Open innovation supply chain configuration and performance through a proximity lens Open Access

In the contemporary business landscape, organizations struggle to operate in turbulent environments marked by evolving market requirements (Zimmermann et al., 2020). Faced with the need to adapt to competitive markets, companies must deliver innovative products and services. This awareness drives organizations to embrace supply chain collaboration and open innovation strategies (Solaimani and van der Veen, 2022), both of which foster collaboration with external actors to remain competitive (Solaimani and van der Veen, 2022; Bogers et al., 2018).

In addition to collaboration with traditional supply chain partners, such as customers and suppliers, open innovation also involves collaboration with expert organizations outside the supply chain, such as universities (Solaimani and van der Veen, 2022; Gibson et al., 2016). Supply chain collaboration can drive new business opportunities, particularly in technology-intensive industries (Patrucco et al., 2022). Open innovation offers benefits such as knowledge acquisition, problem-solving (Bishop et al., 2011), organizational learning and new business opportunities (Johnston, 2022). Given the strong similarities between the concepts, collaborative innovation will henceforth be included in open innovation. Open innovation aims to leverage external ideas to complement an organization’s internal efforts (Zimmermann et al., 2016) through “purposive inflows and outflows of knowledge” (Chesbrough et al., 2006, p.1).

Concretely, appropriate partners must be engaged, which is referred to as external knowledge sourcing, which was highlighted as a research need by Brunswicker and Vanhaverbeke (2015). Inspired by Zhang et al. (2024), we position external knowledge sourcing within the configuration phase. After appropriate partners are engaged, Shamsuzzoha and Helo (2018) describe a product development lifecycle as consisting of a conceptual design and engineering phase where the product and its architecture are specified. Due to their similarity, design and engineering are treated as a single phase. Hence, this paper addresses open innovation during the configuration and design and engineering phases. In open innovation settings, managing supply chain interfaces is still under-researched (Patrucco et al., 2022; Solaimani and van der Veen, 2022). Since supply chains have multiple interfaces for collaboration, and open innovation increasingly involves multiple suppliers, Patrucco et al. (2022) have called for studies beyond dyadic scopes. This study focuses on an extended supply chain consisting of several actors, from sourcing at suppliers to the ultimate customer, in line with Spekman et al. (1998). To emphasize our focus on the supply chain itself, we label it an open innovation supply chain (OISC), in which various actors collaborate to innovate before regular products, orders and deliveries exist.

Open innovation is understood to be facilitated by proximities between actors (Johnston, 2022; Zahoor and AL-Tabbaa, 2020). As proximities, which are used to assess the similarities between actors in open innovation (Johnston, 2022), are generally understood to affect knowledge sharing and uptake (Knoben and Oerlemans, 2006), this study applies the concept of proximities as a lens to gain a deeper understanding of the OISC. Since proximities can affect OISC configuration (Johnston, 2022), and external knowledge sourcing is a key aspect of configuration, we contribute to the ongoing discussion on how knowledge is sourced (Brunswicker and Vanhaverbeke, 2015). The first research question is formulated:

OISC configuration sets the stage for collaborative capabilities in the OISC, which can be observed during the design and engineering phase. Building on prior research (e.g., Oyedijo et al., 2024; Johnston, 2022), proximities can be studied in the design and engineering phase to better understand how they relate to OISC performance. The limited understanding of the antecedents, such as collaborative capabilities (Solaimani and van der Veen, 2022) of successful open innovation, was highlighted by Zhang et al. (2024) and Zahoor and AL-Tabbaa (2020). This gap becomes even more apparent given that potential barriers to open innovation in supply chains remain unexplored, and a deeper understanding of how a lack of collaborative capabilities negatively affects the OISC and its performance is lacking (Anderson et al., 2023; Shamsuzzoha and Helo, 2018). The second research question is formulated:

The purpose of this paper is to deepen understanding of how proximities, knowledge sourcing and collaborative capabilities relate to OISC configuration and performance.

A complex OISC is investigated in a longitudinal, in-depth study, using observations and interviews for data collection. An important aspect of the study is the shift of focus from innovation to the supply chain through which the innovation is realized. Collective OISC performance is captured through delays. The findings point to the importance of social, organizational and technological proximity for OISC performance, showing how a lack of proximity in one dimension can hinder the development of proximity in another. The findings highlight the importance of individuals as bearers of social proximity and the risks that non-purposive knowledge sourcing and OISC reconfiguration pose to collaborative capability. The contributions offer potential for further research and managerial implications for OISC.

The remainder of the paper is organized as follows. A frame of reference is presented, followed by the methodology. Section 4 presents the empirical findings and results related to the two RQs, while Section 5 includes conclusions, implications and directions for further research.

This section builds on the identified gaps in understanding OISC configuration and performance. It introduces three constructs that are considered to particularly relate to OISC configuration and performance: various dimensions of proximity as explored in subsection 2.1, and knowledge sourcing and collaborative capabilities as addressed in 2.2. Subsection 2.3 presents how to capture OISC performance, while subsection 2.4 presents the analytical framework.

Extant research identifies proximities—relational or cognitive similarities between actors—as factors affecting collaboration and innovation performance (Johnston, 2022; Jespersen et al., 2018; Cassi and Plunket, 2014). Studies have explored proximities in various contexts, such as between universities and SMEs (Johnston, 2022), among SMEs (Jespersen et al., 2018), and among universities (Fernández et al., 2021). The starting point is Knoben and Oerlemans’ (2006) structured literature review, which identifies geographical, organizational and technological proximity as central proximity dimensions, where organizational proximity is understood to include, for example, social and institutional proximity. However, this paper treats social proximity separately, as it focuses on personal relationships rather than organizational similarities (in line with Boschma, 2005; Johnston, 2022; Lauvås and Steinmo, 2021) and based on its potential effects on actor inclusion in the OISC during configuration (see, e.g. Johnston, 2022). Therefore, the four proximity dimensions discussed here are geographical, social, organizational and technological proximity.

