Designing a Digital Environment to Support the Co-production of Public Services: Balancing Multiple Requirements and Governance Concepts

Co-production of public services means that services are not only delivered by professional and managerial staff in public agencies but also collaboratively produced and then delivered by citizens and communities [8]. For this reason, co-production is a more inclusive way of developing and delivering public goods and services, a new form of governance in which citizens and private actors - including non-governmental and civil society organizations - are involved in the overall creation and execution process, as well as in the long-term sustainability of services [76]. Co-production emphasizes people as active agents, not passive beneficiaries; and, in large part because of this alternative process, it tends to lead towards better outcomes in the long-term. This growing interest towards new and more participative governance models has emerged because of the failure of the traditional top-down system for providing public goods and services [27] and because of the recent socio-economic crisis and the related challenges public services are facing, such as increased demand, rising expectations and, in many cases, reduced budgets [12]. Linders [50] points out that the distribution of power and responsibility among public and private actors may vary significantly, giving rise to different governance models underlying co-production. “Citizen Sourcing” (or Citizen to Government, C2G) happens when citizens help the government improve public performance and more effectively satisfy the public's wants and needs. “Public-Civic Partnership” (or Government with Citizen, G+C) occurs when the government and citizens share equal power and responsibility. “Government as a Platform” (or Government to Citizen, G2C) occurs when the government helps citizens improve their day-to-day productivity, decision-making, and well-being by offering a facilitating platform. “Do It Yourself Government” (or Citizen to Citizen, C2C) occurs when citizens help themselves and one another; in this case, the government plays no active role in day-to-day activities but may provide a facilitating framework. Different governance models imply different collaborative processes, which therefore can be facilitated and impacted by the introduction of ICT in a varied way [50].

Despite the growing recognition of the benefits of co-production, studies suggest that co-production complicates governance and accountability, and needs high levels of trust among the various parties to successfully achieve results [78]. To achieve co-production processes and effectively impact policies and mainstream public services, several requirements have been identified. Firstly, the right organizational structures and culture should be developed, with low centralization and high connectedness among actors, and with an ethic of mutuality and equality between the different actors involved, including civil society [73]. Three main interconnected sets of resources and capabilities are required to carry out successful co-production initiatives: (i) analytical resources to ensure policy action; (ii) operational resources to mobilize and deploy means, to carry out coordinated actions, and engage policy networks, communities, and individuals; and (iii) political support and resources to ensure political legitimacy and two-way communication with non-state actors [38]. Finally, co-production calls for citizen participation and for a more active citizenship in service delivery, in addition to a more active role for third sector organizations [60].

The work presented in this paper particularly focuses on operational resources to facilitate coordinated action and to monitor and manage the involvement of heterogeneous networks of stakeholders. We also take a technological stance to investigate the supportive role of ICT.

Co-production encompasses a diverse range of practices and activities. Depending on the level of citizens' involvement, on the type of service, and on who is taking the initiative, different strategies, activities, and responsibilities may be needed. The identification of the actual sequence of phases to be followed by networks of public-private stakeholders to co-produce services is a subject of debate and research. Besides the main co-design [83] and co-deployment phases [11, 1], studies recognize the crucial importance of the preliminary phases of commissioning [55], and of political sponsorship [48], stakeholders’ engagement [83], assessment [55], and sustainable service maintenance in the long-term [7]. Although it is not possible to freeze co-production into a one-fits-all model of programmatic steps to perform, it is possible to identify phases and activities that are crucial for every co-production initiative (like formulating a shared problem definition or identifying and engaging stakeholders), phases that should iteratively occur periodically (like evaluation points and go-no-go moments), and optional phases that may be included in certain co-production contexts (like the planning of maintenance for co-produced digital services). Brandsen et al. [7] propose an analysis of co-production phases and the related governance issues that stakeholders should consider and solve. Sample phases and subphases are shown in Table 1. For example, governance issues for the subphase “Identification and engagement of stakeholders” are: Which stakeholders should be involved in the project? Who would use/provide/be affected by the new service? When to involve stakeholders?

In our ICT-based research, the need to accommodate the variety of co-production models and processes stemming from different governance contexts and application domains has been considered as an important requirement for flexibility and guidance: digital tools need to help co-producers build capacity and find the tailored path that solves their specific problem.

