The Role of Management Controls in New Product Development: Codifying a Collective Source of Creativity

Abstract

This paper contributes to the understanding of the role of Management Controls (MCs) in mediating and stimulating creative processes. Adopting a knowledge codification perspective, it demonstrates how actors create multiple MCs that open different ‘spaces’ serving distinctive functions, while also being interrelated in the creative co-creation process. A ‘core MC’ creates space for recognizing, localizing, and developing new connections between items of knowledge, culminating in a new shared understanding of NPD, i.e. a synthesis; a ‘supplementary MC’ stimulates actions and interactions on the micro-level that can challenge existing knowledge structures and cause-and-effect relationships, potentially resulting in the creation of an antithesis. Therefore, we elaborate on how MCs themselves can serve as a collective resource of creativity. Additionally, we add to the debate concerning the aspects determining the assessment of controls as enabling, arguing that the design characteristics of a single MC do not determine its enabling assessment, but rather it is the dynamic material nexus that MCs form and their fluid interplay that determine their assessment. Finally, our insights allow us to contribute to the relational view of knowledge objects such as MCs by evidencing a transition of an epistemic into a technical object caused by mutual interaction.

Keywords:

• Management control
• Enabling controls
• Codification
• Collective sense-making
• Creativity

1. Introduction

New product development (NPD) is a vital organizational process, which, due to its inherent uncertainty, is difficult to manage (Davila, 2000). Early research into the role of management controls (MCs) has primarily argued that they impede innovation (e.g. Abernethy & Stoelwinder, 1991). However, more recent literature has identified a positive impact, that MCs themselves can become an engine for innovation by luring people into action, through tracing past and current courses of action, which allow actors to envision future developments (Revellino & Mouritsen, 2015), creating a space in which actors can perform creative work (Busco & Quattrone, 2015) or triggering a prompt and sustainable search for new or better solutions as they are incomplete visual representations (Busco & Quattrone, 2018).

This paper sheds additional light onto features of MCs that make them an integral part of creative co-creation processes such as NPD. This will contribute to the transfer of scholarly attention away from the provision of content to the role of MCs as mediator and stimulator in creative processes (Quattrone, 2017; Revellino & Mouritsen, 2015). In particular, we address how actors create different performable spaces by codifying knowledge in MCs, and how this can generate moments of ‘collective creativity’ (Hargadon & Bechky, 2006) potentially culminating in an antithesis and ‘creative synthesis’ (Harvey, 2014). Based on a single in-depth case study of a manufacturing firm undergoing a fundamental restructuring of its new NPD process, we describe how actors can develop different MCs to support collective creativity when developing new products. Specifically, we demonstrate how actors commence their collective sense-making process when they begin to develop an initial ‘core MC’. During the first step, they enact and select information to categorize knowledge, followed by the second step being the creation of a shared language. The core MC provides intellectual grounding and opens a space in which questions can be raised, definitions of objectives be reviewed, and connections between different information ascertained and understood as well new order and responsibilities of the NPD activities discussed. However, we found that actors were challenged by the core MCs’ ‘incompleteness’ (Busco & Quattrone, 2015) when engaging in NPD. This incompleteness was necessary though to reach a synthesis, i.e. a new way of the actors’ understanding what NPD is. Arising from the background of this synthesis, the actors developed a ‘supplementary MC’ when performing NPD on a daily basis. In the supplementary MC, a space is opened for sharing operational impressions by providing fixed instruments that show the individual steps to be accomplished during the NPD process and prompt actors to monitor and deliver the expected outcomes. The visualization of task interrelatedness through supplementary MCs improved actors’ creative capacity to process cognitive content during collective sense-making processes. As a result, the supplementary MC created a space for organizational reflexivity that, in some cases, led to the development of an antithesis.

Against this background, this study makes several contributions to current understanding of how MCs serve as engines for innovation. Firstly, the study sheds light on how the creation of MCs can serve as sources of innovation (e.g. Busco & Quattrone, 2015, 2018). More precisely, our study illustrates how actors create multiple MCs to develop different ‘spaces’ serving distinctive functions that remain interrelated within the creative co-creation process. A ‘core MC’ creates space for recognizing, localizing, and developing new connections between items of knowledge, culminating in a new shared understanding of NPD, i.e. a synthesis. In addition, actors develop a ‘supplementary MC’ which stimulates actions and interactions on the micro-level that might challenge existing knowledge structures and authorities. In other words, it creates space for collectively creating an antithesis by developing exemplars of new products based on the synthesis.

Secondly, this study contributes to advancing the understanding of specific consequences of MCs’ interrelatedness for developing enabling control systems (e.g. Englund & Gerdin, 2015; Goretzki et al., 2018), product innovation (Carlsson-Wall et al., 2021), and MCs that form a MC system more generally (e.g. Friis et al., 2015). Regarding the enabling assessment of control, our results clarify how actors construct visual connections between different MCs forming a nexus of controls, and how these connections can influence the overall perception of the nexus as enabling. Therefore, we argue that neither the design characteristics of the core, nor the supplementary MC alone, can define their assessment as enabling or coercive. Instead, our results suggest that their enabling nature is determined by the dynamic material nexus that both MCs form and their fluid interplay.

These insights also contribute to prior research demonstrating how NPD activities can be managed by a hierarchically organized MC infrastructure (Carlsson-Wall et al., 2021), in particular by submitting that the visualization of the nexus of controls also plays crucial role when investigating the inequality of MCs and how they form a hierarchy for managing product innovation. Our insights therefore also complement previous research into the interrelatedness characterizing MC systems (e.g. Friis et al., 2015) by illustrating how the process of codification shapes the development of MCs’ interrelatedness. Specifically, we show how a MC system can arise through visual connections and the structural similarities of an MC system’s components.

Thirdly, our insights when investigating the codification of MCs forming a nexus of control also allow us to contribute to understanding of the relationships between epistemic¹ and technical objects (e.g. Baldessarelli et al., 2022). As we provide a case in which actors ground the technical on the epistemic object, but add multiple additional details, our results challenge the mutual exclusiveness of those two types of knowledge objects (Ewenstein & Whyte, 2009) by evidencing the transition of an epistemic into a technical object caused by mutual interaction and, therefore, supporting a relational view (Scarbrough et al., 2015) of knowledge objects.

The remainder of this paper is organized as follows. The next two sections outline the theoretical background to the study. Then, Section 4 examines the methodological approach employed and Section 5 presents the results from the in-depth single case study. Finally, Section 6 discusses the results of this study, and its contributions to existing research.

2. The Role of MC as Epistemic Objects in Creative Collaboration

NPD is a vital, but complex, organizational practice requiring collaboration between cross-functional² teams (e.g. Miura & Hida, 2004). Despite the benefits offered by teams comprised of actors representing different ‘thought worlds’³ (Douglas, 1987), problems can arise when engaging in ‘creative collaboration’ (e.g. Gephart et al., 2010). This is due to individual differences adversely influencing information sharing, interpretation, classification, categorization, and attribution (Bechky, 2003; Edmondson & Nembhard, 2009). The above challenges can be amplified in NPD settings, as creative collaboration requires more than simply the transference of knowledge across knowledge boundaries (Carlile & Rebentisch, 2003), demanding that actors ‘transform their own existing knowledge into new knowledge that complements and stimulates the knowledge transformation from others, in a process of mutual influence and collaborative emergence (Carlile, 2002, 2004; Hargadon & Bechky, 2006; Tsoukas, 2009)’ (cited in Majchrzak et al., 2012, p. 951). Within such settings, actors may need to challenge their own assumptions concerning how their interpretations of certain activities may be understood by actors operating in other thought worlds (Dougherty & Tolboom, 2008; Skilton & Dooley, 2010).