Proximities exist on a spectrum rather than as binary constructs. While a high degree of proximity among collaborating actors may prevent misunderstanding, suspicion and conflict, it may also hinder creativity and innovation, as sources of novelty may be lacking. Too little proximity may lead to opportunism (Boschma, 2005). A balanced diversity of actors (Solaimani and van der Veen, 2022) or a suitable degree of proximity (Boschma, 2005) is favorable for learning and innovation. Proximities can be present at the start of a collaboration but may also evolve (Balland et al., 2015; Johnston, 2022), with actors adopting similar approaches to the collaboration.

Some authors see geographical proximity, focusing on distances in which OISC actors operate, as instrumental in fostering collaboration (Sjöö and Hellström, 2019; Bishop et al., 2011). It facilitates face-to-face meetings, enhances visibility and allows robust evaluations of partners’ effectiveness (Sjöö and Hellström, 2019), thereby cultivating stronger cooperative ties (Cassi and Plunket, 2014). Co-location in a research center setting, enabling frequent interactions, is identified as a prerequisite for frontier research in excellence centers with long-term funding for advanced interdisciplinary innovations (Hellström et al., 2018). Bishop et al. (2011) emphasize geographical proximity’s role in joint problem-solving and in transferring tacit and context-specific knowledge. However, Boschma (2005) and Cassi and Plunket (2014) conclude that geographical proximity is neither necessary nor sufficient for learning and innovation, though it does facilitate interactive learning and strengthen other proximity dimensions. Johnston (2022) found that it had no significant impact on either the formation or function of open-innovation collaboration. Geographical proximity can be captured by the relative distance based on transportation means or the actors’ perceptions of these distances (Knoben and Oerlemans, 2006).

Social proximity is created by connections through friendships, shared networks and professional as well as private interactions (Boschma, 2005; Johnston, 2022). Prior interactions can develop a sense of familiarity that encourages collaboration and stimulates interactive learning (Boschma, 2005), and are a strong indicator of future or continued collaboration (Sjöö and Hellström, 2019; Lauvås and Steinmo, 2021; Balland et al., 2015). Social proximity plays an important role in forming collaborations, as it facilitates partner selection by enabling the assessment of actor capabilities, their potential contribution and compatibility with other actors, because their knowledge, skills and approach to work are known (Sjöö and Hellström, 2019; Balland et al., 2015). Initiating collaborations with partners outside of traditional networks exposes companies to high levels of relationship uncertainties, and Melander and Pazirandeh (2019) suggest conducting extra careful qualification processes. Trust is crucial for value co-creation and innovation, as it enables resource sharing and combining (Fawcett et al., 2017; Zimmermann et al., 2016). It is reflected through confidence in the partner, indicating a readiness to refrain from opportunistic behavior or the belief that the partner will consistently act as promised (Spekman et al., 1998). Specifically, Fawcett et al. (2017) emphasize that in a supply chain setting, such as in an OISC, trust comprises credibility (or performance capability) and relationship commitment.

Social proximity can be seen as a measure of relationship strength (Jespersen et al., 2018), and empirically, it is often measured by previous interactions in research or collaborative projects (Fernández et al., 2021). It can emerge from direct interactions, such as collaboration and cooperation (Spekman et al., 1998), as well as from arm’s-length relationships or indirect connections (Balland et al., 2015). Project failures or unmet goals can lead to decreased proximity between actors (Balland et al., 2015). AL-Tabbaa and Ankrah (2016) demonstrate that previous interactions do not play a significant role in initiating state-driven collaborations; instead, a strong sponsor, in their case, a government body, replaced previous relationships.

Organizational proximity refers to the level of similarity in working culture, methods and practices or routines (Johnston and Huggins, 2021; Knoben and Oerlemans, 2006) in supplier-supplier and/or supplier-customer relationships (Dallasega and Sarkis, 2018). Little organizational proximity is expected between universities and small companies (Johnston and Huggins, 2021), particularly in terms of flexibility, time perspective and practices (Gibson et al., 2016). Universities often handle certain practices, such as initiating OISC, while companies handle other activities, such as planning (Goel et al., 2017). Routines act as unique bundles of capabilities that rely on regularity for efficiency. Familiarity with these routines and the establishment of common concepts and shared language enhance effective communication, mitigating interpretation discrepancies and supporting smoother collaboration and knowledge exchange (Östbring et al., 2017; AL-Tabbaa and Ankrah, 2016; Balland et al., 2015).

Sharing relevant, accurate and timely information, such as resource availability and operational innovation, facilitates effective collaboration and innovation (Nguyen et al., 2019; Zimmermann et al., 2016). The communication structure in the supply chain can affect both information exchange and communication effectiveness (Mattsson, 2012), which is a central aspect to consider. In a parallel communication structure, the different functions of the supply chain actors and partners communicate directly with the potential of fast and function-specific understanding. In a sequential structure, one person represents the respective actor and partner. This person indirectly forwards information to the respective functions, with the risk of slower and more general understanding, however, with complete control over all information exchange. Similarly, Balland et al. (2015) noted that collaborations in which tasks are centrally organized by a coordinator are less likely to increase the level of proximity between actors. Common understandings shape the collaborative process and facilitate harmonizing interests or motivations and cultivating shared visions (AL-Tabbaa and Ankrah, 2016; Johnston, 2022). Organizational proximity also relates to shared timeframes (Johnston, 2022), and differing perspectives on time can be present between actors. According to Johnston (2022), organizational proximity is pivotal in the configuration and design and engineering of collaborative relationships.

Technological proximity, reflecting similarities in knowledge, expertise and know–how related to specific processes or tools (i.e. knowledge base), can positively influence collaborative relationships by facilitating knowledge exchange and mutual understanding (Johnston, 2022; Balland et al., 2015). A higher degree of proximity in knowledge bases can lead to more detailed discussions about goals, purposes and execution, enabling more precise specifications of how partners can contribute (Johnston, 2022). However, technological proximity does not necessarily imply homogeneity in knowledge bases. Instead, a balance between having a comparable knowledge base (proximity) to recognize collaboration and innovation opportunities and maintaining a diversified knowledge base (distance) fosters innovation and facilitates knowledge absorption (Fernández et al., 2021; Jespersen et al., 2018). Technological proximity can increase as actors learn from each other through interaction, exchange and joint knowledge production. A receiving actor’s knowledge base can also expand and align more closely with that of other actors (Balland et al., 2015), leading to fewer challenges in comprehending the words and actions of their partners in later phases of the collaboration (Lauvås and Steinmo, 2021). Technological proximity has been shown to be crucial in open-innovation collaboration, as it facilitates small companies’ understanding of the potential of university partners during partner selection and supports the uptake and utilization of university knowledge during the collaboration (Johnston, 2022). Technological and organizational proximity complement one another, with organizational proximity facilitating understanding between actors (Johnston, 2022; Östbring et al., 2017). This alignment influences effective knowledge transfer through the working methods and motivations of their partners (how they interact, i.e. organizational proximity) and the similarities in the knowledge bases (what they exchange and the potential value of those exchanges, i.e. technological proximity).