Governments increasingly rely on ICTs to encourage the co-production of public services by their administrations and citizens [53]. “ICT-enabled co-production” is the term used to refer to the use of ICTs to support engagement in the co-production of public services [16]. The role of ICT matters in both instrumental and institutional terms [53]: on the one hand technology facilitates practices of co-production; on the other hand, it transforms these practices into more social and ‘playful’ interactions. Digital technology has the potential to facilitate new connections within the community, enabling the establishment of relationships that were previously hindered by geographical dispersion of users. Moreover, it can empower participants by fostering shared sovereignty and responsibilities in the co-design of services [61]. The impact and role of ICT also depends on the policy domain, the institutional situation, and the existence of citizen communities. Lember et al. [47] propose an analytical framework to account for the conditions in which digital technologies impact on various elements of co-production, starting from the assumption that four elements can be identified in co-production: establishing interaction, motivating the people involved, bringing resources, and shared decision-making. A literature review on ICT-enabled co-production [87] suggests that ICTs are often used by PAs with the objective to improve effectiveness, efficiency, engagement, and inclusiveness in co-production. In fact, ICTs may improve the understanding of public needs due to the possibility of more easily collecting and exploiting large amounts of input from citizens (effectiveness). They also help reduce PAs' costs for service design and delivery by scaling-up networks of public involvement and taking advantage of citizens' and other stakeholders' time, knowledge, and expertise (efficiency). Furthermore, they allow democratic participation and interaction (engagement and inclusiveness). Yuan also notices that, when combined with the Internet of Things technology, ICTs also offer new ways to collect and process large amounts of data useful for delivering public services, both with the active collaboration of citizens or in an automatic way [87]. However, authors also suggest that it is hard to predict whether ICT will positively impact co-production and that the associated effects should be studied considering the different groups of actors involved as well as the type of digital technology considered [47]. As a matter of fact, the quality of citizens' contributions, the digital divide, the governments' commitment, and the capabilities to integrate citizens' contributions, might influence the effect of ICT-based co-production [87].

Studies also identified several challenges and barriers that prevent PAs, companies, and citizens from fully exploiting the potential of ICT to co-produce and co-deliver services. Clifton et al. [16] identified barriers such as the shortage of financial resources required to set-up and maintain an ICT-enabled co-production initiative, and the lack of a skilled workforce able to coordinate and execute related activities. Additionally, the lack of planning for the day-to-day ICT use may also lead initiatives to fail. Other barriers are related to the complexity of legal regulations that can prevent governments from adopting ICTs and can cause delays in the deployment of solutions. Finally, the lack of technical literacy along with cultural barriers - such as a negative attitude toward ICTs - could lead to a scarce participation of both citizens and government staff: while government staff may perceive ICTs as potentially controlling and intrusive, introducing too much rigidity to the organization, citizens might fear that ICTs could invade their privacy. Other barriers identified by Sorrentino et al. [76] are related to the often informal and experimental nature of most co-production initiatives, and to “the difficult scaling up and dissemination of the experiences”.

Brandsen et al. [6] elaborated an interesting comparative analysis of eight recent digital platforms supporting co-production. What emerged from the study is that often the devised digital solutions are tailored to support specific phases of co-production, either co-design or co-delivery, for example to help users create groups to unite people around specific themes, projects, or cases and let public organizations actively search for local residents and experts who can contribute to policy-making processes (co-design) [81]; or to offer citizens and companies a smart infrastructure to carry out and fund bottom-up services and activities in their local neighborhood, also leveraging open public data and energy management (co-delivery) [77]. Digital platforms spanning across different co-production phases do exist but are usually customized to work in specific application domains, like participatory urban planning [41] or collaborative information provision and refinement about public services [62]. The concept of providing flexible digital support throughout the entire co-production process, for diverse governance contexts, and for diverse application domains, is still underexplored.

In the research presented in this paper, we explain how the challenges and barriers to ICT-enabled co-production emerged from previous research have been translated into requirements for digital functionalities and properties to guide the design of a collaborative environment operationalizing co-production processes (in reply to research question RQ1). To contribute to filling current gaps in understanding how ICT impacts co-production, an iterative design-develop-evaluate approach was adopted involving PAs not only as initial informants, but also as essential evaluators and feedback providers of the implemented solutions through two evaluation iterations (to provide input for research question RQ2).

As principles related to collaborative governance models are essential to understand how technology should be designed to support co-production practices and cross-organizational collaboration, a systematic literature review was conducted on the influential factors and methods of citizen engagement within co-production processes, which attempts to summarize and combine existing research on the topic of digital and non-digital co-production and identify both findings and gaps [7]. Further study work was focused on the challenges and opportunities for ICT in co-production [16, 38, 47, 48, 53, 61, 76, 87]. The findings that emerged from the literature were analyzed and translated into high-level requirements for the Collaborative Environment, as follows:

R1. Raise awareness and build capacity in co-production. PAs who are new to co-production might have difficulties in understanding if a collaborative approach is suitable for them. This is related to the novelty that participatory approaches represent for PAs, to their scarce experience, and to the lack of good practices and success stories that help build capacity [49]. The CE should provide orientation information to help PAs better frame their context and understand how co-production can help.