MCs play a crucial role in managing collective collaboration, i.e. a social process in which creativity occurs through dialectic negotiations and the integration of actors’ opinions and perspectives (Hargadon & Bechky, 2006). In general, this social process is supported by the organizational nexus of material arrangements (Schatzki, 2006)–such as MCs–in which it proceeds and that provides some organizational memory for it. More specifically, MCs support actors in their collaborative creative endeavors by supplying information (Davila, 2000), connecting and balancing NPD activities with wider organizational concerns (e.g. Carlsson-Wall et al., 2021; Davila & Ditillo, 2017; Mouritsen et al., 2009), generating dynamic tensions (Curtis & Sweeney, 2017), and integrating different knowledge sources (Ditillo, 2004). In this view, MCs are seen to have agency as they can affect the state of affairs in creative collaborations (Latour, 2005).

MCs can unfold this kind of agency as they are ‘epistemic objects’ (Cetina, 1997) which not only represent the ontology⁴ and epistemology of the actors who created them (D'Adderio, 2001; Knorr-Cetina, 1999), to support their conversational practices and generate knowledge (Bechky, 2003; Carlile, 2002), but also contain material aspects that frame the actions of people (Hutchby, 2001). Thus, MCs are knowledge objects, possessed of material instantiations, but consisting of ‘processes and projections rather than definite things’ (Knorr-Cetina, 1999, p. 6) and ‘representations of a more basic lack of object’ (Knorr-Cetina, 2001, p. 181), characterized by ‘irreducible vagueness’ (Rheinberger, 1997). The incompleteness (or indeterminacy) of MCs raises several questions with the potential to develop into avenues for further exploration and trigger a desire among actors to fill in these knowledge gaps (Busco & Quattrone, 2015; Knorr-Cetina, 2001). This infers that MCs must simultaneously be conceived of as ‘unfolding structures of absence, i.e. as things that continually ‘explode’ and ‘mutate’ into something else, and that are as much defined by what they are not (but will, at some point have become) than by what they are’ (Knorr-Cetina, 2001, p. 182).

While the material instantiations of MCs facilitate a perception of the differences between varying epistemologies characterizing different thought worlds, their incompleteness leaves sufficient space for negotiation (Busco & Quattrone, 2018). This means, MCs offer a visual performable space (i.e. a schematic form) that engages users by offering them a space in which different voices can be raised, as well as questions asked, and ideas critically reflected by the collective (Busco & Quattrone, 2015; Quattrone, 2017). This makes it possible to integrate knowledge arising from diverging thought worlds (Carlile, 2004; Nicolini et al., 2012). Accordingly, MCs perform by influencing how knowledge is ordered and compounded, rather than simply providing information, (e.g. MacKenzie, 2006), as well as being created and shared within groups (Turner & Makhija, 2006). Thus, MCs also control the specific relationships between individual actors or groups, establishing whose knowledge is relevant and whose is not (Turner & Makhija, 2006).

Understanding MCs as epistemic objects supports the shift of scholarly attention from the content provided by MCs toward their role in stimulating creative processes. Quattrone (2009) was one of the first to consider this issue, investigating how accounting’s reliance on images and visual impact (i.e. the relationship between accounting as a medium and its materialized form) informs its role as method to classify organizational thinking and knowledge. His study revealed that accounting creates a performable space by intertwining accounting texts and images, a framework, and a vision which can be reified by actors who practice accounting (Quattrone, 2009). Furthermore, Busco and Quattrone (2015) considered that accounting tools, such as the balanced scorecard, form ‘visual performable space[s] that is, a schematic visualization that engages users’ (Busco & Quattrone, 2015, p. 1246) by offering them space in which to perform creative work, i.e. imagining various meanings for abstract strategic imperatives (Czarniawska, 1997).

Thus, accounting acts more as a guideline or a framework, as opposed to precisely determining the actions to be undertaken. Similarly, Quattrone (2017) argued that spaces created by accounting must leave adequate room to embrace ambiguity and uncertainty, as a means of supporting the management of organizational tasks. This indicates that the inherent incompleteness of MCs, along with the resulting tensions, can encourage actors to strive continually for (unreachable) perfection (Busco & Quattrone, 2018). However, the process of searching itself fosters questioning and debate, and thereby creativity (Busco & Quattrone, 2018).

Similarly, Mouritsen et al. (2009) argued that management accounting calculations consist of a partial, rather than a total calculation of a firm’s values and affairs, particularly as calculations allow the mobilization of alternative propositions. This results in management accounting calculations acting as mediator between innovation and firm-wide concerns (Mouritsen et al., 2009). In addition, it prompts management accounting to add perspective to innovation and causes transformation rather than being a mere representation of innovative properties. Furthermore, Revellino and Mouritsen (2015) demonstrated that the calculative practices included in MCs can, of themselves, become an engine for innovation, by luring people into action as they track ongoing changes within an organization. As innovations are heterogeneous objects, which continuously adapt, tracing and visualizing previous and current courses of action helps actors envision future directions for development (Revellino & Mouritsen, 2015). However, visually tracing novel actions and outcomes can result in a need to reconfigure and transform MCs themselves, which can, in turn, transform them into engines for innovation.

Against this backdrop, the aim of this paper is to clarify how the spaces opened up by MCs are created and codified. More specifically, how actors develop and codify MCs, thereby creating performable spaces laying the groundwork for conversational practices and collective sense-making during the process of developing new products. In particular, it examines how the ability of emerging material instantiations of MCs already shape actors’ behavior and future interaction with those MCs, particularly in collective sense-making processes such as NPD.

3. Codifying MCs and Creative Synthesis

Following recent accounting studies focusing on the role of MCs in NPD (Busco & Quattrone, 2015, 2018; Quattrone, 2017), this paper understands MCs as collective epistemic objects creating performable spaces. To explain the creation of such spaces, we have chosen a knowledge codification perspective as a theoretical point of departure (Prencipe & Tell, 2001; Steinmueller, 2000). This allows us to focus on the role of MCs in dialectical processes, alongside the accompanying difficulties when integrating knowledge from different thought worlds into material representations of MCs.

In particular, this paper considers how actors engage in social interactions determined by their understanding of unique social contexts at a particular time (Bechky, 2003; Berger & Luckmann, 1966), when set against the background of prior experience (Jelinek & Schoonhoven, 1990), as well as how the nexus⁵ of material arrangements (Schatzki, 2006)–such as MCs–shapes such interactions. Accordingly, rather than black boxing them and treating MCs as ‘objects’, the knowledge codification perspective enables us to consider the social and material elements of MCs. This facilitates comprehension of how the resulting differences and sometimes inconsistent perspectives of actors are reconciled during the codification of epistemic objects, as they become transformed into a creative synthesis (Bechky, 2003; Harvey, 2014).

Codification is a process supporting ‘the inscription of knowledge in symbolic form’ (Cacciatori et al., 2012, p. 311). It extends beyond the mere knowledge articulation involving cognitive efforts associated with it, and including the screening of multiple alternatives, as well as establishing selection processes that can transform selected results into linguistic and symbolic representations of underlying phenomena (Prencipe & Tell, 2001; Zollo & Winter, 2002). More precisely, the initial stages of codification require actors from diverse thought worlds to build a common classification and categorization of basic information; i.e. so called ‘crude knowledge’ (Ancori et al., 2000, p. 266). This first step relates more to selecting what information to consider, rather than how it is selected (Bowker & Star, 1999).

Discussing and attaching meaning to crude knowledge is viewed as the second step in actors’ process of forming joint knowledge (Ancori et al., 2000). In this step, actors begin by establishing sense-making processes, in which they: (i) enact–i.e. create, rearrange, select, withdraw and delete some of the objective features of their surroundings; (ii) select–i.e. identify the cause-and-effect relationship that makes most sense regarding their individual perspective; and (iii) retain–i.e. memorize the successful sense-making of novel, crude knowledge (Weick, 1979). The second step does not infer a simple transmission of crude knowledge from one functional background to the other, but rather suggests a re-engineering process integrating different cognitive contexts (Ancori et al., 2000), thereby, reconciling opposing rationalities and understanding of situations. Accordingly, during the second step, actors are not only confronted with antitheses to their current theses (e.g. during enacting) but are also required to reach a synthesis (e.g. during selecting), to form a common understanding of developing new products. Attaching meaning to crude knowledge also infers creating a shared language with which to exchange information and reflect discursively on established practices and procedures (Cacciatori et al., 2012), against the individualized background of different actors (Ancori et al., 2000). Developing a common language requires actors to not only understand how to communicate knowledge (i.e. master codes and language), but also ways individual knowledge can be converted into collective, and reciprocal, knowledge (Ancori et al., 2000). Learning how knowledge is converted also means learning how and why different (and previously unrelated) pieces of knowledge are connected (Ancori et al., 2000), as a means to understanding how to manipulate existing knowledge (Carlile, 2002) and, finally, develop new patterns and capture new understandings to create new knowledge (Harvey, 2014).