In addition to the incremental innovation often identified in the supply chain collaboration literature, open innovation literature highlights collaboration with actors outside the supply chain, such as universities, which typically focus on radical innovation (Solaimani and van der Veen, 2022). Various external knowledge-sourcing strategies for open innovation are also available (Brunswicker and Vanhaverbeke, 2015). They can be labeled as follows: supply chain sourcing, which involves knowledge sourcing from direct customers or suppliers; technology-oriented sourcing, such as sourcing from universities to access new technologies early (see also Gibson et al., 2016); application-oriented sourcing, in which knowledge is sourced from customers and users to understand the expected application, and full-scope sourcing, which involves sourcing from all types of actors. The latter two are most clearly linked to successful innovations. While external knowledge sourcing can be a purposive choice, it may be hindered by factors such as actors’ limited size and lack of resources (Brunswicker and Vanhaverbeke, 2015). Gibson et al. (2016) provide other examples of non-purposive sourcing, showing how funding “regimes” can unintentionally hinder university-industry collaboration by shaping partner selection.

Solaimani and van der Veen (2022), one of the few identified studies on open innovation in supply chain literature, highlight three key collaborative capabilities that can foster successful innovation during configuration and design and engineering. The first capability, ambidextrous collaboration purpose, involves the simultaneous collective exploitation of knowledge and proactive exploration of new opportunities to expand the knowledge base, building on shared visions for the future. Exploitation focuses on leveraging existing skills and resources, while exploration involves seeking external resources through partnerships to create innovation. This is often perceived as paradoxical, as they compete for the same resources (Iftikhar et al., 2024). According to Solaimani and van der Veen (2022), commitment to co-innovation can be achieved through equal collaboration among actors or through more centralized leadership by a key actor. The second capability, ambidextrous collaboration span, encompasses collaboration with supply chain partners as well as actors not usually involved in the supply chain, thus overlapping conceptually with knowledge sourcing previously discussed. For successful innovation, the group of actors must be large and diverse enough to encourage varied perspectives yet small and cohesive enough to foster a sense of shared responsibility and minimize communication challenges. Any power asymmetries between the actors should be used to facilitate and coordinate them in pursuing innovation (Solaimani and van der Veen, 2022). Finally, ambidextrous collaboration orientation emphasizes the pursuit of both incremental improvements and radical innovation through shifts in products, processes, markets or business models (Solaimani and van der Veen, 2022). Nguyen et al. (2019) showed that information sharing and decision synchronization were critical for achieving radical innovation, whereas incentive alignment supported incremental innovation. Continuous engagement, experimentation and willingness to learn, supported by a culture and mindset that foster these activities, are important.

Innovation performance is presented under different labels in the literature. It is captured as the number of patent applications (e.g. Zhang et al., 2024). Similarly, Brunswicker and Vanhaverbeke (2015) captured innovation success by the successful launch of innovation projects and innovation performance as financial income from new products and services. Such approaches require ex post studies, since it can take years after the innovation work has concluded for revenue from new products to be realized. A firm’s successful innovation performance was further discussed in terms of reduced costs, shortened time to market and improved product design (Patrucco et al., 2022; Shamsuzzoha and Helo, 2018). Zimmermann et al. (2016) discussed long-lasting results from open innovation, such as early problem identification and reduced rework. Solaimani and van der Veen (2022) captured successful innovation as effective, broad, fast or synergetic without digging deeper into those concepts. Our study, however, examines innovation from a supply chain perspective. Therefore, it is important to proactively understand OISC performance during an ongoing product development lifecycle. Furthermore, both Zhang et al. (2024) and Brunswicker and Vanhaverbeke (2015) address an individual firm’s innovation performance. Although the studies acknowledge the inclusion of customers and suppliers, they do not address collective performance. Combining the concepts of effectiveness and speed with time to market (Patrucco et al., 2022), OISC performance naturally translates into aspects of timing relative to plan, pragmatically captured as delays. Capturing delays is also suitable when studying ongoing collective OISC performance, which is why it is applied in this study.

The constructs of the study are related to the research questions and to each other in Figure 1. Linking knowledge sourcing to configuration follows Zhang et al. (2024). OISC configuration sets the stage for collaborative capabilities, which are essential, and first become observable once initial knowledge sourcing is complete. This makes collaborative capabilities central during design and engineering. The link between proximities and both phases draws on sources such as Johnston (2022).

This section presents the research design in 3.1, sampling and data collection in 3.2 and analysis and coding approaches in 3.3. Research quality and rigor are discussed throughout the section.

An externally funded project for developing an advanced technological innovation—a prototype for a superconducting magnet, intended for application in research facilities, was studied from idea to finished prototype. As no single actor could achieve this alone, the combination of competencies and resources from multiple actors was necessary, making this an open innovation project. A qualitative research approach is suitable when the purpose is to gain understanding (Bryman and Bell, 2022). The overall purpose was formulated in the funding application based on practical relevance—for which support can be found in e.g. Toffel (2016)—was influenced by funding agencies, and adapted by the researchers to ensure theoretical relevance in this study. Case studies are often appropriate when understanding detailed, complex or unique phenomena is central (Yin, 2014), which is why a qualitative case study research design was adopted. Several partly overlapping, collaborative research approaches exist within case studies, such as interactive research and action research (Sandberg et al., 2022; Näslund et al., 2010). Anderson et al. (2023) suggest action research to capture supply chain dynamics in open innovation, while Sandberg et al. (2022) highlight interactive research for innovation studies. Action research involves a higher degree of researcher intervention. In contrast, the adopted research design is primarily interactive, with close collaboration at designated workshops during project meetings and more extensive face-to-face workshops throughout the innovation project (similar to what is suggested by Sandberg et al., 2022; Svensson et al., 2015). Interactive research can simultaneously contribute to research and practice (Svensson et al., 2015). Interactive research was enabled by unique access and reciprocal trust (Sandberg et al., 2022), created as the same funding agencies financed experienced researchers in the project to study collaboration during the three-year project. Therefore, interactive research criticism and skepticism could be avoided, specifically shallow participation, untrained researchers and a short-term perspective (in line with Näslund et al., 2010). Sampling this single case, typical in interactive research settings, is consequently based upon the potential to learn from a specific case, which aligns with Yin (2014). The case study is furthermore longitudinal; the same case is studied at several points, as conditions are expected to change over time (Yin, 2014). Overall, the research is considered well designed for the purpose.