R2. Overcome cultural barriers by providing clear added value for co-production. Negative perceptions and fears on the part of government staff about ICTs may significantly undermine their active participation in ICT-enabled co-production [16]. Flexible and adaptable solutions are required that do not introduce additional bureaucracy to processes. New digital tools should provide information and guidance closely related to existing work practices and to the specific co-production task users need to undertake, so that they can clearly perceive the added value with respect to general-purpose tools they already use [49].

R3. Guide users in managing the process. There may be different co-production paths according to different government models, stakeholders involved, and types of services [9, 11, 19, 1, 55]. To avoid PAs getting lost in the process, the CE should provide effective step-by-step guidance by suggesting actions to perform and appropriate resources to use to help complete such tasks, even though flexibility should be guaranteed.

R4. Support the management of the network. The lack of proper management and leadership of the collaboration network, and the fact that nobody “owns” the outcome results might lead to low commitment to the project [64]. This is particularly amplified by the specific inter-organizational and intra-organization specificities of public bodies and the complexities of institutional relationships [49]. A collaborative venture needs a reputable stakeholder to play the role of initiator and facilitator of the collaboration, i.e., the "convener" of the collaborative network [80, 45]. The CE should guide coordinators (or conveners) in understanding which skills are needed in the different phases of the process, defining which the related roles and responsibilities are and making them transparent to other members, supporting the network of stakeholders in sharing knowledge, open discussion and decision making. A clear coordinator role should be defined, with access to advanced functionalities that are hidden from other participants.

R5. Offer a catalogue of reusable resources to lower the financial impact of ICT use. Shortage of finance is a common barrier to a government's promotion of ICT-enabled co-production and financial impediments to the deployment of ICTs in co-production are reported as lower when governments or public agencies opt to use low-cost ICTs [16]. This approach implies the risk of fragmenting the digital support to co-production into many different tools, thus increasing complexity and conflicting with requirement R2 above. As an alternative approach, the CE should experiment with a centralization of resources, so that PAs dealing with co-production can find in a unique reference place reusable resources to ease their task. Digital functionalities should be made available within the platform to support collaboration dynamics. At the same time, a selected number of reusable software components should be made available to lower the implementation costs of the targeted public services. Guidance must be provided on sustainability issues, and on how to select the most appropriate third-party ICTs considering different aspects, like maintenance costs and regulatory framework.

R6. Contribute to technical capacity with knowledge resources. The lack of a skilled workforce constitutes a barrier to co-production. Furthermore, the lack of planning for the day-to-day ICT use may lead to the failure of ICT-enabled initiatives [16]. The CE should provide information on how ICT can be exploited throughout the whole co-production process (e.g., practical recommendations, best practices, risks related to potential technical failures, etc.). Furthermore, the CE should incorporate targeted knowledge resources to meet the specific requirements of the PAs that are opening their processes of public service design to relevant stakeholders [56, 83]. The objective is to operationalize knowledge by combining co-production know-how with a digital platform that guides and supports its use [79].

R7. Enforce transparency and ensure privacy. A lack of trust in the government tends to reduce the participation of citizens in ICT-enabled co-production. Additionally, citizens could fear that ICTs may invade their privacy [16]. The CE should ensure the transparency of collaborative processes and deploy privacy-preserving solutions that are clearly communicated to users.

R8. Provide legal framework. Another barrier experienced by governments is the complexity of legal issues that can prevent them to take-up ICTs [16]. For instance, the deployment of ICT solutions in co-production faces significant delays due to complex challenges related to legal regulations concerning data protection and privacy. The digital platform should support privacy by design [15] and offer guidance on how reusable software may be deployed and managed in compliance with norms and regulations.

R9. Usability of solutions and training. Lack of technical skills, along with a negative attitude toward ICTs tends to reduce the participation of citizens in ICT-enabled co-production [16]. More in general, the adoption of a new technology is affected by many factors. Usability is an important factor, but other factors play a crucial role: accuracy, price, physical appearance, security, function, interoperability, and robustness are all independent factors affecting user acceptance [42]. The CE should promote bottom-up processes and ICT training as measures to encourage stakeholders’ participation. Besides, the CE should ensure high usability and acceptability through a design process that iteratively evaluates the developed solution with end-users and incorporates their feedback.

