Innovate the Innovation : An exploratory study about why digital innovation units are being closed and have to constantly reinvent themselves

MASTER THESIS WITHIN: Business Administration NUMBER OF CREDITS: 30

PROGRAMME OF STUDY: Digital Business (M.Sc.) AUTHORS: Lea Sophie Lindemann and Marcel Wiegand JÖNKÖPING May 2021

Innovate the Innovation

An exploratory study about why digital innovation units are

being closed and have to constantly reinvent themselves

Acknowledgments

First, we would like to thank our interviewees for their support and the time they took for us, and especially for their openness towards such a sensitive topic and the valuable insights they provided. Without them, our research would not have been possible.

Second, a special thanks to Infront Consulting & Management GmbH as our partner company, especially to Dr. Felix Lau, Lea-Theresa Münch and Thomas Sindemann. Without their knowledge, expertise and network, the thesis would not have been at the same level as it is today. Additionally, we want to thank Felix for reviewing our thesis and always believing in us and Infront for giving us the possibility to publish a practical study from our results.

Third thank goes to Ryan Michael Rumble, our thesis supervisor, for the great and helpful supervision, for the constructive thesis seminars and tutoring sessions and especially for his always motivating manner, guiding us throughout the thesis process.

Last but not least, we want to thank our friends for the great friendship throughout our studies, the mutual support and for the continuous encouragement of each other.

Thank you – Tack så mycket – Teşekkürler – Gracias – Vielen Dank

for wonderful two years of study!

Lea & Marcel May 24, 2021

Master Thesis in Business Administration

Title: Innovate the Innovation – An exploratory study about why digital innovation units are being closed and have to constantly reinvent themselves

Authors: Lea Sophie Lindemann and Marcel Wiegand Tutor: Ryan Michael Rumble

Date: 2021-05-24

Key terms: Digital Innovation, Corporate Innovation, Ambidexterity, Corporate Entrepreneurship, Digital Innovation Units, Organizational Failure

Abstract

Background: The advancement of digital technologies and hence the emergence of new

opportunities but also changing customer needs force incumbent firms to increase their innovation capacities. Therefore, they must find ways to explore the opportunities of digital innovation, while simultaneously exploiting their core business and ensuring a secured revenue stream. This phenomenon, called organizational ambidexterity, can amongst others be achieved by implementing a digital innovation unit (DIU). DIUs, structurally separated units, especially within Germany, Switzerland and Austria are declared to be successful instruments to increase incumbent’s innovativeness but also to foster an innovation culture and advance the corporate digital transformation while the core organization can focus on core business activities.

Research Problem: Despite the prevalent perception of DIUs being successful initiatives to

find remedy for incumbents’ conflict of being innovative while simultaneously advancing the core business, in recent years more and more DIUs were closed. This development is particularly interesting as the amount of DIUs is still increasing while on the other hand especially business press claims the end of these units. However, regarding this paradox, academic literature is still rare and has until now primarily focused on the challenges of DIUs but not on their closure. Correspondingly, DIU closure poses a significant research gap.

Research Purpose: The study primarily aims to clarify why DIUs are being closed and derive

a general theory explaining this phenomenon. Above that, a secondary purpose is to advance knowledge regarding DIUs as a digital innovation instrument and what characterizes a DIU. Finally, we aim to examine whether a closure can be equated with a failure.

Research Question: Why are Digital Innovation Units being closed?

Research Method: Qualitative, inductive research; Ontology – Relativism; Epistemology –

Social Constructionism; Methodology – Grounded Theory; Data Collection – 14 semi-structured in-depth Interviews; Sampling – Purposive, Snowball, Theoretical; Data Analysis – Grounded Analysis

Conclusion: Resulting from the findings, a model was developed, that explains five major

statements related to DIU closure. On the one hand, we ascertained distinct symptoms that anticipate a closure such as personnel measures and budget cuts. On the other hand, we identified factors that significantly influence the fate of a DIU such as corporate politics or cultural dissonances and identified remnants of closed DIUs. Finally, the immediate reasons why DIUs close can be related to strategic changes within the parent company and to how the DIU is evaluated.