The studied “project actors” include two universities, two small companies (10–50 employees) and one micro-sized company (<10 employees). Where applicable, multiple respondents from each actor participated to cover necessary competencies. Such data triangulation increases validity and rigor (Näslund et al., 2010), enables the collection of rich and varied data and enhances reliability. Informed consent was acquired from all respondents. Table 1 presents the OISC actors, provides brief descriptions of the organizations and their respective knowledge domains and lists the respondents.

Following Näslund et al. (2010), multiple data collection methods were applied, including participant observations by researchers during digital and in-person meetings and workshops. Ninety-four events were documented (in the form of protocols and reflections from the researchers), including 30 project meetings (duration 90–120 min), 56 technical meetings (typically 60 min), six digital lectures (intended to support a common understanding of the involved technologies) and two extended face-to-face workshops including visits to all project-actor sites (with lunch-to-lunch duration). The documented meetings contribute to the study’s overall high validity and reliability.

Relevant to this paper, 39 interviews were conducted during the project’s three-year duration. Semi-structured interview guides (see  Appendix) were distributed in advance. Due to the COVID-19 pandemic restrictions, most interviews were conducted digitally and by both researchers to increase reliability (Yin, 2014). The duration of the interviews was 40–75 min. Interview summaries were compiled and shared with respondents for verification. Short workshops and follow-up interviews were conducted after several meetings, lasting 15–20 min. The follow-up interviews served as “temperature checks” to gain deeper insights into participants’ lived experiences beyond the main interview rounds (see below). Project documents such as the decision on funding, collaboration agreement and meeting minutes are also part of the collected data. Method triangulation was applied to enhance validity and reliability (in line with Yin, 2014), with the researchers’ observations during meetings helping to reduce the risk of response bias. Continuous data collection enabled the collection of rich and varied data.

Interview data were collected in three main rounds of interviews (after 6, 18 and 35 months of project duration), with final interviews timed close to project closure. These rounds reflected anticipated stages or phases where changes were expected to become evident (Yin, 2014). The first round of interviews concentrated on the project’s early stages, emphasizing collaboration experiences, actor inclusion criteria, project organization, working methods, objectives and performance to date. The second round revisited these topics while introducing questions about communication and collaboration challenges, as well as knowledge sharing and learning. The third round continued to explore similar themes from the second round, with a deeper focus on performance. Most questions were retrospective in nature. All data were cross-validated by the researchers, allowing for an in-depth analysis. All data are saved in a database.

Based on the research questions, the unit of analysis is proximities in the interfaces between actors in the OISC. The unit of analysis was refined during the analysis. Throughout this study, the understanding of proximities emerged and was analyzed between different sets of actors within the supply chain. The analysis progressed from examining proximities between two actors (dyads, one-to-one), to between one actor and multiple actors (networks, one-to-many) and finally to between groups of actors within larger networks (many-to-many). As Näslund et al. (2010) observed, the unit of analysis may evolve throughout a project; therefore, addressing this issue is important when it contributes to advancing our scientific and practical knowledge.

Coding-wise, the configuration and design and engineering phases were naturally separated as data were collected longitudinally; when the OISC was configured and the project started, the design and engineering phase began. The semi-structured interview guide facilitated the coding. The well-structured proximity literature was applied during the analysis. Both researchers conducted the data analysis by color-coding the documented statements, observations and other relevant data related to the four proximity dimensions, knowledge sourcing, collaborative capabilities and performance. Any disagreements between the researchers were reconciled. Clear examples of proximities and the lack or loss of proximities, with significant effects on OISC configuration and performance, were selected for further analysis. Therefore, rather than trying to present proximities evenly across all interfaces or all possible network combinations, specific interfaces were given more focus than others. For the researchers, it was important to stay close to the literature in the frame of reference, to ensure both validity and reliability (in line with Näslund et al., 2010; Yin, 2014). All actor data were analyzed together with pattern-matching elements (Bryman and Bell, 2022).

The empirical findings and results are divided into two sections: subsection 4.1 examines the OISC configuration through a proximity lens. Subsection 4.2 similarly addresses the OISC performance, presented as bullet points after each proximity-based analysis.

Knowledge sourcing: University A (UA) initiated the project, bringing its theoretical knowledge of superconducting technologies, including electromagnetic design expertise. The fact that a university, rather than a company, initiated an OISC aligns with the findings of Goel et al. (2017). In addition to contributing its knowledge, UA also acted as a service provider for simulations and testing. The manufacturer is a well-known magnet manufacturer serving research, medical and industrial applications. UA and the manufacturer jointly applied for external funding and were key actors in the OISC. From the manufacturer’s point of view, the collaboration with UA can be considered technology-oriented sourcing aimed at getting early access to new technologies (in line with Brunswicker and Vanhaverbeke, 2015) and learning opportunities.

