Understanding knowledge translation in university–industry research projects: a case analysis in the automotive sector

2.1 U–I research projects: key features affecting knowledge translation

U–I research projects have been studied from a variety of perspectives, including institutional dimensions of the collaboration, organisational culture, inter-organisational relations, intermediaries and brokerage able to facilitate such collaborations (Morandi, 2013).

Amabile et al. (2001) suggested that there are three important features of project collaborations between academic researchers and business practitioners. First, they involve people who are members of different professions (academia and business), second, they may be understood as collaborations between individuals or teams, not between organisations and third, the collaborators are not all members of the same organisation.

A significant part of organising these interdisciplinary research projects is the establishment of a shared language, relatedness of specialised knowledge, shared meanings, and ability to recognise the contribution of each individual knowledge domain (Kellogg et al., 2006).

Many scholars (Rajalo and Vadi, 2017; Schaeffer et al., 2020) investigated the issue of knowledge transfer as a key variable to get success in this kind of projects. Following Christensen (2003), knowledge transfer requires the identification of existing knowledge, the ability to acquire it and the subsequent application of such knowledge to develop new ideas. In the same vein, Simeone et al. (2017, p. 1409) stated that “the objective of a knowledge transfer process which takes place between two or more actors (individuals or organizations) is to enable an actor to acquire the knowledge of another actor”.

A number of researchers have dealt with the interfaces of knowledge transfer among individuals engaged in inter-organisational collaborations across different domains of expertise (Simeone et al., 2018; Axelson and Richtner, 2017). Rajalo and Vadi (2017) studied the different practices adopted during the two main steps of a U–I project (initiation and implementation) to facilitate the management of knowledge. They adopted Rau et al.'s (2012) approach to define practices implemented to fill the gap between the partners to facilitate knowledge transfer. They described boundary-crossing and pragmatic boundary-crossing mechanisms as different practices used to facilitate communication and build a common framework among the partners. Collins and Evans (2007) examined the interactional conditions affecting knowledge flows in and across organisational borders. Specific forms of expertise play an important role in facilitating the coordination and transfer of knowledge across boundaries. When dealing with knowledge transfer at the interface of U–I research projects, Simeone et al. (2018) stated that, given the different objectives and languages prevalent in academic and industrial contexts, translation mechanisms between these two groups of stakeholders are needed to establish a common platform of communication and knowledge sharing, and “[…] there is still a lack of investigations about how to support the alignment and collaboration of different stakeholders, particularly in the context of academic entrepreneurship processes” (Simeone et al., 2018, p. 435). Knowledge translation becomes a (key) glue able foster knowledge transfer (Petrilli, 2003). Collaboration between heterogeneous actors from universities and industries often involves complex knowledge translation acts because each partner has his/her own nomenclature, demands and expectations of innovation (Sandberg et al., 2015).

In a U–I research project, translation is a dynamic process, where each actor operates at the organisational level to produce a specific output (a document, report, video, etc.). This involves adopting others' perspectives and selecting a framework and words that can be easily understood by others. It is therefore produced first at organisational level and negotiated and defined through collaboration (Secundo et al., 2019b).

One of the most critical factors in U–I research projects is the creation of an interactional space, where actors work together to overcome their individual and organisational interests and needs and to negotiate and collaborate to achieve a common goal. Creating this space is vitally important and may need specific effort: every player involved should his/her language to adopt a different framework, moving from his/her “natural environment” (De Rond and Bouchikhi, 2004). Universities often use a more scientific language, and the vocabulary can be hard to understand and share. Academics are used to publishing their work in scientific journals, each of which has its own style and rules for publication, including for language and references. Similarly, in industry, each organisation (or each organisational unit) uses its own language and has a particular way of storing and transferring knowledge.

In the knowledge-translation domain (Liyanage et al., 2009; Simeone et al., 2017), the idea of shedding light on the project as a knowledge translation mechanism is not analysed in depth, even if many studies are related to research contexts, in which project is an elective organizational arrangement.