Table of Contents

1

Introduction and Background ... 1

1.1 Background ... 1

1.2 Research Problem ... 2

1.3 Research Purpose ... 3

2

Literature Review and Frame of Reference ... 5

2.1 Dimensions of Innovation ... 6 2.1.1 Incremental Innovation ... 6 2.1.2 Sustaining Innovation ... 6 2.1.3 Radical Innovation ... 6 2.1.4 Disruptive Innovation ... 7 2.2 Organizing Innovation ... 7 2.3 Digital Innovation ... 8 2.4 Organizational Ambidexterity ... 9 2.4.1.1 Contextual Ambidexterity ... 9 2.4.1.2 Structural Ambidexterity ... 10

2.5 Digital Innovation Units ... 10

2.5.1 Embodiment of Structural Ambidexterity ... 10

2.5.2 DIUs – Origin and Examples ... 11

2.5.3 DIUs – Terminology ... 12

2.5.4 DIUs – Characteristics and Design Dimensions ... 13

2.5.5 DIUs – Design Typologies ... 14

2.5.6 Conclusion and Own Definition ... 15

2.6 Organizational Failure ... 16

2.6.1 Failure of Established Companies ... 16

2.6.2 Failure of Startups ... 17 2.6.3 Failure of DIUs ... 18 2.7 Theoretical Framework ... 19

3

Research Method ... 21

3.3.1 Data Collection ... 25 3.3.2 Sampling Strategy ... 26 3.3.3 Interview Design ... 27 3.4 Data Analysis ... 30 3.5 Research Quality ... 31 3.6 Ethical Implications ... 33

4

Empirical Findings and Analysis ... 36

4.1 Reasons related to Strategic Change within Parent Company ... 39

4.1.1 Top Management Change ... 39

4.1.1.1 Sponsor Departure ... 39

4.1.1.2 CEO Change ... 40

4.1.2 Reactive Strategic Change ... 40

4.1.2.1 Covid-19 Pandemic ... 40

4.1.2.2 Cashflow Difficulties ... 41

4.1.3 Proactive Strategic Change ... 41

4.1.3.1 Focus on Core Business ... 41

4.1.3.2 Periodical Strategic Realignment of Parent Company ... 42

4.2 Reasons related to DIU Evaluation ... 43

4.2.1 DIU Performance Evaluation ... 43

4.2.1.1 Contextual Perception of DIU Failure ... 43

4.2.1.2 Lack of explicit Objective ... 44

4.2.1.3 Lack of Commercialization Success ... 45

4.2.1.4 Strategic Misfit of Innovations ... 45

4.2.1.5 Inadequate Financial Measurements ... 46

4.2.2 DIU inherent Structural Determinants ... 47

4.2.2.1 Designed to be Closed ... 47

4.2.2.2 Disproportion of Goal and Resources ... 48

4.3 Contributing Factors to DIU Closure ... 48

4.3.1 Corporate Peculiarities ... 48

4.3.1.1 Corporate Politics ... 48

4.3.1.2 Leadership Failure ... 49

4.3.1.3 Culture Clash ... 50

4.3.1.5 Internal Competition ... 51

4.3.2 Strategic Realignment of DIU Goals ... 52

4.3.2.1 DIU changed towards Cultural Enabler ... 52

4.3.2.2 Expectations Management ... 52

4.4 Symptoms anticipating DIU Closure ... 53

4.4.1 Personnel Measures ... 53

4.4.2 Lack of Budget for Innovation Projects ... 54

4.4.3 New or unrealistic Objectives ... 54

4.5 Components surviving DIU Closure ... 54

4.5.1 Reintegration of DIU Components ... 54

4.5.2 Spin-Off into New Venture ... 55

4.6 Conclusion of Findings ... 56

5

Discussion ... 59

5.1.4 Success Measurement for Digital Innovation ... 64

5.1.5 Dependence of Parent Company and its Strategy ... 65

5.1.6 Organizational Failure ... 66 5.2 Managerial Implications ... 67 5.2.1 Conceptualization ... 67 5.2.2 Operation ... 67 5.2.3 Closure ... 68 5.3 Social Implications ... 68 5.4 Conclusion of Discussion ... 69