Additionally, knowledge related to conversion is particularly relevant in situations in which actors originate from different thought worlds. This is due to the fact that differing language and codes (i.e. MCs) do not present a neutral mechanism of knowledge transmission (Ancori et al., 2000). Thus, language, codes, and visualizations include representations from different thought worlds (Ancori et al., 2000), and therefore codification can be used to represent knowledge, in particular the kind of knowledge deemed relevant for knowledge generation (D'Adderio, 2003). This also affects who attains authority over past knowledge, the process of generating new knowledge, and the organization of the work itself (Cacciatori, 2008; Goretzki et al., 2018), and, consequently, who can use knowledge effectively and how (D'Adderio, 2003).

Finally, common knowledge can offer a basis for developing differing kinds of ‘organizational reflexivity’ (Prencipe & Tell, 2001; Zollo & Winter, 2002). These have the potential to enable actors to learn collectively, i.e. ‘to enable individuals to rearrange, manipulate and examine symbols and symbolic relationships to transform the underlying knowledge represented in such systems’ (Prencipe & Tell, 2001, p. 1379). In other words, organizational reflexivity may lead to the development of an antithesis, challenging the current thesis, and thereby generating a creative synthesis, i.e. the recognition and development of complex connections between previously unrelated concepts (Harvey, 2014). This notion of synthesis is not a specific idea, rather it is a new collective understanding of what constitutes a new idea (Harvey, 2014).

In consideration of the above, we submit that an investigation into how actors codify knowledge in MCs has the potential to contribute to an understanding of how the process of creating performable spaces can lead to moments of ‘collective creativity’ (Hargadon & Bechky, 2006), culminating in a creative synthesis (Harvey, 2014; Van de Ven & Poole, 1995). This, in turn, can deepen the understanding of how MCs can contribute actively to creativity, rather than by simply reducing uncertainty. Figure 1 summarizes the theoretical framework for this paper.

Figure 1. Theoretical framework derived and adjusted from Van de Ven and Poole (1995).

4. Methods

4.1. Research Setting

We conducted an in-depth single case study between October 2011 and April 2013, to explore the structuring of NPD procedures at a manufacturing firm called LACUS (fictitious name). LACUS is a leading manufacturer of innovative premium and luxury kitchen and bathroom fixtures, employing over 3000 people at its production facilities located in Europe, the United States, and China. In 2012, it generated revenues more than €500 million. LACUS generally launches approximately 300 new products (each consisting of about 200 components) per annum, whose sales form up to 30% of the firm’s annual revenue.

4.2. Data Collection

Our data collection process was based on interviews, observations, and archival material. We conducted 27 interviews with 18 representatives from different backgrounds, to explore different corporate functions associated with NPD (see Table 1). All the interviews were recorded and transcribed verbatim, except for one early interview, during which notes were taken.

Table 1. List of interviewees.

#	Length (min)	Function
1	70	Member of Board of Directors
2	55/55/170/105	Management Accountant
3	10	Product Manager
4	80	Head of Product Planning (HoP)
5	65/30/55	Product Planning
6	55	Head of Construction
7	120	Head of Research & Development
8	65	Head of Product Management
9	40	Head of Product Management
10	95/65/55	Head of Product Planning
11	65/30	Head of Prototyping and Head of Working Team that developed map
12	60	Member of Board of Directors
13	60	Industrial Engineering (IE)
14	20	Product Manager
15	60/35	Head of Quality Management
16	45	Technical Service Center
17	55	Sourcing
18	60	Member of Board of Directors

The interviews focused on gaining an understanding of the processes involved in codifying and collaborating in NPD. After familiarizing ourselves with the firm’s NPD process, the interviews concentrated on exploring changes to the process being conducted to establish a novel structure for workflow, and collaboration based on the two new MCs. In addition to the interviews, one researcher attended several daily NPD work meetings to collect observational data (e.g. bi-weekly meetings at which representatives from different NPD projects discussed problems they have experienced arising within projects). In addition, the same researcher was given workspace on the NPD floor at the case study site, which enabled the observation of daily NPD work, and interactions with NPD participants. During this period, discussions were conducted at workstations, or in groups, providing us with additional contextual information regarding NPD processes and any potential changes. The observations enabled us to study interactions between NPD actors, to ascertain how they engaged in negotiations surrounding the restructuring of NPD processes. Finally, we aimed to reveal how actors codified their knowledge of newly developed MCs, and collated archival material regarding the change process itself, focusing on the development of the MCs in particular. This included taking PowerPoint charts, flip charts, and photographic evidence from workshops, in which actors negotiated changes to the NPD process. Furthermore, we accessed old and new MCs in relation to NPD. We acquired 22 different versions of the new process flow chart, from which we were able to reconstruct the development of the MCs and establish the types of knowledge codified, as well as how this was achieved. We also received three different versions of a self-developed XLS tool, and a final version complete with information concerning the progress realized as part of the NPD project. In addition, we conducted data analysis regarding function-specific MCs, as demonstrated in meetings and interviews. We gathered extensive field notes on function-specific MCs and how they were being used. Finally, we collected information concerning NPD from the firm’s intranet, to review announcements about innovative MCs in the staff magazine.

4.3. Data Analysis

After conducting initial interviews, we reviewed the interview transcripts, field notes, and archival documents, which we collated in an iterative manner, to identify topics associated with the NPD process (Miles et al., 2018). The researchers attended many case analysis meetings to read and re-read the data, as well as discussing the most remarkable aspects (Miles et al., 2018).

We followed an iterative process of data collection and analysis, moving back and forth between theory and data. We arranged key events within LACUS, using visual mapping to structure the data and employed illustration as the basis for a descriptive chronological narrative, as a way to create a foundation for the case story (Langley, 1999). The thick description of the narrative allowed us to examine events and negotiations as part of the case study contributing to the restructuring of the NPD process at LACUS. Following these discussions, we coded the data in simple thematic form, according to themes identified when discussing the narrative (Flick, 2018). To unravel our interviewees’ ‘emic perspective’ (Lukka, 2014), we sought detailed descriptions from the participants’ perspectives. Finally, in a workshop held at LACUS, following the data analysis, we cross-checked our results, clarifying our understanding of the codification, collaboration, and coordination procedures; the differences in the various versions of the MCs, and the connections identified between them; as well as how the MCs affected the creative collaborations.

5. Restructuring LACUS’s NPD Process by Codifying new MCs

5.1. Background to the Case–LACUS old NPD Process

As part of the traditional NPD process, individuals from eleven different corporate functions collaborated on a single NPD project, fulfilling a shared objective, i.e. the completion of a new product. LACUS’s NPD process had been relatively simple prior to 2011, when there had been far fewer projects developed simultaneously, and NPD projects were accomplished in an ad hoc and hands-on manner. LACUS possessed a formal MC in the form of a NPD manual to control the process, but was missing some activities, represented a ‘SAP-ontology’ (but SAP was never used for NPD), had a static format (i.e. unchangeable PDF file) and its length of 155 pages impeded actors from discovering connections between working packages across their functional boundaries, and, thus, the actors did not use it. At this time, the principal challenge concerned the integration of NPD and mass production at LACUS, because new products were expected to be of high-quality, incurring costs over a period of approximately two years, thus conforming to the typical time span LACUS’s competitors require to imitate its products.