2.2 Project as a knowledge-translation mechanism

Projects are able to engage with heterogeneous inputs from participants in order to foster knowledge translation. They allow the involved actors to clearly spell out their contribution towards a common endeavour. In this sense, they corroborate claims such as “this is what my knowledge means for you” (Seaton, 2002; Liyanage et al., 2009) rather than those resonating with putting in common bits of knowledge (“this is what I know so we can mutually integrate our knowledge”).

In projects, participants need “good representations of what the others know […and] may thus, based on quite a minimalist base of shared knowledge, develop a pattern of interaction and the collective competence needed” (Lindkvist, 2005, p. 1200). The project may represent a guidance as to the knowledge needed to accomplish the goal, the feasibility of the solutions to be implemented within limits set as to the allocated resources. In our view, knowledge translation becomes then central in mutually clarifying knowledge inputs for the different participants in order to collectively act with a shared purpose.

In the management literature, projects are considered as a way to achieve novel objectives by reducing uncertainty (De Meyer et al., 2002). In many complex projects, problems emerge and are identified and solved by use of appropriate tools or through a web of interactions across organisational and disciplinary boundaries (Brown and Eisenhardt, 1995). Project management is therefore crucial to sustained problem-solving and understanding how activities in the project are organised and implemented. In the organisational-driven literature, management of projects requires consistent mechanisms over their lifetime.

Appropriate mechanisms are required to ensure that project activities are carried out, recognising existing interdependences. For instance, joint planning, task allocation, meetings and social structures all enable people to share and create organisational knowledge. Through standardisation, negotiations among participants and social interactions, projects therefore have full potential to be seen as the heart of the knowledge-translation process. Project teams have access to different knowledge bases but must also solve translation problems arising from multiple perspectives. In fact, at this interface, projects are entities that can provide for both complexity reduction and negotiation (Engwall, 2003).

Key project elements may have profound repercussions on the knowledge translation outcomes. Actors in the U–I projects are thrown away from their routine processes and structures to develop joint products, processes or solutions. Such separation may be of help for achieving the common effort and contribute to minimize obstacles to implementation of the collaborating entities engaged in a shared enterprise.

Goals push towards a mutual understanding, re-activate previous experiences and may be re-inforcing role-based coordination (Valentine and Edmondson, 2015). The temporal dimension is also obviously central in project-based reasoning because projects rely heavily on the use of the timeline as both a device to draw support and resources and for managing dispersed time orientations and local knowledge processes as time becomes a pervasive issue to prevent participants from being guided by concurrent orientations (Lindkvist et al., 1998, p. 948). Time constraint is also a form of control (Lindkvist et al., 1998) allowing for execution to proceed without external interference.

In sum, goal definition and time constraints represent two basic traits of the project as organizational arrangement, that are often critical in U–I research contexts and that can help in understanding knowledge translation outcomes.

However, projects account for specific knowledge translation mechanisms that may use both rationalistic tools and social interactions among participants. The first deals with conventional issues such as breaking the project into smaller components, ensuring careful planning, scheduling, estimating and execution of project tasks and striving for cost and time efficiency throughout the project to achieve the optimum outcome. To manage sources of uncertainty requires also close attention to interactional aspects. Managing meetings and deploying “soft” tools to enhance socialisation around technical issues (such as visualisation artefacts) may be crucial. There is a clear trade-off between reducing complexity of knowledge transfer through standardisation and more time-consuming issues of negotiations among participants requiring time and relationship-building (Engstrand and Enberg, 2020).

It is also useful to note that, within projects, focussing on planning, organizing, coordinating and controlling does not always fully reflect organizational reality as messy, ambiguous, fragmented and political in character (Hodgson and Cicmil, 2006). To this extent, widely accepted conventional views of the relationships between project and knowledge tend to underplay political and social processes pointing instead towards more simple issues and problems (Cicmil, 2006) that can be addressed through an appropriate corresponding action and behaviour.