6

Limitations ... 71

7

Future Research ... 72

Reference List ... 74

Appendix ... 86

List of Figures

Figure 1 Taxonomy of DIUs ... 14

Figure 2 Coding Tree - Part 1 ... 37

Figure 3 Coding Tree - Part 2 ... 38

Figure 4 Comprehensive Framework for DIU Closure ... 58

List of Tables

Appendix

Appendix 2 Research Teaser ... 87

Appendix 3 Interview Topic Guide (Initial Version) ... 88

Appendix 4 Interview Topic Guide – Expert Interviews ... 89

Appendix 5 Interview Topic Guide (Revised Version) ... 90

Appendix 6 GDPR Participant Consent Form ... 91

List of Abbreviations

BU Business Unit

CC Corporate Culture

CDI Corporate Digital Innovation

CDO Chief Digital Officer

CEO Chief Executive Officer

DI Digital Innovation

DIU Digital Innovation Unit

GA Grounded Analysis

GT Grounded Theory

HR Human Resources

IC Innovation Culture

INT Interviewee

ITG Interview topic guide

KPI Key Performance Indicator

NIHS Not-Invented-Here-Syndrome

OrFa Organizational Failure

PaCo Parent Company

PARC Palo Alto Research Centre

R&D Research & Development

ROI Return on Investment

1 Introduction and Background

______________________________________________________________________ The purpose of this chapter is to lay the contextual foundation of the research by illustrating the challenge for incumbent firms to be innovative, becoming ambidextrous for that purpose and employing digital innovation units to implement this concept. Additionally, the demonstration of the research problem and associated research gaps leads to the research purpose and the resulting research question.

______________________________________________________________________

1.1 Background

The innovation capacity of established corporations is today seen as a focal determinant for improving a company’s competitiveness (Ferreira et al., 2019). One instrument to increase innovativeness, especially utilized in the region comprising Germany, Switzerland, and Austria (hereinafter referred to as DACH-region), is the digital innovation unit (DIU) (Infront, 2017). While startups focus on building radical or disruptive innovations, incumbent firms are more successful in continuous innovation activities to serve existing customers. However, this ability withholds incumbents from exploring new opportunities through radical or even disruptive innovations and increases their risk to underestimate the threat of disruption (Christensen, 2013). Accordingly, incumbent firms have two main strategic goals. First, incumbents are striving to exploit their core business through continuously developing existing products and services. Second, incumbents intend to explore new opportunities or markets by developing radical innovations enabled by digital technologies, which simultaneously reduces the risk of being disrupted by new market entrants (O'Reilly III & Tushman, 2013). The first goal is achieved through efficiency projects and incremental innovation. The second goal is addressed by systemically facilitating explorative activities (Birkinshaw & Gibson, 2004). The challenge of targeting both strategic goals, lies in the difference of the required approaches to address them (O'Reilly III & Tushman, 2004). Structurally ambidextrous organizations are characterized by the separation between innovation and core business implemented within the organizational structure (O'Reilly III & Tushman, 2013; Birkinshaw & Gibson, 2004). To achieve structural ambidexterity, companies employ DIUs, which focus on e.g., breakthrough technologies and market-oriented digital innovation (DI). DIUs are widely adopted in the DACH-region by large and

medium-sized corporations (Holotiuk, 2020). The term DIU represents different types of separated entities, which have different names according to their functioning such as corporate accelerators, corporate incubators, and digital labs. Presumed a focus on DI, these corporate units can all be grouped into the collective term DIUs (Infront, 2017). According to industry reports, around 250 DIUs exist throughout the DACH-region (Infront, 2019) and 29 out of 30 companies from the German stock market index DAX employ on average three of them (Infront, 2019; BCG Digital Ventures, 2019).