The increased complexity of NPD (including additional products created within shorter periods of time, globally dispersed production, and using new materials), culminated in a project that ran out of control as a result of micro-cracks,⁶ and which ultimately became a game changer in the face of considerable financial losses. The failure of the project prompted the executive board to hire an external consulting firm to initiate the restructuring of the NPD process. After holding several workshops with all NPD contributors, the consultants’ primary advice was to restructure and formalize the NPD process by codifying knowledge by means of a ‘stage gate model’ (Cooper, 1990) and formulating clear criteria for subtasks for displaying the status quo of the NPD process, to generate a shared understanding of NPD at LACUS. LACUS’s senior management concurred with these recommendations, as did the actors involved, leading to the founding of several internal working teams, who were instructed to adopt the consultants’ ideas for subsequent adjustment according to LACUS’s specifications.

5.2. Codifying the ‘Core MC’ – Opening the Space for Synthesis

The members of the executive board assigned the Head of Prototyping (HoP) to develop the stage gate model, referred to by LACUS’s actors as the map. During the initial phase, he examined the original NPD manual, i.e. a 155-page PDF file characterized by textual description of working packages (i.e. a bundle of individual but related tasks) of the previous NPD. The HoP initiated to selecting which kind of information to include into the new NPD process by eliminating packages he considered repetitive or irrelevant. Consequently, he reduced the number of working packages from 250 to 90 which became the starting point for further enactment with NPD by the involved actors. HoP interpreted this significant reduction as an indication that knowledge relating to NPD was not generally available across the boundaries of the different involved corporate functions.

After creating a starting point for selecting what information to include, the HoP assigned a team–comprised of two members each from product planning, product management, and construction–to discuss, select, rearrange, withdraw, and attach meaning, as well as to visualize the new NPD and share the information with all NPD involved actors. HoP made a conscious decision to assign a cross-functional team to undertake this task, as he wanted them to ‘pull together properly from the beginning’ (HoP), to facilitate cross-functional sense-making. This was vital for restructuring the NPD due to the need to integrate all functions into NPD to secure an effective transition to mass production.

Following the advice of the consultancy, the project team started to make sense of the working packages using a very simple skeletal visual template of a stage gate model to rearrange, select, and create an order of the remaining 90 working packages displayed as ‘bubbles’ (see Figure 4 for the final version of the map). It was considered vital that the principal objective of visualizing the entire process and all tasks was presented on a single page:

Normally, you find huge manuals … but for us, it was essential to outline the entire process (which generally takes between one and two years), in a schematic form on a single page. That was very important to us … because the [product planners] required a clear schema … a clear process flow from us to be able to understand the connections [between the working packages] and take over the responsibility for a project … To achieve this, we started with a workshop in which we could reproach each other. Then, the project team reviewed the results and began to define new standards. (HoP)

The above reveals that the actors felt the need to ‘reproach each other’, meaning that they required a space in which to raise their different voices, challenge one another’s underlying assumptions, and critically reflect upon the current NPD as a collective. As a result of this discussion, they externalized their highly specific knowledge in a visually simple and compact way (i.e. on one page) to enable all the NPD involved actors to firstly create a common and cross-functional understanding of the overall NPD and underlying connections between individual tasks, and secondly, to facilitate retention of the agreed NPD process. This demanded the actors to identify both the shared and unshared aspects of the knowledge categories, followed by reflecting on, and selecting, which aspects of knowledge to include and omit. This led to identifying the cause-and-effect relationships making the most sense across different thought worlds. Additionally, this collective sense-making process prompted actors to introduce new categories of how to structure the entire NPD process.

We started from the end, meaning that we wanted a new product with a turnover of, let’s say, 120 million Euros. We subsequently tried to identify which steps would be necessary, and how we could order ‘the bubbles’ chronologically. We tried to identify different phases. Together with production, and all other [corporate] functions, we tried to complement and sharpen the bubbles that needed to be included, as well as identify how they should be ordered. Finally, we developed a phase model with design, technical conceptualization, construction, sourcing, pilot series, and serial production. (HoP)

The participants discussed, and agreed on, the six phases of NPD shown at the top of the map; how to re-organize the work and re-allocate the responsibilities for working packages (changing colors Figure 2 and y-axis on the left of Figure 4). This sense-making process was shaped by the skeletal visual template, as discussed by the actors:

[W]here to locate a particular working package [on the map]? Shall we locate it at the front, above, below, later? … Shall we keep it at all [on the map]? … There were interminable discussions over many days. (HoP)

Figure 2. Analysis of the changes of the map between versions 1–4. (Please see online version for colour interpretation)

The indeterminacy of the simple, skeletal stage gate model opened a space in which questions could be raised, definitions of objectives could be reviewed, and connections between different information ascertained and understood as well as new order and responsibilities of the NPD activities disputed. For example, actors debated the underlying assumptions of locating the bubble ‘sampling’. The product planner (PP) critically raised the question:

Must ‘sampling’ be located at the beginning or at the end of the sourcing phase? In the old NPD, it was located in the pilot series phase. (PP)

However, the Head of Application Technology found the conclusion of the sourcing phase as occurring too late, because:

For PPs, the new product development is done when the new product is launched but for us this is the point where our job actually begins. When a new product is launched, and becomes part of our product portfolio, we must guarantee further purchases up to 15 years. Therefore, we need to ensure spare parts are already available … We find it much more difficult if we have only start sourcing the spare parts once the product is launched. Therefore, we need to undertake the sampling and define all parts of the final product (including spare parts) as early as possible. (Head of Application Technology)

The above shows how tensions resulting from different rationalities unraveled because of the indeterminacy of the map.

To solve the tension and reconcile opposing views, the project team had to re-engineer the understanding of the NPD process and resulting responsibilities. Therefore, the project team introduced both bubbles (representing working packages) and lines (representing the overall responsibilities of a corporate function) (see Figure 4), along with the colors of the bubbles indicating the corporate function specifically responsible for each individual task. This allowed the project team to negotiate where to locate on the line indicating the wider area of responsibility of a function (e.g. quality assurance), while simultaneously showing a separate color (e.g. blue for ‘constructor’) indicating that another function was responsible for this specific task. Re-engineering the NPD processes and re-allocating responsibilities created a room for many controversies that led to intensive debates as the many changes of red color of the bubbles show in Figure 2.

Additionally, in reducing indeterminacy, actors were not only confronted with different rationalities requiring re-engineering into a new NPD process, but also the appearance of new knowledge. This means that connections between working packages became apparent across functional boundaries that have not been known before. To make this knowledge visible and, thereby, accessible for all members of the NPD, the project team integrated a further feature into the map: Boxes around working packages that indicate connections between working packages and subtasks [see light gray boxes in Figure 3(a) and Figure 4]. Juxtaposing working packages with other items and visually connecting them, allowed actors also to collectively uncover new cause-and-effect relationships which, in turn, questioned the current organization of NPD, whose knowledge is relevant, and, finally, who should be responsible for what. For example, IE critically reflected its consideration in general and its timing in particular:

Just when the new product was launched and should have been [mass-] produced, the [PP] realized the capacity would not be sufficient … but you must know how [production] lines are set up, to be able to understand that, you need 1.5 years sometimes to build them. (Head of IE)

Figure 3. (a) Third version of the map. (b) Excerpt from the tab ‘all working packages’. (Please see online version for colour interpretation)

Figure 4. Final version of the map (including milestone relevant working packages). (Please see online version for colour interpretation)

This indicates that IE challenged their assumptions about, and interpretations of, the NPD process, to ensure PP and the other functions realized the limitations of their local focus. Specifically, a conflict arose between IE and other NPD members concerning the criteria for assessing if and where new tools for production were developed, so impeding the development of a collective understanding concerning how to create new products:

[T]here are different cases if we develop a new product. If we have a new product where just one part (like the helve) is different, we agree with the PP that PP can still be in charge of utilities … However, we also have cases in which we have completely new products, and we do not know where we want to develop the tools to produce them, to be able to create new products with sufficient capacities … These are cases for which we need to intensify discussions with the PPs … and, finally, we have cases where we know right from the beginning that we want to produce 200,000 or 300,000 parts per year, which means we need completely different technologies. Here, it would be definitely inappropriate for the PP to decide in isolation. (Head of IE)

By confronting other functions with an antithesis, IE used the space that opened when visualizing the map to engage in a re-engineering of the NPD process and to inscribe its world view into the map. Overall, IE was able to successfully re-engineer NPD as it was not present at all in the first (abstract) version of the map, but in the final version it was responsible for seven working packages, of which four were milestones.