As more knowledge and resources were necessary to develop and manufacture the innovation, two suppliers to the manufacturer were recruited: S1, which performed advanced processing, and S2, which manufactured high-precision components. Suppliers are key enablers of open innovation from an SCM perspective, particularly in technology-intensive industries (Patrucco et al., 2022), like the one studied. The manufacturer’s knowledge sourcing from suppliers aligns with Brunswicker and Vanhaverbeke’s (2015) supply chain sourcing strategy, which involves interaction through traditional supply chain linkages. University B (UB), a supplier of mechanical design, CAD construction, testing and tooling required during manufacturing, was included in the OISC based on a request from the regional funding agency. Since the funder required the inclusion of a local university to support regional research and development through learning and knowledge retention, we argue that in this case, the knowledge sourcing was less purposive (Brunswicker and Vanhaverbeke, 2015) and influenced by factors outside the OISC (as highlighted by Gibson et al., 2016). Although not part of the externally funded project, the prospective customer contributed valuable knowledge of superconducting technologies and expertise in practical applications of superconducting magnets, sharing this knowledge through lectures, a study visit and consultation on various occasions. Therefore, interacting with the prospective customer constituted application-oriented knowledge sourcing to better understand the expected innovation application. It also represented an additional form of technology-oriented sourcing. Given the broad range of knowledge sourcing strategies, classified as full-scope sourcing, encompassing technology-, supply chain- and application-oriented strategies, the resulting configuration is likely to generate certain collaborative capabilities. This provides a strong foundation for successful innovation and OISC performance, building on the findings of Brunswicker and Vanhaverbeke (2015). The following sections analyze how proximities influenced the configuration.

Geographical proximity was not actively considered by the initiating actors (UA and the manufacturer) as a factor for inclusion in the OISC, aligning with Johnston’s (2022) findings. Although high geographical proximity prevailed between UB and the project companies, other proximities—primarily social and technological, as discussed later—along with the funding agency’s requirements (similar to AL-Tabbaa and Ankrah, 2016) affected OISC configuration.

Social proximity was a substantial contributing factor in initiating the project between UA and the manufacturer and in the inclusion of S1 and S2. Between UA and the manufacturer, it can be argued that it compensated for the lack of geographical proximity. For similar substitution effects, see, e.g. Cassi and Plunket (2014) and Johnston (2022). Based on previous collaborations with essentially the same people involved, the actors were confident in the prototype’s successful innovation and delivery.

Also, UAPL1 expresses the importance of social proximity:

The quotes also convey trust (Fawcett et al., 2017; Zimmermann et al., 2016) in the other actors, related to their capabilities and commitment to the task, thereby expressing the perception of the other actor as a good fit for the OISC. The social proximity among the project companies was founded in their prior interactions (Johnston, 2022; Boschma, 2005) and collaborations (Lauvås and Steinmo, 2021; Balland et al., 2015; Spekman et al., 1998) in customer and supplier relationships and was expressed in beliefs that the other actors would be able to perform in their respective roles. MMD expresses,

The initial contact person at UB retired after the project application was granted and the collaboration agreement was signed. Consequently, after the configuration phase, the project responsibility at UB was transferred to a second person with whom no social proximity had been established. According to Melander and Pazirandeh (2019), such situations require extra care in qualification. The funding agency’s inclusion of UB without social proximity to other actors is similar to AL-Tabbaa and Ankrah’s (2016) findings of a government body superseding prior interactions.

Organizational proximity can be seen as a “silent” inclusion criterion that did not receive much attention in contrast to findings by Johnston (2022) and Cassi and Plunket (2014):

Based on the social proximity between UA and the companies, the assumption was that each actor would share a working culture and methods (Johnston and Huggins, 2021; Dallasega and Sarkis, 2018; Östbring et al., 2017), enabling them to perform as expected.

The collaboration agreement, signed by all actors, was perceived as an important way of formalizing the project and as a “statement of intent” for the actors to follow the directed pathway. Therefore, for UB as well:

Based on the literature, significant differences in working culture and methods can be expected between universities and small companies (Johnston and Huggins, 2021; Knoben and Oerlemans, 2006). The companies expressed concerns regarding potential university bureaucracy, as well as UB taking over parts initially planned for the manufacturer:

However, this did not affect the OISC configuration. Positive expectations were also conveyed:

Technological proximity, reflected in complementary rather than overlapping knowledge domains, competencies and technologies, was actively considered during OISC configuration. A certain degree of distance was preferred over a high degree of proximity (in line with Fernández et al., 2021; Jespersen et al., 2018). The following quote illustrates the interdependence between the actors:

The universities’ much more active service supplier role in the OISC than customary in research projects (following Johnston, 2022) stems from their unique expertise in domains such as testing (in line with Figueiredo and Fernandes, 2020) and design. Such complementary knowledge facilitates interdisciplinary research and development (Hellström et al., 2018). Equally important, the level of competencies, skills and access to advanced production technologies were critical inclusion criteria at this stage. Consistent with Patrucco et al. (2022), the importance of multiple suppliers’ expertise was emphasized. UAPL1 highlights the competencies and resources of S1 and S2:

The small companies stand out with unusually high education levels—S1 and S2 are technology-intensive firms, and the manufacturer is a university spin-off—which enhances technological proximity to the universities (Johnston, 2022). Like the manufacturer, S2MD had actively participated in earlier university collaborations. A high level of competence and familiarity with research institutions is important, as this can provide a foundation for the necessary technological proximity to develop.

Collaborative capabilities: Three years of collaboration, trials, errors and numerous iterations of material and information flows across numerous supply chain interfaces were still insufficient to move from idea to finished prototype. The project was delayed by several months—first through an extended deadline, then by some actors deciding to complete the prototype outside the project’s timeframe and budget. Given the project’s strong prerequisites, including an innovative idea, access to state-of-the-art expertise (demonstrated through supply chain, technological and application knowledge sourcing) and long-term project funding, this creates an opportunity to explore how proximities and changes in collaborative capabilities may have contributed to the delays during design and engineering. An ambidextrous collaboration purpose (Solaimani and van der Veen, 2022) was reflected in the actors’ joint commitment to explore new knowledge and exploit the existing knowledge base. This commitment was grounded in a shared vision and equal collaboration. Aspects of centralized leadership were expected of UA, such as coordinating and advancing the innovation process. In terms of collaboration span, working with supply chain partners and other external actors (universities) was seen as essential to complete the innovation. To encourage diverse perspectives and achieve a “balanced diversity” of actors (Solaimani and van der Veen, 2022), the actor span was considered sufficient for learning and innovation. Regarding ambidextrous collaboration orientation, the externally funded project enabled these small companies to engage in both incremental and radical innovation—an opportunity that would otherwise have been out of reach.