1.2 Research Problem

BCG reports that since 2012 the amount of different innovation vehicles such as DIUs is rapidly increasing (BCG Digital Ventures, 2019). DIUs are claimed to successfully foster corporate digital innovation (CDI) (Infront, 2020; Svahn et al., 2017). Despite this trend, in the DACH-region the first DIUs are already being closed or sold by the parent company (PaCo) (Buchenau, 2020; DHL, 2021). Other units are strategically realigned (BCG Digital Ventures, 2019; Unger, 2018). Building on the claim that DIUs are successful instruments to foster CDI, it should be investigated why despite this wide acceptance of DIUs as DI enablers within the DACH-region, the number of DIUs that are being closed is also increasing. Despite reports of significant contributions from DIUs (Infront, 2019), their impact is beginning to be questioned and some business magazines even daringly proclaim the failure of DIUs (Kyriasoglou, 2020; Capgemini, 2015b). Beyond the ambiguity of DIU perception and the paradox of the increasing number of closed and newly opened DIUs, DIUs have only recently received wider attention by scholars and academic literature on DIUs is still rare, highlighted by no mentions of DIUs in peer-reviewed journals (Holotiuk, 2020; Raabe et al., 2020a). Accordingly, there are multiple research gaps, ranging from an established definition of the term DIU to a consensus regarding taxonomy and typology for DIUs. Despite this underrepresentation in academia, the wide adoption within the DACH-region justifies further research of DIUs. This holds especially true, as DIUs attempt to contribute to CDI efforts which pose an increasingly important challenge for established organizations (Nambisan et al., 2017). Their perception as panaceas to fuel the digital transformation strategy further requires an examination of whether DIUs have the projected impact or disappoint in achieving expected results (Kleske et al., 2016). DIUs therefore represent a new and relevant research stream for corporate innovation research and for corporate entrepreneurship

scholars, alike. While popular studies such as INFRONT (2020),ETVENTURE (2020)and

BCGDIGITAL VENTURES (2019)investigate similar phenomena, their focal subjects are

not yet aligned with academic literature. Research and reports on DIUs highlight their multitude of challenges, which can eventually cause their failure (Raabe et al., 2020b; etventure 2020a). Some innovation experts assume that between 80% and 90% of DIUs fail (Capgemini, 2015a). Potential reasons for their failure range from a lack of communication, missing support of the PaCo, inappropriate skillset of employees, to the absence of financial resources (Capgemini, 2015a; Raabe et al., 2020b). RAABE ET AL.

(2020b)furthermore constitute for two DIUs, that they “were closed because their leads left the unit” (p. 5). This exemplary quote and reports on DIU failures show that DIU failure and closure are often used synonymously and that hence failure is seen as the main reason for the disappearance of DIUs (Unger, 2018; Raabe et al., 2020b). However, failure in organizational contexts is difficult to measure (Mellahi & Wilkinson, 2004) and, moreover, has not yet been empirically validated as the main reason for DIU closure. On the other hand, the closure of DIUs is a clearly delimited event that is even reported by DIUs themselves (DHL, 2021). Accordingly, we see a significant research gap regarding the reasons behind DIU closure.

1.3 Research Purpose

Looking at this research problem, we derive a set of purposes for our study. First and foremost, this research attempts to develop an understanding for why DIUs close. We want to provide knowledge for the striking paradox between an increasing number of closures, and the proclaimed significance of DIUs for the CDI efforts of established companies. By researching the reasons behind the closure of individual cases and deriving a general theory explaining the reasons for DIU closure, we aim to advance the understanding of DIUs in general and of their significance for CDI. Accordingly, our main Research Question (RQ) is:

“Why are Digital Innovation Units being closed?”