Due to the increasing cognition of the relevance of connections between tasks, actors also started to discuss how to measure the success or progress of a certain task because their connection with other tasks made different dimensions of the tasks and dependency more obvious than before (see also Figure 2 changes in milestones). Accordingly, the codification created tensions between the different actors because they disagreed about the assessment of tasks. For example, the task ‘selecting supplier’ was intensively debated by the PP responsible for shower pipes and a representative from Disposition. The following observation reveals how the tension stimulated a debate that unraveled opposing rationalities and understandings (Observation, June 2012):

You are responsible for selecting suppliers and deciding what, and how much, will be supplied by a single supplier. You assess and select suppliers based on [the measurement of] on-time-delivery. However, the problem we face is that the reliability of delivery [from our suppliers] is not good from my perspective. This means that our delivery time [to our customers] is, for no good reason, far too long for many new products, simply because not enough has been produced. This is particularly the case for shower pipes, which have been precisely developed according to the NPD plan, and increase in sales, but we are now experiencing delivery problems. (PP)

In response, the individual from Disposition responsible for selecting the suppliers argued:

My target figures are okay. I select and measure suppliers based on on-time-delivery. The supplier we are talking about is extremely good. (Disposition member)

The PP for shower pipes acknowledged her viewpoint, along with the responsibility held by Disposition, but countered that:

[On-time-delivery] is also important, of course, but delivery reliability must be good, too. Because, if you produce more shower pipes, you can sell more, finally leading to an increase in turnover. (PP)

To overcome the disagreement and reach a synthesis between both rationalities and decide on one measurement, the product manager involved LACUS’s board of directors to prioritize the different tasks:

I discussed this with our board, and they agreed with my point of view. So, Disposition [i.e. the person responsible for disposition] must change the way they measure and perform their task, in a way that focuses more prominently on delivery reliability. This is because, in the end, it is the sales figures that determine the success of a new product. (PP)

The person in charge of Disposition agreed with this decision because she also had the target to develop best-selling new products.

The above shows how codifying the map engaged users to raise their different voices and critically reflect their common understanding of NPD. Additionally, the quote also represents one example of the disagreements and disputes concerning the location, order, and responsibility of tasks, as reflected in many versions of the map developed over a relatively short period of time (see Figure 2 and Appendix for a detailed overview of these changes). For example, just five working packages were added between versions two and four of the map (see Figure 2, blue color), while there were more than 14 changes made to the identified responsibilities (see Figure 2, red color).

Re-engineering NPD was also accompanied by the need to create a shared language to not only exchange information, but also re-arrange actors’ own cognitive contexts into other cognitive contexts, thereby converting individual into collective knowledge. The project team, therefore, had to find a way to display the required data in a common language while taking care not to overload the map and contradict its synoptic nature. The team’s solution was to use an already existing visualization of the NPD, with which all functions agreed, i.e. an Excel file, in which the map was saved (see Figure 3a). To ensure further knowledge, the project team added a separate tab for each function (see Figure 3b for an example), while linking the relevant knowledge covered by the tabs with the one-page of the map. In addition, the project team pre-set a fixed classification scheme–i.e. the ontology of this shared epistemic object–for the tabs that should reflect a shared perspective, to overcome linguistic differences resulting from the divergent thought worlds, as well as create a common language, and support the synthesis of differing views, thereby demanding less function-specific ontology (see Figure 3b):

[I]t is difficult to clearly define it across all functions. It starts with how to describe the working packages linguistically … Some functions had real difficulties with that [i.e. descriptions] … They did not submit their descriptions for a long time because they were wondering ‘what shall I fill in?’. (PP)

The project team individually supported the functions when they had to fill in their descriptions to support the creation of a common language. While collecting the local information, the team also developed new tabs that included, for example, shared names for the tasks and subtasks, thereby introducing a common language.

This shared language enabled the actors to learn from each other by clarifying, completing, and extending their knowledge across different thought worlds, with the aim of developing a new shared understanding of NPD, i.e. a synthesis. For example, we observed the following:

12 actors from different corporate functions involved in NPD met … A designer and Head of Technology discussed the ‘push open technology’ of a new sink plug using a prototype during the meeting … The clicking mechanism is reversed in the current design. It works, but particles of dirt could accumulate, potentially leading to malfunction. The parties discuss whether they should change the mechanism prior to the market launch, as defined by the map, which would result in a delayed launch or if they change it during the production which would deviate from codified process and accompanied by overrunning the target costs. The head of innovation pointed out that should the clicker fail, customers could attribute the problem to the material (i.e. plastic) rather than to the mechanism, potentially resulting in negative spillover to other products containing plastic parts, which, though, play a significant strategic role for LACUS in the future. The Head of Innovation and a designer began discussing the issue, specifically the trade-off between design and reputation of plastic parts. As the expected, summarized profit of plastic parts clearly outweighs that of the single product with the inversed mechanism, the Head of Innovation was awarded the right to reverse the mechanism. The Head of Innovation was convinced that, despite delaying the product launch, this was the correct decision. (Observation, NPD meeting, July 2011)

The shared language introduced by creating the map opened a space for all NPD contributors for controversies that challenged the current status quo (‘shall we change the design that late in the process’). These dialectical negotiations enabled collective learning (‘plastic will be made responsible for malfunctions that result from a design decision’ and ‘plastic is a key feature of many future new products’) to reach synthesis affording a new understanding of NPD. The collective learning culminated in a synthesis (i.e. the final version of the map, as shown in Figure 4) upon which all actors agreed and referred to when discussing NPD. Therefore, we refer to the map as ‘core MC’, as it emerged from NPD practice and sustains cross-functional collaboration by forming the visual infrastructure for the new NPD understanding to which actors can collectively refer to and build upon.

Figure 5. (a) Toolbox – overview page (b) Toolbox – example tab.

Additionally, the synthesis was also reflected in the milestones of NPD (see also Figure 4). Thus, instead of translating function-specific criteria into financial milestones, to reduce the (function-specific) knowledge required for their understanding, this allowed original technical criteria to be retained, as a result of developing the common knowledge and synthesis surrounding the codification of the map.

5.3. Codifying the ‘Supplementary MC’ – Opening the Space for Antithesis

After the team had agreed on the final version of the map, it was tested as a tool for managing the NPD process with two prototype projects. After two months, the following conclusion was reached:

[W]e realized that the map is just a masterplan, which may outline how the NPD process proceeds, but is not a working tool. We were unable to use it for developing a single product … Therefore, we … developed ‘the toolbox’ [i.e. supplementary MC] to create new products. (HoP)

The map created a space in which the voices of different thought worlds could be raised, questions elaborated, and the general connections between previously unrelated pieces of knowledge recognized and developed. This was reflected by describing it as the masterplan. The incompleteness of the core MC created sufficient space for integrating and connecting actors’ different perspectives, as well as learning to develop a new understanding of NPD, in particular through dialectical negotiations, i.e. a synthesis about NPD. However, due to the room for interpretation created by the core MC, the actors’ precise actions for creative co-creation in NPD were not accessible at the intersection of the different functions. Specifically, the team had not fully considered the concrete function-specific connections between individual tasks in the templates stored in the tabs of the map, i.e. individual knowledge failed to be adequately converted into collective knowledge.

Therefore, the project team chose to develop a separate MC to supplement the map and allow for a more detailed integration of different cognitive contexts. The project team termed the supplementary MC ‘the toolbox’ (see Figure 5a), reflecting its role as a device supporting actors in accomplishing a specific task, rather than as a map offering a visual representation of the relative position of the different activities of NPD.