From the manufacturer’s perspective, the innovation can be considered radical. Its first product combining new processes and knowledge (mainly from UA and the prospective customer) with existing capabilities, aimed at attracting new customers in new markets (following Solaimani and van der Veen, 2022). Based on the assessment of collaborative capabilities (purpose, span and orientation), the configured OISC should provide a strong foundation for successful innovation and OISC performance. Beyond the externally funded project and the prospective customer, other “non-project” OISC actors, such as surface treatment and superconducting cable suppliers, were necessary to complete the prototype.

Geographical proximity, evident between the companies and UB, was understood as “convenient” but not significant until the later stages of the design and engineering phase. Until then, the Zoom-enabled knowledge-sharing and relationship-building, together with instances of temporary geographical proximity, were perceived as broadly satisfactory.

However, a lack of geographical proximity between UA and the manufacturer negatively impacted the completion of the prototype. To complete the prototype within the timeframe, UATE and UAPL2 wanted to work on the manufacturer’s site to address the remaining problems and support the final assembly, with at least a few weeks remaining. However, this was not possible:

This supports the findings of Bishop et al. (2011) regarding the benefits of geographical proximity for joint problem-solving. The lack of geographical proximity negatively affected OISC performance, causing delays due to:

• hindered joint problem-solving

Social proximity between actors can shift abruptly, significantly influencing collaboration and OISC performance. The relationship and trust levels (Fawcett et al., 2017; Zimmermann et al., 2016) between the project actors and the prospective customer were largely disrupted at the onset of design and engineering, as the management team of the prospective customer underwent significant changes.

This disruption delayed design and engineering in the OISC. It is considered a critical incident within the project and is understood to have set off a chain of delays. toward the end of the project, a similar change occurred: S2MD was replaced, and the new MD, without consulting the other OISC actors, decided to abandon the agreed-upon manufacturing technique and deprioritize the project.

The change in MD affected the actor’s willingness to deliver on the promise. Both examples demonstrate how previously strong social proximity can be lost. This underscores the close connection between social proximity and individuals, in line with Boschma (2005) and Lauvås and Steinmo (2021). Similar shifts occurred both in small and large organizations, illustrating that an organization’s credibility and relational commitment often depend on specific individuals, regardless of organizational size. Though rarely addressed in the literature, social proximity can shift abruptly as individuals enter or exit a collaboration, altering the nature of OISC interactions. In this case, the interaction shifted overnight from collaboration to an arm’s-length buyer-supplier relationship. The loss of social proximity, as bearers of social proximity were exchanged for non-proximate individuals, negatively affected OISC performance in terms of delays due to:

• the need for additional documentation and agreements to compensate for reduced trust

• the project being deprioritized

Lack of organizational proximity in terms of working culture (e.g., Johnston and Huggins, 2021; Dallasega and Sarkis, 2018) had a significantly negative impact on the OISC performance. The lack of flexibility in university working culture, as indicated by Johnston and Huggins (2021), was evident in UB, where fixed teaching periods and inflexible staffing stood in stark contrast to the flexibility and adaptability of the companies. The prototype was significantly delayed due to UB’s rigid structure, a problem that was compounded by the initial delays, as key personnel could not perform their tasks at the required time.

The project was structured for all participants to meet monthly, with biweekly technical meetings for those directly involved in development.

UB did not fully uphold this structure. Based on the limited availability of UB’s participants, UBDPL reasoned that full participation in meetings would be too time-consuming. Instead, (s)he was often the only UB participant in crucial technical meetings. This created a sequential communication structure (as discussed by Mattsson, 2012) with communication through a coordinator (as in Balland et al., 2015) and serves as an example of one participant taking over an entire actor interface. This communication structure prevented UBCAD and others from developing technical, social or organizational proximity with other actors, leading to misinformation and misunderstandings, and ultimately causing confusion, repeated efforts and delays in the innovation’s development.

UB’s inability to deliver on commitments affected operations across the OISC, preventing trust between UB and other actors from developing. UBDPL’s perspective was instead that UA was unnecessarily controlling.

A lack of responsibility was experienced on the part of UB despite the project group’s relatively small size (5 actors). Communication and information-sharing challenges prevailed; from this point of view, the collaboration span could no longer be understood as fully ambidextrous.

Differing perceptions of time (Johnston, 2022) and its criticality were also identified between universities and companies, as expressed by S2MD regarding the experienced delays:

UBDPL also expressed awareness of this lack of organizational proximity, though this awareness did not prevent it from delaying the innovation:

Demonstrating significant flexibility, such as running production overnight or on weekends when necessary, and a strong focus on meeting deadlines, the companies exhibited a markedly different working culture, as is often the case for small companies (in line with Johnston and Huggins, 2021). This flexibility allowed them to recover from some delays. Planning stood out as an area of limited organizational proximity, both in terms of understanding the necessity of planning and how it should be carried out. The companies were frustrated with the lack of details, such as activity plans.

The companies expected UA, initially perceived as the only actor with sufficient knowledge, to supply such a plan. On the other hand, UA was unfamiliar with its expected active production roles in the OISC. In line with the observations of Goel et al. (2017), the manufacturer initiated the planning. The OISC was well into the design and engineering phase before the universities realized the need to maintain a detailed plan. This challenge was addressed by recruiting UAPS, a person with strong production planning expertise. This highlights the added complexity of open innovation when collaboration is required between organizations with differing characteristics. The lack of organizational proximity in working culture and routines negatively affected OISC performance by causing delays due to:

• inflexibility among actors, resulting in an inability to perform tasks on time

• a sequential communication structure, leading to misinformation, misunderstandings and repeated efforts

• differing perspectives on time, which created varying levels of urgency and drive among actors

The lack of technological proximity led to misunderstandings and lengthy discussions in search of clarity, construction errors and time-consuming remanufacturing and testing operations. Throughout design and engineering, several iterations between actors were necessary, highlighting the importance of each actor having a degree of understanding and, therefore, technological proximity to other actors. The willingness or ability to learn (Nguyen et al., 2019) was evident within companies and UA, while it was less obvious within UB. Therefore, the collaboration orientation was not understood as fully ambidextrous. Regarding the innovation, insufficient proximity was identified specifically between UB and UA/the manufacturer. This caused several misunderstandings, leading to mechanical design and CAD construction mistakes. UB was aware of the lack of technological proximity between themselves and UA/the manufacturer (a one-to-many interface):

Despite this awareness, UB could not expand its knowledge base through learning:

The expected increase in technological proximity over time (Balland et al., 2015) did not occur rapidly in this interface. The lack of organizational proximity, such as limited meeting participation and a sequential communication structure, hindered the development of technological proximity between UB and other actors, aligning with the findings of Johnston (2022) and Östbring et al. (2017).