The contribution of a theory explaining the closure of DIUs is relevant for practitioners and scholars alike, as the reasons can be representative for the question whether DIUs are as promising as proclaimed and whether they achieve their objectives. Simultaneously, the research of DIU closure explores factors influencing how the DIU evolves over time. With our main purpose of creating knowledge about the reasons for DIU closure we aim

to add nuance to the discussion about DIUs, especially regarding the common notion that closure equals failure. This contribution can eventually result in reframing the perception of DIUs and their impact. We furthermore aim to contribute to existing knowledge of how DI can be institutionalized into existing organizational structures. From a practical perspective, this study aims to show sponsors and leaders of active DIUs, why DIUs are being closed and how they can strategically circumvent this development. For agents within DIUs, we want to provide recommendations for actions they can take to prevent their DIU from closing. Finally, for top management within established companies, we intend to provide recommendations for how they can nurture their DI efforts, by learning from the mistakes and pitfalls of others.

2 Literature Review and Frame of Reference

______________________________________________________________________ The purpose of this chapter is to present the state of our knowledge and build the theoretical foundation to the topic of DIUs. More specifically we aim to understand what DIUs are, what their purpose is and to frame what has been done so far within the field of DIUs. Furthermore, this part sheds light on potential reasons for DIU closure by summarizing literature’s suggestions for main reasons of organizational failure.

______________________________________________________________________ We initially approached relevant literature by examining the databases Web of Science, Google Scholar and Primo according to the underlying topics “corporate digital innovation”, “digital innovation units”, “organizational failure”, “ambidexterity”, as well as the German term “Digitale Innovationseinheit”. This procedure allowed us to assess the status of predominant literature and to determine that the topic has not yet been widely studied academically. Accordingly, we extensively gathered and reviewed grey literature such as consultancy studies, reports, dissertations, and online magazine articles to develop a substantial theoretical background and an in-depth understanding of our research topic and the related concepts. Subsequently, we identified distinct relevant theoretical areas and specified the focus of our literature review to find relevant peer-reviewed articles. Thus, we explicitly looked for literature explaining the dimensions of innovation, how these relate to how organizations can organize innovation and hence shedding light on organizational ambidexterity, finally resulting in the construct of DIUs in relation to organizational failure. Before conducting our empirical study, we ensured that we captured the most relevant literature so that no relevant concepts and theories were overlooked. Correspondingly, reviewing literature was especially challenging due to the non-uniform terminology of DIUs and the resulting difficulty in finding the right search items. However, eventually, by enhancing peer-reviewed articles with trustworthy and relevant grey literature, we developed a holistic theoretical construct that helped to understand our research topics, proved a significant research gap and that in turn built an adequate foundation for our empirical research.

2.1 Dimensions of Innovation

To clarify the role DIUs play in the organizational context and to illustrate their objectives regarding innovation, it is necessary to shed light on the distinct dimensions of innovation. The way innovation is organized, and the corresponding organizational and technological capabilities determine the degree of companies’ innovativeness and lay the foundation of this work (Christensen, 2013; Henderson & Clark, 1990).

2.1.1 Incremental Innovation

Incremental innovation can be described as a gradual improvement of an existing product or technology to constantly enhance the market position and strengthen companies’ capabilities. Industrial and more traditional companies primarily rely on incremental innovation as the main driver to sustain their competitiveness (Richter et al., 2018; Hendernson & Clark, 1990).

2.1.2 Sustaining Innovation

Similarly, sustaining innovation focuses on continuing to satisfy existing customers by incrementally improving the product or adding small features. Being especially relevant in established market, it targets customers with high expectations and the willingness to pay a higher price for an improved product. This results in a profitable market for the producer, while on the other hand freeing the lower market for new market entrants (Bower & Christensen, 1995).

2.1.3 Radical Innovation

Radical innovation in contrast emphasizes creating products that have fundamentally new meanings, often pushed by new technologies, and that target new markets. They force organizations to develop new capabilities, technologies, and novel problem-solving mechanisms and often pose great challenges for established firms as they prevalently lack the capabilities to go beyond their core business. Correspondingly, radical innovations are often developed and implemented by startups as they possess the required dynamic capabilities and are not yet blinded by a distinct industry or presumed customer needs (Richter et al., 2018; Henderson & Clark, 1990; Ahuja & Lampert, 2001; Tripsas & Gavetti, 2000).