To ensure the design of the CE was user-centered, a scenario-based design approach [14, 68], enriched with the use of personas [17], was adopted to foster an active participation of stakeholders and end-users in the definition of requirements. Scenario-based design consists in presenting and discussing stories that represent a specific problem or technology in use with different purposes [10], namely to identify needs and problems; to present and situate potential solutions; and to illustrate and discuss alternative solutions.

Three use cases of PAs in Europe in need of co-producing a public service in collaboration with a heterogeneous network of public-private stakeholders were involved in the requirements elicitation process. The three use cases were selected to represent a variety of intents and dynamics: e.g., central vs. local PAs; low vs. high citizens involvement; and governance models involving different shares of power and responsibility. This variety ensured that distilled requirements incorporated the flexibility inherent in the co-production process, as discussed in Sections 2.1 and 2.2 above. 

The three analyzed use cases highlight the potentially large heterogeneity of co-production processes that may differ by objectives of PAs in relation to co-production; actors involved; level of participation of citizens or associations of citizens; type of service to be co-produced; and resources needed to carry out the co-production process.

During an initial focus group, 15 PA representatives (10 for MEF, 3 for VARAM, and 2 for ZGZ) were invited to reflect on their co-production objective and on the social and organizational context in which co-production would take place. A two-page template was distributed to collect information on the public service to be co-produced; on the desired future scenario in which the service would be delivered; key actors and roles; related initiatives; available resources, desired additional resources and desired digital support to sustain the co-production of the public service.

After this preparatory activity, use case coordinators were guided through the generation of personas (6 for MEF, 6 for VARAM, 4 for ZGZ, taking into account different actors engaged in co-production) and the description of their relationships. Personas, i.e., fictitious profiles of service end-users and stakeholders [69] that act as a benchmark for design, encouraged teams to reflect on the actors involved in the service to be co-produced, i.e., on the potential co-producers. The generated personas were used to describe three scenarios of co-production, i.e., the process by which a network of stakeholders collaborates to create and deliver a new public service. PA representatives were invited to imagine co-production needs that could be solved by means of digital technology and reusable knowledge resources. Table 2 summarizes the requirements that PAs mentioned as useful in their use case. As it can be observed, some of the requirements refer to functionalities that could be integrated directly inside the Collaborative Environment, such as tools for workplan scheduling or tools for managing incentives and rewards to sustain participation (U1, U9). Other requirements call for software modules to be used to support specific tasks, such as an editor to collaboratively revise texts appearing in web pages (U8, U11–U14). A third class of requirements refers to reusable knowledge (guidelines and template documents) to bootstrap activities (U2–U7, U10). 

Reusable knowledge and reusable software may vary significantly across use cases and a rich and extendable catalogue of such resources, which we call INTERLINKERs, should be made available.

Janowski [39] argues that the Digital Government concept has evolved across four main stages: Digitization (Technology in Government), Transformation (Electronic Government), Engagement (Electronic Governance) and Contextualization (Policy-Driven Electronic Governance). ICT-based co-production falls within the category of technology that involves the engagement of a plurality of public and private actors in the generation of solutions and the co-creation of public value, thus enabling government to make a transformation to a more open and participatory model that they see as more desirable [54]. Therefore, ICT-based co-production can be seen as spanning across Electronic Government and Electronic Governance, implying a transformation of the internal working and structures of government as well as its relationships with citizens, businesses, and other stakeholders [39]. Previous studies and projects on Digital Government let emerge particular requirements that must be taken into account, not only during the implementation phase but also at the early design or architecture modeling phases. Technical requirements refer to all those features that the platform should implement and the constraints it should satisfy to be interoperable and compliant with national and international regulations. Table 3 summarizes a set of common characteristics of Digital Government architectures that have emerged from a literature review by Baheer et al. [2].  Most of these requirements are also explicitly endorsed by national and international regulations and recommendations, such as for example the EU eGovernment Action Plan that provides guidance for all countries in the European Union [25]. Specific requirements address the issue of ensuring that the platform can be easily adapted and used by different PAs and with a cross-sector and cross-border dimension.