The project team commenced a series of interviews with all associated corporate functions, to comprehend the different cognitive contexts and establish associations between the concrete tasks across functions. Additionally, they invited interviewees to demonstrate and explain their processes and supporting local tools:

Developing the toolbox was possible … as [local] best practices were made accessible for all parties involved in NPD … However, we could not consider the hundreds of possible activities … So, we had to ensure that everyone has the same starting point, affording a clear overview of the status quo of the single new product, but [can] examine the relevant details. Therefore, we integrated and connected with local tools in the toolbox, starting with the map as basis. (PP)

This demonstrates that the project team decided to visualize the interrelatedness of tasks and ensure their compatibility, as well as integrating local contingencies, by building upon the latest version of the map and combining it with pre-existing function-specific MCs.

The landing page of the toolbox (see Figure 5a) affords a comprehensive, operative overview of the status of a concrete NPD project. The page is visually dominated by four separate elements: (1) A traffic light on the above right corner, which signals if the project is still on time (green), delayed but still with the potential to meet the planned deadline (yellow), or if the project is unable to meet the deadline (red). (2) A status overview concerning the different stage gate relevant working packages, with green tick or yellow respectively red bubble on the left-hand side. (3) On the above right-hand side, a risk map showing the current most relevant risks and their probabilities. (4) On the bottom, a cost overview related to the main parts of the new product, structured according to the differing stages of NPD. Such visual highlighting of these elements through the use of color (i.e. red) and symbols (i.e. traffic lights) signaling urgency, prompted actors to engage in a reflective dialog to find solutions, including for, for example, current time delays:

Suddenly, I had a red traffic light and if I had not solved the problem, I would have to postpone the product launch for three to four months … and the costs would be ‘striking’ … Therefore, triggered by the red traffic light and our analysis, we inquired specifically the construction [department] … the constructors offered no solution … two weeks later, we addressed the topic again because we left the task unsolved in the toolbox and the traffic light was still red. After confronting the constructors once more with this problem, [name of the responsible person] said that he had a possible solution to the problem … This was a great performance on the part of the constructors, but it was only due to the ongoing questioning, questioning, and questioning. (HoP)

Juxtaposing and linking different and sometimes opposing rationalities on the more micro-level in the toolbox, forced actors to find new solutions. In other words, the red traffic light and detailed display of the cause-and-effect relationships enabled actors to prompt specific other NPD contributors to develop an antithesis to the current synthesis to proceed in NPD process.

Although the landing page displayed the overall status of the project, it was created by summarizing the underlying function-specific tasks of the involved functions displayed in different tabs of the toolbox. These separate tabs include detailed monitoring of the status of function-specific working packages (such as start and end dates), the goals of the package, risk management and solutions, input from previous actors, etc. (see for an example Figure 5b). For instance, the status quo of ‘proof of function’ is created by an Excel-formula that consolidates all tasks and subtasks required for fulfilling this task. The corporate functions are responsible for keeping their tasks and subtasks’ statuses up to date and, thereby, the overall status for the entire NPD project. Because of the Excel formulas, it is possible for all members of NPD to identify who is responsible for potential delays or cost overruns. As a consequence of the links visualized between function-specific tasks, the assessment of the status quo of tasks is taken for granted for the functions, permitting the project team to not only integrate different individual ontologies, but also to facilitate learning across functions, in particular regarding the connections between different tasks as they were explicitly codified via formulas set out in the toolbox. Thus, the linkages in the toolbox converted individual into collective knowledge.

We combined the map with the local tools in the toolbox […] The relevant points in the stage gate [i.e. tasks] were interrelated with the toolbox. We constructed this form of interrelatedness [Excel formulas] to ensure people would understand how and why they should follow the order of the new [NPD] process. (HoP)

Thus, while the map supported learning regarding the relevant knowledge and responsibilities, in particular by opening a space for questioning and discussing (and so reaching a synthesis), the toolbox supplemented a shared operational impression by providing fixed instruments that demonstrated the individual steps accomplished during the NPD process, as well as prompting actors to monitor and deliver the expected outcomes (see Figure 5b). Therefore, we refer to the toolbox as ‘supplementary MC’.

The actors’ capacity to process cognitive content (e.g. assessing the risks of planned NPD activities) was also extended by visualizing the results of tasks:

If we are discussing ideas, I use my laptop to enter these directly into the [toolbox] during the NPD meeting … For example, when we discuss risks, I immediately enter them into the toolbox, because then one can directly assess the probabilities of their occurrence and impacts. This is actually a nice feature, because [before implementing the toolbox] I had to think of how to assess the risk, but now I can do it, and show it immediately during the meeting. (PP)

This infers that the physical materiality of the toolbox not only renders the cognitive process visible, but also transforms it into a creative resource transmitting creative impulses between MCs, by enabling actors to cross-functionally manipulate and visualize potential future outcomes if they realized their ideas. Accordingly, the toolbox opened a space within which actors formed different thought worlds to engage in a co-creation process, resulting in momentum promoting collective creativity, including leading to completely new ways of creating products, i.e. an antithesis to the currently existing synthesis (see the example of construction). However, this did not take place in every case. Some reassessments of the map against the background of the potential antithesis provided by the toolbox resulted in reorientations towards original synthesis, i.e. the map ‘absorbed’ impulses from local units. For example:

[D]uring our NPD meetings … we discussed functional risks, for example, because of the design … and the toolbox linked that with financial risks. So, we could see how new design ideas would influence our financial risk … Compared to the map’s suggested process, the summarized risk of the deviations led us to cancel some projects that were already in the initial procurement phase. (PP)

This demonstrates how creative impulses from new patterns for designing products were linked to the overall risk of the entire project. The real-time evaluation of financial risks results from the built-in connections between functions-specific tabs, the risk-reporting tab, and the overall risk matrix illustrated in the toolbox. More precisely, every function-specific tab includes the mandatory ‘risk management’ field (see Figure 5b). All subtask risks are consolidated in a working package risk that, in turn, is linked via formulas to the tab ‘risk reporting’, which is the basis for visual presentation of the top ten risks shown in the toolbox (see Figure 5a). Meanwhile the functions individually quantify the probability of functional risk but the PP complements together with the respective function the expected financial risk, which is calculated by adapting the costs at the single component level as calculated by the preceding function. The result for the expected financial value of the risk is then displayed in the adjusted ‘forecast’ in the toolbox (see Figure 5a on the middle right). Risks will be accepted if the forecast remains below the investment total or not if it exceeds the planned budget. Although the risk itself is sometimes disputed, actors rarely discuss the underlying assumptions or the systematic capture of risks. This is because the link between financial and functional risk is created based on the function specific assessment of the functional risk; i.e. on the rationality of the respective function’s own thought world which is taken for granted by them, and the joint assessment–PP and function–of adjusting the expected production costs. Integrating the diverse rationalities stemming from different thought worlds by connecting and summing them up via visible formulas to evaluate risks, led to relatively unquestioned mapping of the risks in the toolbox. Consequently, the connections visualized for the different risks, as well as the impact of combined risk, were able to reorient actors’ ideas towards a collectively shared understanding of the significance of NPD to the team, i.e. the existing synthesis.

6. Discussion and Conclusion

This study investigated the role of MCs created by actors while re-engineering the NPD process, using an in-depth single case study examining the fundamental NPD restructuring process alongside mobilizing ideas from the knowledge codification literature (D'Adderio, 2011; Majchrzak et al., 2012; Prencipe & Tell, 2001). By focusing on codification, rather than on the content, of MCs, this study offers several contributions to the understanding of the role of MCs as mediators and stimulators in collaborative creative processes (e.g. Busco & Quattrone, 2015, 2018; Quattrone, 2009, 2017; Revellino & Mouritsen, 2015).

Firstly, the study reveals how the creation and codification of MCs can serve as sources of innovation in creative settings. Prior research illustrated ways existing visualizations of MCs (i.e. budgets or a balanced scorecard) are able to create spaces in which actors can perform creative work, including linking various meanings to abstract concepts (Busco & Quattrone, 2015; Quattrone, 2017). Our study adds to this perspective by illustrating how the process of codifying MCs shapes actors’ creation of a common language of interrogation and synthesis that forges conditions for further engagement. More precisely, our study illustrates how actors create and visualize multiple MCs, to develop different ‘spaces’ to serve distinctive functions that remain interrelated within the creative co-creation process.