3.1 Research context

The empirical study was based on a qualitative investigation. We carried out the analysis iteratively with a semi-grounded approach, and our data informed the selection of theories (Van Maanen et al., 2007). We chose an example of a setting where project management was used in relation to the interplay of various stakeholders in a U–I collaboration (Eisenhardt, 1989). In line with the inductive approach adopted, we focussed on a single case study of particular interest rather than studying several cases (Yin, 1994). This should not be viewed as a limitation of our research, as shown by Cunningham et al. (2017), who conducted a review of qualitative studies in the field of knowledge transfer. The in-depth case study approach is advised when the boundaries between phenomenon and context are not entirely obvious, for example, the knowledge translation mechanisms used within a U–I partnership (Simeone et al., 2018; Dubois and Gadde, 2002). This is in line with Dyer and Wilkins (1991, p. 614), who advocated using a case study approach, and recommending that researchers should be aware of the benefits of “the careful study of a single case that leads [them] to see new theoretical relationships and question old ones”.

We analysed empirical material collected within a research project involving a partnership between Fiat-Chrysler Automotive (FCA), a multi-brand auto manufacturer with a product range across different market segments and two universities, the University of Sannio (Department of Management, Italy) and Portsmouth Business School (Department of Operations and System Management, UK). The project's main aim was to develop and implement a multi-criteria method to modify the quality control cycle of assembly-line production. The researchers from the two universities had different and complementary expertise in innovation management. The Portsmouth researchers had expertise on topics such as data envelopment analysis and multi-criteria decision-making. The Italian researchers had skills in organisational design and human resource management. The project was funded by the National Operative Programme FSE-FESR Research and Innovation, Action 1 – “Dottorati Innovativi con caratterizzazione industriale”.

At the time of drafting this case study, the project (formally presented in December 2017) was in an intermediate stage (October 2018–July 2019). It followed the decision of the Italian Ministry of Education, University and Research to finance the project activities for a three-year period and was at the end of the third project phase (see Table 1).

The case selection phase used two main criteria. First, the specific FCA plant with which the partnership has developed shows the adoption of knowledge management methods and tools as enablers for the development of process and product innovation (Canonico et al., 2018). This study used empirical material collected at FCA's Pratola Serra Plant, in Avellino, Italy. The Pratola Serra Plant was founded in 1991 and has received many certifications and awards, but it has also experienced some problems and important organisational changes. From a performance perspective, it moved from producing 570,000 engines in 2007 to 170,000 in 2015, back to 300,000 in 2016. Recent trends and forecasts are encouraging enough to think this figure may increase again. Recently, FCA has refined the organisational design at the plant using the Work Place Integration initiative, a broad program of activities designed to streamline manufacturing processes, using a lean product development approach.

Second, as is common in literature, we selected a case study in which the phenomena of interest – in this case, the use of knowledge translation mechanisms in a U–I project setting – were transparently observable. The case study provided a consistent, differentiated and information-rich setting for studying the phenomenon of knowledge translation. One of the authors of this paper was also directly involved in the project team.

3.2 Data collection

We used three data collection techniques: internal document analysis, participant observation and semi-structured interviews. Documentary analysis enabled an understanding of the project structure and tools (for example, flow diagram, GANTT charts, budget, work breakdown structure scheduling, and diagrams). Participant observation was used to explore knowledge translation issues among the different partners of the project. During her fieldwork, the third author spent from 8 a.m. to 2 p.m. in the Plant for one to two days a week during alternate weeks. Having free access to the Plant premises, she was able to make numerous formal and informal contacts and become relatively familiar with the management. Finally, in-depth interviews with the management of FCA were the main source of data for this study. Five semi-structured interviews were conducted with the world class manufacturing Plant coordinator, the Pratola Serra Plant manager, the Plant HRM manager, a manufacturing engineering manager, the Plant quality manager and the Portsmouth expert in data mining and multi-criteria decision-making. Each interview was conducted by at least two authors of this paper to reduce interviewer bias (such as first-impression error, non-verbal influences or negative emphasis). The interviews were based on an open, wide-ranging protocol (Czarniaswka, 2004), provided one week before the first interview. This invited the interviewees to explain how they had worked in the different stages of the research project. The interviews were guided by a questionnaire including questions about (see Appendix):

