H, J., Keenan,
M, Miles, I, and Popper, R, The Handbook of Technology Foresight, Cheltenham:
Edward Elgar
• Popper, R, Keenan, M, Miles, I, Butter, M & Sainz, G (2007) Global Mapping Report
2007, European Foresight Monitoring Network report to the EC www.efmn.eu
• Saritas, O (2006) Systems Thinking for Foresight, Unpublished Ph.D. Thesis, PREST,
University of Manchester
• Van der Meulen, B (2007) Looking Beyond the Endless Frontier, Strasbourg: European
Science Foundation
 Thank you!

michael.keenan@manchester.ac.uk
michael.keenan@oecd.org