The technological proximity between the prospective customer and the companies was initially insufficient, making it difficult for actors to understand each other and move beyond their own knowledge domains. This is illustrated by MDE seeking to understand why a specific solution was not considered viable, expressing frustration:

In those instances, UA acted as an interpreter or mediator, limiting the negative effects. However, the technological proximity did develop over time among the project actors. A joint study visit to the prospective customer halfway into the project is considered crucial.

This illustrates the effect an “episode of geographical proximity” had on technological proximity. Although technological proximity within the OISC increased, its development was slow in specific interfaces.

A lack of technological proximity was also identified between the non-project suppliers and the project actors (a many-to-many interface). The superconducting cable was one of the innovation’s most distinctive aspects; it was clearly not an off-the-shelf item. Despite discussions with multiple suppliers and pre-selection sampling, the result was a poorly insulated cable. This triggered a chain of events, including testing, quality improvement efforts, re-testing and re-assembly, leading to significant delays. UBDPL reflects on the possibility that the project may not have had the required buyer competency:

Similarly, the selected surface treatment supplier’s quality deficiencies caused irreparable damage to the magnet sub-systems and necessitated a second round of supplier assessment and selection. This led to additional time-consuming re-manufacturing operations, with significant delays in finalizing the innovation.

Including multiple actors in innovation creation is beneficial (Iftikhar et al., 2024; Zimmermann et al., 2016). However, this approach also involves risks. The non-project actors did not commit to collaborative innovation and had lower levels of involvement, which resulted in a less ambidextrous collaboration purpose during design and engineering than during configuration. This was reflected in the lack of proximity, which negatively affected OISC performance. According to Solaimani and van der Veen (2022), commitment to co-innovation can be achieved through equal collaboration among actors, but this was not the case with the non-project actors. The lack of technological proximity negatively affected OISC performance in terms of delays due to:

• construction errors and time-consuming remanufacturing

• quality deficiencies and, in some cases, irreparable damage to magnet parts, leading to renewed time-consuming supplier selection, testing and remanufacturing.

To deepen understanding of how proximities, knowledge sourcing and collaborative capabilities relate to OISC configuration and performance, social and technological proximities are identified as the most important dimensions affecting OISC configuration. Social proximity is the foundation for actor inclusion, as it serves as a bearer of trust based on previous collaborations and as a vehicle for assessing the suitability of potential partners. In contrast, the degree of technological proximity can be understood as the actual inclusion criterion, together with the actors’ resources and competence levels. Organizational proximity was overlooked during the OISC configuration, based on the assumption that actors would adopt the project’s working methods. During design and engineering, the lack of both technological and organizational proximity led to miscommunication, which negatively affected the OISC through errors and rework, resulting in delays. The lack of organizational proximity further exacerbated the existing lack of technological proximity and hindered the development of social proximity. Brunswicker and Vanhaverbeke (2015) advocate for application-oriented or full-scope knowledge sourcing to establish a strong foundation for collaborative capabilities and successful innovation. Despite the deliberate use of application sourcing—and, by extension, full-scope sourcing—the innovation work in the OISC remained a challenge. This suggests that understanding OISC performance through the proximity lens complements these knowledge-sourcing strategies. Findings also indicate that the beneficial collaborative capabilities present during configuration were lost as knowledge from additional actors had to be sourced during design and engineering. Those actors had no opportunity to commit to an ambidextrous collaborative purpose and orientation, as recommended by Solaimani and van der Veen (2022).

The study addresses the limited understanding of open innovation in supply chains, as highlighted by Zahoor and AL-Tabbaa (2020), Patrucco et al. (2022) and Solaimani and van der Veen (2022). By shifting the focus from the innovation to the supply chain through which the innovation is realized, this study contributes to the supply chain management literature. While established in other fields, the lens of proximities was introduced here to supply chain studies, where little overlap between proximities and the supply chain management literature has been identified. Notable exceptions include Oyedijo et al. (2024), who explore geographical and what might be referred to as supply chain proximity, and Solaimani and van der Veen’s (2022) idea of balanced diversity, based on proximities between actors. Our study builds on and empirically grounds Solaimani and van der Veen’s (2022) primarily conceptual study of collaborative capabilities. Therefore, the lens of proximities can provide a deeper understanding of and contribution to supply chain management. Based on this study, we find support for treating social and organizational proximities as separate dimensions, given social proximity’s impact during configuration (e.g. Johnston, 2022) and how it changes as individuals enter or exit the collaboration. Therefore, the departure from Knoben and Oerlemans’ (2006) dimensional structure is understood as valid.

The studied OISC, which comprised five project actors and a handful of non-project actors, answered the call from Patrucco et al. (2022) for studies involving larger sets of collaborators. Including many actors in the analysis ensured multiple perspectives were considered. In doing so, the study extends the internal or dyadic focus of earlier proximity research (e.g. Johnston, 2022; Nguyen et al., 2019) to a supply chain, covering sourcing from suppliers through companies and universities with connection to the ultimate customer, following Spekman et al. (1998). As a study object, the supply chain revealed a complex OISC where proximities were first assessed in one-to-one interfaces. Over time, one-to-many and many-to-many interfaces were also identified. This finding is in line with Näslund et al. (2010), who acknowledged that study objects can develop over time, which increases the understanding of proximities. This understanding of different OISC interfaces in which to assess proximities is a contribution to both proximity and supply chain management literature.

By exploring how actors are included in an OISC, this study contributes to Brunswicker and Vanhaverbeke (2015) regarding how external knowledge sources are acquired. Here, aspects of knowledge sourcing that overruled proximities were revealed, such as requirements from funding agencies. This highlights that proximities can be impacted by decisions made by actors outside the OISC, complementing Gibson et al. (2016). Such requirements have been identified as negatively affecting proximities during design and engineering and causing delays in the OISC. These findings expand the knowledge of the potential barriers to collaborative innovation in supply chains, addressing a research gap identified by Anderson et al. (2023).