1. The overall organisation of the project (for example, in how many phases was the project articulated? Could you describe the main objectives, working methods and critical points of each phase? What were your main tasks?).

2. Organisational practices within the cross-boundary team (What were the main coordination mechanisms used among project members? Could you describe an episode where there was a misunderstanding between researchers and managers? Can you describe a specific incident? What actually happened? Can you provide an example of a problem that you solved and how? How was the meeting space used? How frequently were formal or virtual meetings held to discuss new ideas or process upgrading? How was project management software used?).

3. Knowledge management mechanisms and practices (What were the main tools to exchange data, information and knowledge between project members at different phases of the project? How were those tools actually used? What kind of knowledge did you need to work effectively?).

The protocol aimed to stimulate interviewees' interest in this participative research process and to promote a narrative approach, crucial for the success of the interviews. The collection of data was carried out between October 2018 and early May 2019.

3.3 Data analysis

Our analysis was in stages. The interviews were taped and roughly analysed, and we wrote memos to coordinate and develop a common view among the authors. In the second step, all five authors listened to the recordings. We compared the raw material and the written notes taken during the interviews with the recordings. The personal involvement of one of the authors in the research project allowed us to inquire into the process of knowledge translation from the inside (Evered and Louis, 2001).

We first listened to all the interviews to understand the general ideas expressed. We then reviewed the recordings, transcribing the interviews verbatim, stopping the recording after every sentence, to be sure that the meaning was clear. We did this immediately after recording and read the transcripts several times. Sections dealing with the practices, tools and artefacts used for knowledge transfer during the different stages of the project were noted. This reminded us of the events, words and looks and helped us to remember what happened in each interview. Most of the conversations during interviews were not spontaneous, but we were interested in showing the managers' and practitioners' voices and attitudes, to enable readers to experience the interviews more effectively. We also used a triangulation strategy, comparing the evidence from different sources (interviews, participant observation and formal sources). In the final step, we used the interviews as texts, with knowledge translation theory as a guide. We sought to understand the most frequent affirmations about cross-boundary interactions, project tools (e.g. project template, project charter, diagrams and virtual design tools), formal and informal coordination mechanisms and procedural and routine patterns. This provided the ‘text' for interpretation. The data were therefore interpreted by identifying how the translation process was developed using the set of project management methods and practices.

4.1 Project planning

The most common coordination solution in the early stage of the project planning phase was methods and practices devised by the partners. They largely focussed on communication and direct comparisons between partners. During a series of meetings, the theoretical and operational perspectives of the study were discussed and defined. The expected goals and outputs were negotiated, and the investigation methodology was developed.

Afterwards, in the definition of the project proposal, the comparison between the partners was channelled into a formal path as the project documentation was drafted. An important role was played by formal evaluation and funding tools (project template, flow diagram, GANTT charts, timeline, budgetary control, work breakdown structure scheduling, status table and fishbone diagram). The template for the project (which must be followed to obtain public funding) had been a part of similar previous tenders. This constituted the main frame in the implementation of this phase. Following the guidelines to make the work eligible and attractive for funding meant explaining the research in “project language”, setting out responsibilities, resource allocation, work packages and reporting structures. These were all central components in the toolbox and had to be included so that the application would pass the first component of the assessment process. The standardised forms for planning purposes were a way to specify and control the content to be shared among project members. This gave structure to the discussions in the project meetings.