These technological requirements guided the baseline architectural choices of the platform and the development process. For example, to put in practice technological neutrality and platform independence, a web-solution was preferred (T15, T16), with the possibility for the platform to work both in single-tenant configuration (i.e., with one instance of the platform for each PA) and in multiple-tenant configuration (i.e., with the same server safely managing different PAs), to help reduce operation costs to a minimum (T14). An authentication and authorization service implementing the Single Sign-On protocol¹ was put in place (T8), supporting European identification standards, thus also complying with cross-border characteristics (T11). To comply with the requirement of traceability and accountability of actions, methods for logging user interactions and analyzing behavior were conceived (T9). Multilinguality of the user interfaces and the possibility of customizing them according to PAs' needs was included in accordance with the adaptability and flexibility requirements (T3, T4). As explained in Section 5, an extendable catalogue of reusable resources facilitating co-production was envisaged to cater for the requirements of reusability, resilience, and scalability (T5, T6, T12). To ensure interoperability and integrability, the formalization of reusable software resources was inspired by the Digital Europe Building Blocks methodology [28] (T1, T7). FAIR principles and the General Data Protection Regulation (GDPR) constraints for data and metadata management [86, 30, 75] were percolated through the system architecture to comply with the security and legality requirements (T2, T13). The user-centered design approach adopted as a baseline methodology for the overall project, ensured citizen-centric and usability requirements right from the start (T10, T19). Tutorials and user manuals addressed the comprehensibility and minimal learning curve for all potential users of the platform (T17, T18).

The following section describes the first version of the Collaborative Environment that was designed and implemented according to the collected multiple types of requirements, namely, top-down governance requirements, bottom-up use case requirements and technical requirements.

Public or private entities wishing to coordinate a new initiative for co-producing a public service might have difficulties in understanding how to structure and manage the collaboration process. For this reason, a special “Overview” area is available to process coordinators with a checklist on how to proceed, as advocated by governance requirements (R1–R3).

Coordinators are instructed (i) to create a title and goal description for their process; (ii) to assign administrators to the newly configured process; (iii) to select the schema that best describes the collaborative process matching their co-production goals; and (iv) to invite teams of other participants to join the process. Step (iii) in the sequence is particularly important, as it offers coordinators examples of co-production process models, which are called “schemas” in the platform. These schemas are represented in the form of suggested sequences of phases for the co-creation of a public service (e.g., Engage, Design, Deploy and Sustain), which can expand into more detailed objectives (e.g., Engage expands into Identify Stakeholders, and Engage Stakeholders), and tasks to perform to reach each objective (e.g., “Understand the different types of stakeholders” or “Map stakeholders, analyze motivation, skills, expectations”). Figure 3 illustrates how co-production schemas are graphically represented in the user interface. The devised Collaborative Environment offers alternative co-production schemas to choose from, with the possibility of additionally customizing templates by removing steps (i.e., phases, objectives, or tasks) that are not relevant for a specific use case, or by adding new ones (in compliance with the flexibility advocated by requirement R3). Figure 3 and Figure 4 illustrate two examples of different co-production schemas, one with default steps for the co-design and co-delivery of a public service where citizens are involved in all phases (Figure 3), and one with suggested steps to foster the ideation of social innovation through hackathons involving citizens (Figure 4). The set of predefined co-production schemas offered by the platform was identified through governance research [7] and through the analysis of co-production use cases [48, 18].

Once the coordinator has chosen the co-production schema and has possibly further customized it, the “Guide” view helps the network of stakeholders orientate within the process by visualizing the suggested list of tasks and what is required to accomplish them (requirement R3). For each task, a work area is available where the co-production team can share resources, like documents and links (under the “Resources” tab on the right part of the interface in Figure 3). A button allows users to visualize a list of recommended knowledge and/or software enablers (called INTERLINKERs) that could help to complete the task (see the button “Instantiate task resource through recommended INTERLINKERs” in Figure 3). For example, for the task “Ideas collection for service definition” that happens during the Design phase, a bespoke knowledge resource is suggested containing guidelines on how to organize focus groups. The public resource can be reused by cloning and editing it for the purposes of the specific co-production project and can then be saved in the task workspace.

A special category of generic software INTERLIKERs is also available offering baseline collaboration functionalities, such as the creation and sharing of documents (via integration with Google Drive APIs²); the creation of surveys; the creation of discussion forums (via integration with Loomio³); and the sharing of external links (see the drop-down menu shown in the right part of the interface in Figure 4).

The catalogue of INTERLINKERs, defined as reusable resources offered to build capacity in co-production according to governance requirements R5, R6 and R8, represents one of the core concepts that characterize INTERLINK's CE. INTERLINKERs can be used for different purposes:

Overall, 60 INTERLINKERs were uploaded in the Catalogue integrated in the first version of the CE. However, the approach used to ground the design and the development of the INTERLINKERs catalogue was purposefully aimed at ensuring future flexibility and extensibility of the catalogue with new enablers of different types (technical requirements T5, T6, T12). In fact, the set of INTERLINKERs resulting from the second development iteration of the platform, after the evaluation described in Section 6, counted more than 70 enablers.