Thus, a ‘core MC’ creates space for recognizing, localizing, and developing new connections between items of knowledge. This space offers room for debates and disputes resulting from the different rationalities of the actors. During the course of the debates, the visual characteristics–such as lines and bubbles for distinguishing different task responsibilities–of the core MC leave adequate room for actors to integrate and connect their (potentially) diverging rationalities and develop a novel understanding of NPD through dialectical negotiations; i.e. a synthesis. By reflecting the new fundamental understanding of the process, the material characteristics of the core MC create common ground on which actors can perform creative work. While these results somewhat corroborate the findings of prior studies (Busco & Quattrone, 2015, 2018; Quattrone, 2017), our insights also reveal that actors can develop a ‘supplementary MC’ with material qualities that differ from the core MC. Compared to the core MC, the supplementary MC consists of material instantiations (e.g. symbols or colors) that prompt or prohibit precise actions, so shaping the collaboration between actors on the micro-level, rather than opening up a space for fundamental re-engineering.

By rendering more tangible the steps for creating an exemplar of a specific new product, the supplementary MC stimulates actions and interactions as well as controversies and tensions on the micro-level that are capable of challenging existing knowledge structures and authorities. In other words, it creates space for collectively reflecting and creating an antithesis by developing exemplars of new products based on a synthesis. These exemplars endow collective knowledge with structure and meaning (Weick et al., 2005) through action. As the supplementary MC maintains the codification of different thought worlds, it creates a space in which actors are exposed to diverse rationalities when solving problems and creating new products. This allows alternatives (including conflicting understandings) to emerge, gaining sufficient momentum to confront and engage the status quo (i.e. the thesis). In these cases, an antithesis might result in the construction of a new creative synthesis, or–if the antithesis does not gain adequate momentum–actors can maintain their thesis. Accordingly, the thesis evolves, as antitheses emerge revealing and shaping new understandings; i.e. creative syntheses.

We, therefore, argue that MCs, in creative contexts, have the potential to serve as repositories of cognitive structures, as well as facilitate more conscious engagement in cognitive operations, including the storage and retrieval of ideas, or the categorization of experiences typically carried out unconsciously (Stigliani & Ravasi, 2012). In this vein, reifying the experiences and ideas of actors in MCs transforms their mental content into ‘creativity resources’, which actors can manipulate, combine, and share with others more readily than in disembodied form (Gephart, 1993). Hence, our results demonstrate that MCs open up differing spaces constituting ‘an intricate, interlinking and overlapping mélange of representation and construction’ (Davison, 2010, p. 168), so influencing the outcome of collective sense-making processes and serving as engine for innovation.

Secondly, this study contributes to advancing the understanding of specific consequences of MCs’ interrelatedness for developing enabling control systems (e.g. Englund & Gerdin, 2015; Goretzki et al., 2018; Wouters & Wilderom, 2008), product innovation (Henri & Wouters, 2019), and the relevance of MC systems more generally (e.g. Friis et al., 2015; Grabner & Moers, 2013). Research into the design process of enabling control systems has revealed that other MCs can influence knowledge integration in development processes of these systems (Englund & Gerdin, 2015; Goretzki et al., 2018; Wouters & Roijmans, 2011; Wouters & Wilderom, 2008). While these studies indicate that a comparison with other MCs as a point of reference influences the assessment of controls as enabling, we emphasize the relevance of the ‘nexus of material arrangements’ (Schatzki, 2006) potentially formed by MCs for their assessment. More precisely, we provide a case study that demonstrates how actors construct connections between different MCs that form a nexus of material arrangements, and that these connections can influence the overall perception of the MC as enabling. In particular, we elaborate how two MCs contain different features and serve different purposes for enabling actors to perform their jobs better. The core MC enables collaboration between actors by codifying the current process and understanding of NPD, including locating the main tasks, clarifying responsibilities, and serving as a foundation on which actors can build additional MCs. Furthermore, the core MC simultaneously sets boundaries for cooperation by embodying the infrastructure for collaboration.

Therefore, we submit that the core MC primarily supports design feature global transparency, i.e. a comprehension of how and where local processes fit into the cross-functional overall process of enabling controls. Thus, the supplementary MC facilitates individual actions by outlining the required input by other actors, aims of the task, and criteria for compliance. Consequently, the supplementary MC primarily features internal transparency, i.e. an understanding of the functioning of local processes. Due to its material characteristics, it also prompts actors to find new ways of resolving emerging problems. However, it is neither solely the core nor the supplementary MC which enables actors to perform their jobs better. Rather, it is the dynamic material nexus formed by both MCs that supports a collaborative co-creation of new products requiring a fluid interplay between individual actions and collaboration.

Against this background, we have deepened the insights of Goretzki et al. (2018), who have demonstrated the influence of local MCs during the development of (global) enabling controls by providing a case study highlighting the way different MCs forming a nexus of controls, which then become the reference point for an assessment that is either enabling or coercive. Accordingly, we argue that actors assess the enabling nature of controls, based not only on their participation during controls’ development process (e.g. Englund & Gerdin, 2015; Wouters & Roijmans, 2011; Wouters & Wilderom, 2008), available reference points for them during the development process (Goretzki et al., 2018), or MCs’ design characteristics (Jørgensen & Messner, 2009). However, actors also consider the nexus of material arrangements of which the MC is a part when assessing it as enabling or coercive control. Consequently, our results suggest a more dynamic and conditional assessment of controls following their development process, because their enabling nature is not only grounded in relatively stable design features or a finished design processes, but is also conditional on their dynamic interplay with other MCs in the relevant nexus of controls.

The relevance of the interdependence of MCs has recently also been emphasized for product innovation (Carlsson-Wall et al., 2021; Henri & Wouters, 2019). In terms of the content provided, Henri and Wouters (2019) demonstrated that, depending on the level of uncertainty, MCs can prove complementary and trigger product innovation, or become substitutes and impede product innovation. Our results intensify content-centric insights by explaining how visual links can create additional connections between MCs that extend beyond content. In addition, we demonstrate that visual connections between different MCs can transmit creative impulses. In particular, visualization of how the actions of one organizational function influence the actions and outcomes of others delivers important implications for organizational creativity properties. The ability to visualize these connections ensures that corresponding knowledge is accessible at the intersection between separate functions, satisfying actors’ requirement to ‘see’ their thoughts and logical connections (Carruthers, 2000), so stimulating ‘non-verbal thinking’ (Ferguson, 1977). Consequently, visual aspects of MCs not only support organizational reflexivity (Prencipe & Tell, 2001; Zollo & Winter, 2002), but also construct the foundation for moments of organizational creativity. We, therefore, submit that the visual connection between MCs in product innovation can support creativity by transmitting impulses, in addition to content-related connections.

In this vein, our insights add to those of Carlsson-Wall et al. (2021) who demonstrated that NPD activities can be managed by a hierarchically organized MC infrastructure governed by a management control anchor practice. Our results corroborate their finding that some MCs seem to be more guiding than others, but we further elaborate the dynamic relationship between the different components of the ‘management control infrastructure’ (Carlsson-Wall et al., 2021). In theoretically departing from a knowledge codification perspective, we have developed the argument that actors create the core and supplementary MCs, and their connections, in a way that ensures creative impulses are prompted, transmitted, and stored across functional boundaries. Therefore, actors, while explicating their knowledge, choose a codification that both allows them to extend their capacity to process cognitive content and also ensures it become accessible at the intersection of functional boundaries throughout the NPD process. Against this background, we submit that the codification of the nexus of controls (or ‘management control infrastructure’ as termed by Carlsson-Wall et al., 2021) also plays crucial role when investigating the inequality of MCs and how they form a hierarchy.