4.2 Project launch

After the positive evaluation by the Ministry and obtaining three-year funding for the proposed activities, the group of four university researchers and four company managers carried out a series of meetings and contacts to define the details of the first activities in the project. In this phase, project management software was adopted: a cloud-based platform that allows users to gather ideas, execute tasks, track progress and ultimately achieve goals. The team members started to use this app for project activities like sharing documents and signing off plans.

Two meetings were held at the manufacturing plant to present the project to the employees involved in the work processes studied and to discuss the selection and data collection criteria for the start-up phase of the theoretical decisions model. The Plant Director stated in an interview:

From the beginning, I proposed to choose, as a focus of the research study, the problem of the evaluation of improvement actions that could be designed periodically, after the detection of anomalies during engine assembly. After a long chat with the two professors and the sharing of a visual diagram (a graphical schema presenting error prioritisation) … I became aware of the need to modify our system of classification and evaluation of anomalies during the assembly phase … The process by which academics and engineers acquired a mutual perception of each other's needs in the development of the project was far from easy.

Engineers and academics analysed the sequential patterns of the control processes, and the existing organisational procedures and routines were identified and mapped. After four months of work, the project group had completed the monitoring of the anomalies that emerged in the engine assembly process.

4.3 Building the new multi-criteria decision-making method

In the third phase of the collaboration project between the company and the university, the main effort was dedicated to transferring, understanding and interpreting existing decision-making models in a common form. It was decided to set up a working team from the existing project group, composed of two engineers, two researchers from the University of Sanni and a researcher from the University of Portsmouth. This team was tasked with achieving a radical change in the anomaly analysis decision model and collecting all the data sets from the last two production cycles for the transfer to Portsmouth. It was also decided that one of the Italian researchers would participate in the testing phases at that university.

This team worked together in a specific space at the Plant, where academics and practitioners interacted and collaborated using interactive tools, especially for visualisation. The team's work was enabled by a large-scale visualisation of content and personalised desktop views. The room integrated physical spaces and virtual tools in a visual collaborative environment that supported co-located Italian team members and the UK-based researcher in tight synchronous collaboration. The manufacturing engineering manager commented:

The visualisation of technical data increases awareness of the design and drawings among all those involved. The use of pictures, graphs and colours makes it possible to grasp complex relationships between many different factors very quickly … Visualisation helped to support fruitful communication and decision-making processes, [and] strengthen the intermediate goals of our work ….

5.1 Prioritising errors

In the project launch phase, the specifications for the quality control model were defined. In this phase, an event of particular importance was likely to become a highly critical factor. The model proposed by the researchers used the application of the multi-criteria logic to analyse errors made by the workers during the assembly of the engine. The model, among other things, was intended to classify the errors by level of importance.

FCA already had a system for the analysis of error prioritisation. The system was based on WCM protocols that were adopted by all the companies in the automotive group. A frame called the QA Matrix identified anomalies related to the production process and located them to the phase in which they were generated, providing an index of priority that was the result of different weights given to the frequency of the error, the costs generated, the detection methods and the severity. At the time of the development of the first part of the new model, the company managers involved in the project were neither interested nor available for this type of experimentation. The quality manager asserted:

We have invested heavily in training and WCM-inspired management systems in recent years. Our error prioritisation system allows us to “weigh” errors detected in the assembly line using the four most important criteria for evaluating the performance of the manufacturing process.

The multi-criteria decision making expert commented:

I thought I couldn't convince them! The quality manager stubbornly and repeatedly stressed the lack of interest in experimenting with a different way of classifying production anomalies. Beyond the need to complete my experimental research, I couldn't accept not being able to understand the limits of that error classification approach. Their classification criteria were based on an algorithm similar to that of the average of four variables and did not lead to an accurate analysis of the observed phenomena.