2.1.4 Disruptive Innovation

Disruptive Innovations go beyond radical innovations and often utilize new technologies leveraged with business model innovation to develop fundamentally new products or services (Christensen, 2013). Primarily startups, disrupt industries by systematically addressing a market that is not on the radar of incumbents and subsequently adopting the incumbent’s market (Richter et al., 2018; Christensen, 2013). For DIUs, disruptive innovation is the focal goal, enabled by exploring opportunities through radically new digital technologies to support the PaCo in disrupting the market themselves, instead of being disrupted by new entrants (Infront, 2018).

2.2 Organizing Innovation

Within established organizations, the generic strategic goal to foster CDI requires amongst others an organizational culture that embraces innovation and creativity (Dobni, 2008; Martins & Terblanche, 2003). Such an innovation culture (IC) nurtures innovative ideation and positively influences the commercial success of innovations during the development process (de Brentani, 2001). An IC can according to DOBNI

(2008) be measured by an organization’s intention to innovate, the existence of an innovation infrastructure, its innovation influence, and its ability to implement innovations. Embracing a constructive tolerance for failures furthermore positively affects the firm performance (Mohan et al., 2017). A changing technological landscape and a stronger organizational innovation focus requires established companies to develop new capabilities (Tripsas & Gavetti, 2000). This notion holds true for sustained and disruptive innovation alike. For that purpose, it requires an organizational structure facilitating innovation (Lam, 2006). CHRISTENSEN AND OVERDORF (2000) introduce three

different ways how established firms can address this structural challenge. Incumbents can (1) develop capabilities internally, (2) build a spinout organization or (3) acquire companies that already have the respective capabilities. All three approaches require a new organizational framework dedicated to this goal.CHRISTENSEN AND OVERDORF

(2000) summarize that internally developing capabilities often require new interdisciplinary teams working in a project-oriented way. DIUs usually employ such teams and provide the structural framework and room for unfolding. Moreover, BORCHERT AND HAGENHOFF (2003) describe that the innovation function can be

or decentrally. In a decentralized structure, the BUs of an organization take over the task of innovating. Hybrid models include for instance a central innovation department and innovation officers in the various departments working together with the central function (Borchert & Hagenhoff, 2003).

2.3 Digital Innovation

In the context of corporate innovation, DI can be defined as “the creation of (and consequent change in) market offerings, business processes, or models that result from the use of digital technology” (Nambisan et al., 2017, p. 224). Bringing new products to the market is an important task fostering organizational competitiveness, while DI is also developed through more internal process innovations (Lyytinen et al., 2016; Kohli & Melville, 2019). DI gains increasing academic and practical attention as it positively contributes to the overall firm performance, also of more traditional and industrial companies (Hanelt et al., 2021). The underlying mechanism of DI is the symbiosis of physical and digital components with the potential to radically change the meaning of products and services as well as the nature of innovation through creating truly transformational novelties (Yoo et al., 2012). DIs have a generative, convergent, and

borderless character enabling infinite recombination and reshaping of digital artifacts,

without limitations to a physical form or digital dimensions. This bears opportunities but also poses challenges for companies regarding new ways of working, organizing, and managing DI (Yoo et al., 2012). For incumbents, especially market-oriented DIs require capabilities that usually exceed their existing capabilities focused on continuous improvement (Svahn et al., 2017). Thus, organizations need to find ways to successfully manage the entire DI process (Nylén & Holmström, 2015). Despite a lasting reliance on centralized internal sources for DI, there is an increasing shift towards open innovation and thus more dynamic organizational boundaries that allow the integration of external sources of DI (Lakhani et al., 2012; Capgemini., 2020). Companies increasingly collaborate with different innovation sources such as universities, startups, crowds, and public innovation hubs to complement internal innovation efforts. This emerging distribution of innovation agency is highlighted by central innovation sources still being the primary contributors to the overall innovativeness of enterprises, while the newer sources, such as innovation labs and crowds, combined generate the more significant innovations (Capgemini, 2020). The opportunities of more dynamic organizational