This study further conceptualized OISC as consisting of configuration and design and engineering phases. An important insight is that configuration is not a discrete period confined to the beginning, when OISC actors are first included. Instead, this study reveals that additional knowledge sourcing may be necessary and can become important during design and engineering when current OISC actors lack specific competencies or resources. Similarly, when sourcing knowledge in the configuration phase, collaborative capabilities should be a key aim, underscoring their importance in that phase as well. This finding complements our analytical framework, suggesting that knowledge sourcing and collaborative capabilities should be recognized in parallel, adding nuance to the simplified framework. The finding that distinct types of proximity play key roles in different OISC phases contributes to both supply chain management and proximity literature. This insight was made possible by our longitudinal study, a methodological approach not commonly used in this area of research (e.g. Anderson et al., 2023; Patrucco et al., 2022).

Lack of social proximity—defined as unfamiliarity with supply chain partners—proved significant, as most delays in the OISC were attributed to actors with little social proximity. This finding challenges the core principle of open innovation: knowledge sourcing from external actors. It underscores the risks of failing to assess potential partners’ organizational (i.e. working culture and methods) and technological proximity (i.e. knowledge bases), which negatively impacted the OISC during design and engineering. Another interesting aspect of social proximity, with implications for supply chain management, highlights the dynamic role of individuals in shaping proximity. As the initial contact person, who bore some social proximity, signed the agreement but handed the task over to others, who lacked that proximity, various adverse effects on the OISC developed. This aligns with findings in the literature on transport supply chains, where an LSP secures an agreement and subsequently hands it off to other LSPs or haulers, leading to various challenges (see, e.g. Forslund et al., 2022). The resulting loss of social proximity highlights the value of referencing multiple individuals within each organization and points to the relevance of careful knowledge sourcing and actor selection. In some relationships, social proximity was lost during design and engineering as individuals left their organizations or as new people assumed roles previously held by socially proximate individuals. This adversely affected trust relationships and the nature of interactions taking place within the OISC, again resulting in frustration and delays. The diminished social proximity further compromised collaborative capabilities, such as the collaborative purpose. These findings complement the dynamic development of proximities as the collaboration progresses (Balland et al., 2015; Johnston, 2022), but primarily in a negative way. The collaboration span was expanded by including non-project actors with no social or technological proximity. However, the collaboration purpose was simultaneously weakened as actors lacking commitment to collaborative innovation became involved. This inclusion disrupted the balanced diversity of actors (Solaimani and van der Veen, 2022).

Interplays between proximities were identified during design and engineering. A lack of organizational proximity, reflected in differing working methods and communication structures, hindered the development of social proximity and trust by preventing actors from getting to know and understand one another. Similarly, this lack of organizational proximity also accentuated the lack of technological proximity. While Johnston (2022) and Cassi and Plunket (2014) discuss how certain proximity dimensions can compensate for others, we observed how a lack of organizational proximity undermines the development of other proximities.

Low performance across the entire OISC may result from the loss or lack of any proximity dimension in any interface, ineffective knowledge sourcing or insufficient collaborative capabilities. Miscommunication and misinterpretations (Dallasega and Sarkis, 2018), leading to errors and rework (Zimmermann et al., 2016), were observed and affected OISC performance, captured as delays. Studying collective OISC performance complements the approach of Solaimani and van der Veen (2022), Patrucco et al. (2022), Brunswicker and Vanhaverbeke (2015) and Zhang et al. (2024), who focused on focal firms and their innovation performance. Studying OISC performance also revealed the added complexity of interactions between project and non-project actors (many-to-many). Capturing OISC performance during design and engineering offered a proactive and complementary approach compared to ex-post evaluations (Zhang et al., 2024).

The study contributes to practitioners such as managers in universities and companies, funding agencies, policymakers and non-managerial staff involved in open innovation. Toffel (2016) encouraged focusing on aspects within practitioners’ control to motivate action. An important implication is an increased understanding of the structured proximity concept and the importance of considering social, organizational and technological proximity in decision-making. The finding that an OISC consists of a configuration phase and a design and engineering phase, and the structuring of the effects of proximities on both the OISC configuration and performance, reinforces the awareness that proximities central to configuration are not necessarily the most important during design and engineering. The effects of low organizational proximity first became evident during design and engineering, however, highlighting the need to address this issue already during OISC configuration. To support strong OISC performance, supply chain practitioners can proactively ensure necessary proximities, as a lack of proximity in key interfaces can translate into delays for the entire OISC. The study highlights both opportunities to build proximity and the risks of losing it within the OISC. Knowledge sourcing and collaborative capabilities provide useful vocabulary and deeper insights. The study also highlights the need to involve supply chain management professionals in knowledge sourcing, given their expertise in sourcing-related domains.

The study informs departmental heads and deans, especially in teaching-intensive universities, on how organizational flexibility can support successful research and development outcomes. It also alerts funding agencies to the risks and potential consequences of actor inclusion decisions in research projects. Altogether, the study recognizes that proximities can foster strategic decision-making and collaboration, offering insights for external policymakers aiming to foster innovation-friendly environments.

Access to a unique single case enabled rich longitudinal data collection and deep qualitative insights, offering a strong foundation for early stages of theory building. However, the study’s context limits the transferability of its findings. While they may apply to OISCs that include research- and teaching-intensive universities and technologically advanced small companies, they may not extend to industries beyond technology-intensive manufacturing. To test the transferability of the findings and continue the process of theory building, they could be investigated in a broader sample in further quantitative research. Such a study could include large companies, other industries and potentially geographic contexts outside Sweden. Additional research, with a quantitative approach, can also provide a deeper explanation of how proximities are linked to various dependent variables, including collective innovation performance, such as efficiency, innovativeness or success in the OISC, in line with the call for further research by Brunswicker and Vanhaverbeke (2015). With a qualitative approach, further research can also systematically focus on identifying supporting and hindering factors in OISCs. It can also distinguish between different phases—in greater detail than simply configuration and design and engineering—of an OISC. Treating the design and engineering phases as a single entity, following Shamsuzzoha and Helo (2018), may be a simplification that overlooks essential details.