The working group had four working sessions on this topic. In the third of these sessions, the use of a visual diagram made by the academics proved to be an unexpected turning point (see Figure 1). The visual representation (a visual translation) of the methodological advances – termed the glasses of wine – showed the weaknesses of the existing error prioritisation method based on boolean logic (weakness a. No definition of weights; b. Aggregation of criteria) and intuitively illustrated the main advantages of the multi-criteria method (based on a fuzzy logic). The diagram was accompanied by a series of sketches made with an electronic pen by mathematicians.

As a result of this presentation, the working group as a whole became more aware of the proposal made by the university partners. The new multi-criteria method made it possible to assign priority indices to assembly errors, evaluating anomalies more precisely without losing valuable information, the weakness of average-based algorithms. The last of the four work sessions followed a series of chats and interim reports exchanged through the shared project management software platform. It was attended partly in person and partly remotely, through videoconferencing.

The group closed the session by deciding to adopt the new multi-criteria model for error prioritisation but also planned to work in the next testing phase to better define weighting standards. The organisational academic stated:

The time and energy spent in the dialogue together with the intuition of elaborating “the diagram of the glasses” was the real foundation for our ability to understand each other. The glasses and the image of the trapeziums seemed to be “magic keys” to open a new channel of communication. The company understood the richness of the analysis we proposed and we understood the criticality of changing a practice that affected the evaluation of people's mistakes on the assembly line. Now that we are in the experimental application phase, it is as if we had a new common subject to shape!.

5.2 Building the portfolio of critical processes

A second critical episode in terms of communication and comparison between the project partners emerged when the forms of experimentation around the new quality control model (phase three of the project) had to be defined. After reaching an agreement on the “prioritisation” of anomalies, academics and managers addressed the issue of identifying organisational improvement actions capable of reducing errors in the assembly process.

It took two weeks of activity during which three meetings were scheduled. The academic partners proposed applying the multi-criteria logic to the evaluation of alternative organisational development actions. The company already had a consolidated methodology but was willing to develop a more effective one within the general rules of the WCM system. The existing method was based on the concept of “critical process”: in practice, on the idea of modifying the part of the process in which serious anomalies most often occurred. The concept of “critical process” was new to the academics, or at least not entirely clear compared to the field of applications that had already been tested in other research projects. The project team therefore initiated a discussion about the possibility of using multi-criteria analysis as a basis for decision-making for organisational change, adapting the model to the WCM pillars.

The possibility of visualising a decision-making grid (DMG) method during group meetings significantly facilitated comparison between the positions of managers and academics. This generated a new convergence of ideas on the concept of “critical process”. Through the matrix, it was possible to identify “pieces” of the process, distinguished in green, yellow and red (see Figure 2). The production engineers made short videos (attached to the platform) to show which assembly activities corresponded to those different “pieces”.

Using this method, the project team generated an original way of researching the process in which to invest in quality improvement. Rather than identifying the process in which errors occurred most frequently, they identified processes in which re-design interventions could be carried out (even in parallel), and where the effects could be effectively evaluated. The project team had therefore identified a new concept of critical process: from a process in which more serious errors occurred, to a process where the company could intervene with greater benefits, selecting the nature of the benefits (for example, economic, employee safety, implementation time or economies of scale).

At the end of the discussion at the last meeting, the mathematical decision-maker said:

I am pleased that one of the most powerful properties of multi-criteria methods has emerged, namely the possibility of considering a diverse set of criteria at the same time and involving the decision maker more closely in the choice of improvement. This application opens new horizons for future research activities.

Even the quality manager recognised:

With the multi-criteria approach to error analysis, we are no longer looking for a unique and optimising result, but trying to identify the elements useful to clarify the priorities on which to base the choices […] even if this does increase and aggravate the level of participation and responsibility of the manager in the analysis of alternatives.

The organisational academic added:

I think we have achieved an additional advantage … Having developed the concept of a “sensitive process” together, the DMG method will allow us to build a portfolio of processes on which to focus investment … and in the subsequent testing phase, we will strengthen this perspective!.