boundaries and a variety of DI sources also question existing suppositions of DI agency (Nambisan et al., 2017). The rise of co-creation and open collaborative innovation requires companies to increasingly orchestrate DI, rather than innovate themselves and leverage multiple sources of innovation (Ind et al., 2013; Baldwin & von Hippel, 2011).

2.4 Organizational Ambidexterity

Companies try to be successful within their core business while simultaneously preparing themselves for pursuing new business opportunities (Tushman & O'Reilly III, 1996). One major challenge is balancing between these two activities, hence exploration and

exploitation. DUNCAN (1976) introduces us to the co-existence of both strategic goals as

organizational ambidexterity. Organizations should pursue the exploitation by

incrementally innovating, and by aiming for process efficiency and operational excellence. On the other hand, organizations should focus on the identification of future opportunities arising from a changing business and technology landscape, hence creating radical and disruptive innovations (Duncan, 1976). MARCH (1991) furthermore highlights

the conflict between the two organizational goals of exploration and exploitation and reveals different approaches to target both. While exploitation presupposes involved actors to have detailed knowledge of the existing core business, explorative activities require more creativity and openness to potential future business (March, 1991). To achieve both, organizations need to become ambidextrous. Ambidextrous organizations are differentiated by being able to simultaneously achieve both, which is today further enabled by increasing DI (Birkinshaw & Gibson, 2004; O'Reilly III & Tushman, 2004). Regarding organizational ambidexterity, the two major research streams contextual

ambidexterity and structural ambidexterity prevail. Supporters of each of the

approaches controversially discuss the optimal implementation of organizational ambidexterity (O'Reilly III & Tushman, 2013).

2.4.1.1 Contextual Ambidexterity

BIRKINSHAW AND GIBSON (2004) argue for contextual ambidexterity which unifies the

simultaneous exploration and exploitation within individual employees, who split their work time accordingly. A reported advantage is the situational focus ensuring that employees who dedicate a certain amount of their time doing both, are aware of the requirements of the core business and can include this knowledge within the exploration activities (Birkinshaw & Gibson, 2004). Acknowledging the advantages, KAUPPILA

(2010) on the other hand, highlights the lack of explicitness of this approach regarding how exploratory activities are realized.

2.4.1.2 Structural Ambidexterity

O’REILLY III AND TUSHMAN (2004) emphasize the benefits of structural ambidexterity as

the stricter separation of exploration and exploitation. Embedded in the organizational structure, this approach ensures superior performance for the exploration and exploitation alike and hence a better overall company performance. They further argue that distinct organizational entities with the sole focus on the exploration of new opportunities should be linked to the core organization via top management. This approach enables more concentrated innovation activities, yielding better explorative results for the PaCo (O'Reilly III & Tushman, 2013). While structural ambidexterity ensures the explicitness of responsibilities, missing in contextual ambidexterity, it also requires leadership to integrate external knowledge and to circumvent isolation of the respective activities (Kauppila, 2010). However, the integrative character of structurally separated activities is focal to the approach, as the missing alignment can critically impede the overall success of ambidextrous organizations (Donada et al., 2021).

2.5 Digital Innovation Units

This part explains the research subject DIUs and delivers a definition of the term as well as a description of configuration options for DIUs and their peculiarities. Accordingly, this part summarizes the state of our knowledge regarding DIUs.