In view of the above, our insights also complement previous research investigating the interrelatedness between MC components forming MC systems more generally (Davila, 2005; Davila et al., 2009; Friis et al., 2015), by illustrating how the codification process shapes the development of MCs’ interrelatedness. In addition to a connection established by means of content (e.g. Bedford, 2015; O’Grady et al., 2016), interrelatedness between MCs–which characterizes them as a system (Grabner & Moers, 2013)–can also arise through visual connections and the structural similarities of the MC system’s components. Specifically, our findings suggest a chronology in the creation of MC interrelatedness. At the outset, actors establish and visualize an ‘infrastructure’ (Nicolini et al., 2012) for their collaboration (i.e. the core MC), by creating a common classification of knowledge and language, as well as a general understanding of how to organize their work. This essential codified infrastructure not only sets boundaries for collaboration, but also serves as the foundation to enable actors to build additional MCs. More precisely, actors devise a common language with which to inscribe their specific knowledge into additional MCs based on already codified categories. Consequently, the structural elements and material characteristics of the first MC spill out onto additional MCs. Therefore, the material characteristics (in addition to the content) of the infrastructure-building scope of a MC influence its subsequent course during their development. Thus, it is vital to consider the process of establishing single MCs and their mutual influence when creating them individually, to understand interrelatedness within a MC system, as well as its information processing properties.

Thirdly, our insights derived as a consequence of investigating the codification of MCs forming a nexus of material arrangements allow us to contribute to understanding of the relationships between different types of knowledge objects (e.g. Kertcher & Coslor, 2020; McGivern et al., 2018). Therefore, we have departed from the theoretical understanding that MCs can be interpreted as epistemic objects triggering actors to fill in the knowledge gaps (Busco & Quattrone, 2015, 2018; Quattrone, 2017), instead providing evidence of a particular MC (i.e. the supplementary MC), which possesses additional features, which classify it as a ‘technical object’. Our results show how actors, while striving to perfect the epistemic object (i.e. the core MC), developed a new tool (i.e. the supplementary MC), which can be classified as a technical object. More precisely, actors grounded the technical on the epistemic object contributing numerous additional details, connections, and graphical elements, capable of determining a precise frame for the objects of inquiry (in our case, developing new products). Thus, our results challenge the mutual exclusivity between these two types of knowledge objects, as propagated in prior studies (Ewenstein & Whyte, 2009) to some degree by providing evidence for a transition of an epistemic into a technical object. Specifically, we demonstrate that transition of a knowledge object from an epistemic to a technical object cannot only be caused by changes in its use and interpretation (Bijker et al., 2012) but also through its interaction with another knowledge object. Therefore, we contribute to the relational view of knowledge objects (e.g. Scarbrough et al., 2015) and submit that considering the way knowledge objects are created in relation to other objects, i.e. their nexus of material arrangements, is critical for establishing their influence on knowledge creation and classification as alternate types of knowledge object.

In this vein, we also observed technical objects embody a wide range of knowledge in a mobile and easy to share way, supporting actors when mastering their daily tasks and facilitating shared understanding. That is, technical objects stimulated actions and interactions on the micro-level that challenged existing knowledge structures and authorities codified in the epistemic object. Thus, although it generates answers about the epistemic object that reveal what ‘one does not know yet’ (Nicolini et al., 2012, p. 614), the technical object, simultaneously, fostered a lack of completeness of the epistemic object. Therefore, we extended the insights about the general role of technical objects in knowledge creation (e.g. Baralou & Tsoukas, 2015; Nicolini et al., 2012), specifically, in cross-functional settings (Kaplan et al., 2017), by evidencing that technical objects can submit creative impulses to the epistemic object, compelling it to change and thereby evolve. Thus, we propose that it is not only the lack and incomplete nature of MC as an epistemic object (e.g. Busco and Quattrone, 2018; Nicolini et al., 2012), but also its relationship with the MC that features characteristics of a technical object, which plays an important role in knowledge development. Accordingly, we complement the literature on epistemic objects (e.g. Baldessarelli et al., 2022; Boxenbaum et al., 2018) by demonstrating how the interaction between epistemic and technical objects can generate creativity, knowledge, and learning.

Aside from these contributions, this study has introduced several limitations that we consider could be subsequently addressed in future research. Although the process of codification might increase organizational reflexivity, regardless of whether or not the outputs of knowledge codification are used (Cacciatori et al., 2012; Prencipe & Tell, 2001), it was impossible to study the two MCs beyond the first implementation over a longer timeframe. However, organizations not only learn by codifying knowledge, but also by implementing, replicating, and adapting that knowledge (Prencipe & Tell, 2001). Therefore, it could prove beneficial to examine the cognitive implications of codified shared knowledge in terms of how to better perform innovative tasks such as NPD, i.e. how the outcomes of the codification process also afford input for future projects.

Furthermore, the influence of the consulting firm accompanying the initiation of the NPD change process was marginal. However, consulting firms can potentially have significant impacts on change processes through the provision of knowledge, serving as a flexible resource, or task- and goal focused agents (e.g. Nippa & Petzold, 2002).⁷ Therefore, we submit that investigating a case in which the influence of a consulting firm is more palpable than in our case could complement the insights set out in this paper.

Additionally, we may have underemphasized the role of informal controls and actors’ personal relationships as factors controlling complex NPD settings in this study, due to selecting knowledge codification literature as a theoretical point of departure. However, it appears to be a reasonable assumption that cross-functional knowledge management is simpler if the thought worlds of those functions share common ground (for example, industrial engineering and manufacturing might share certain technical expertise). Accordingly, it could be interesting to study cases during which there is a variation in collaborative functions.

Moreover, the role of disputes and debates for collective learning has not been as prominent in our case as it has been in prior studies (e.g. Chenhall et al., 2013). The reason for this could be that MCs have developed from and as such are connected with function specific tools that reflect the taken-for-granted rationality of the respective function’s own thought world. Therefore, we submit that future research might investigate how variations of this aspect might influence the relevance of disputes and debates in cases of collective learning.

Finally, the results suggest a positive impact from the joint development process of formal MCs, one that outweighs the cost of joint development, so rendering it a superior option. However, the lesser the knowledge asymmetry between the different actors and the designer of the MC, the lower the requirement for knowledge explication, and the less beneficial a joint development of MCs. Therefore, it could prove beneficial for future research to investigate how actors evaluate the costs and benefits of joint development in a variety of task environments.

Acknowledgements

We acknowledge the very constructive feedback of the editor and two anonymous reviewers, the participants of the EAA 40th Annual Congress, and the 10th ENROAC Conference. Stefanie Malz gratefully acknowledges funding from Konrad Adenauer Foundation.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental data

Supplemental data for this article can be accessed on the Taylor & Francis website, doi:10.1080/09638180.2022.2082994

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1 Epistemic objects are (knowledge) objects of inquiry and pursuit. They are defined at once by what they are and what they are not, i.e. a lack of completeness; technical objects, instead, provide a frame for the objects of inquiry. They are defined by a closed and ready to be used form (e.g., Ewenstein & Whyte, 2009).

2 We use the term ‘cross-functional’ to refer to ‘individual’s specialized knowledge from his or her past or current intrapersonal experience or tenure diversity and his or her intrapersonal educational discipline diversity.’ (Majchrzak et al., 2012, p. 966)

3 A thought world can be seen as a community of actors engaged in a specific activity, who have a common understanding about that activity, which provides internally consistent cause-and-effect relationships arising from analytical criteria and priorities (Dougherty, 2001).

4 Ontology means the categorization that reflects an actor’s conceptualization of her/his cognitive context. Epistemology means how actors produce knowledge about their specific context from their point of view (Knorr-Cetina, 1999).

5 Although the idea of the ‘nexus of material arrangements’ shares some similarities with an ANT perspective, it conceptualizes material arrangements rather as structure that surrounds the interaction between human actors and does not treat material arrangements as non-human actors.

6 The function ‘application technology’ was responsible for the micro-cracks as it did not test the pilot series under batch or mass production conditions because of time delays in prior steps of the old NPD.

7 We thank one of the two anonymous reviewers for this idea.