2.5.1 Embodiment of Structural Ambidexterity

Preceding the term DIU, the concept of separated units within ambidextrous organizations was already well researched, whether as central research and development (R&D) units for incremental innovation or e.g., corporate incubators (Chiesa, 2000; Becker & Gassmann, 2006). As a focal instrument for structural ambidexterity, they also prevail as advantageous for radical innovations within literature (O'Reilly III & Tushman, 2013; Arnold, 1992). BORCHERT AND HAGENHOFF (2003) compare different organizational

forms for innovation and highlight that centralized innovation units channel the organization’s innovation capabilities. LAM (2006) points out that project-focused

adhocracies employing interdisciplinary teams tend to yield more radical innovations. Adhocratic project teams are like heavyweight teams for innovation tasks. However, the major difference between both, is that an adhocratic organization systematically employs

interdisciplinary project teams for a variety of tasks, increasing the entire organization’s innovativeness, while heavyweight teams are often deployed for one specific task and are dissolved after fulfilling this assignment (Lam, 2006; Christensen & Overdorf, 2000; Clark & Wheelwright, 1992). With the example of a newspaper, GILBERT (2006)provides

empirical evidence for the advantages of structurally differentiating between a unit for exploring the opportunities of digital journalism and exploitative tasks of the threatened traditional newspaper business. Accordingly, an overlap between those tasks is counterproductive for both endeavors (Gilbert, 2006). TUMBAS ET AL.(2018)agree on the

potential of increasing the innovation pace by decoupling new DI initiatives from the core business functions and emphasize the decreased risk of conflict between innovation and core business. Beyond exploring DI, these initiatives help developing dynamic digital capabilities that need to be reintegrated into the PaCo to holistically prepare an organization for future challenges, limiting the time of the decoupled approach (Tumbas et al., 2018; Svahn et al.,2017;Wildhirt & Bub, 2018). HOLOTIUK AND BEIMBORN (2018)

describe DI labs as the embodiment of organizational ambidexterity and explain that they represent the “organizational structure for the development of digital innovation” to foster explorative tasks within incumbent firms (Holotiuk & Beimborn, 2018, p. 3). As representatives for structural ambidexterity, they provide a locus for innovation within established firms, while providing the freedom to develop DIs (Poguntke, 2019; Turrin, 2019; Lau, 2019).

2.5.2 DIUs – Origin and Examples

The term DIU as a collective term for various innovation vehicles was coined by the German consultancy Infront Consulting & Management GmbH (hereinafter referred to as Infront) in their first study about corporate innovation units in 2017 (Infront, 2017). The first mentions of the term, however, can be found in the context of describing innovation units in the field of journalism (Heurist, 2009; BBC, 2014). This term origin aligns with the comparably early research regarding a separate innovation unit focused on digitalization from GILBERT (2006),who investigated a newspapers organization. One

of the first examples of a form of a DIU was the Palo Alto Research Centre (PARC) of Xerox, established in the 1970s. Strategically located in today’s famous Silicon Valley, near the Stanford University, the absence of traditional guiding structures caused by the distance from Xerox’s corporate headquarters in New York was supposed to unleash the full potential of creativity. PARC developed potentially revolutionary ideas for the

computer industry. However, it eventually failed to diffuse and commercialize these ideas (Poguntke, 2019). Another prominent example highlightin the potential of DIUs is X, the moonshot factory of Google’s PaCo, Alphabet, that has already developed innovations such as the first delivery drone with a commercial flying license in the USA (Levin, 2019; etventure, 2020b). Within the DACH-region two successful examples are the Audi Denkwerkstatt (German for thinking garage) and the SAP Innovation Center Network that both were awarded the prize for best DIUs in 2019 by the Capital magazine and Infront (Infront, 2019; Kreimeier, 2019).

2.5.3 DIUs – Terminology

Academically, the term DIU is not yet widely established and has not been used in peer-reviewed journal articles so far. Overall, research regarding DIUs is still scarce but currently gains popularity, especially in Germany. While FUCHS ET AL.(2019)research

digital units, a subsequent article concerned DIUs, moving from digitally capable separate units to entities focusing on facilitating DI (Barthel et al., 2020). They describe DIUs “as organizational units with the overall goal to foster organizational digital transformation by performing Digital Innovation activities for existing and novel business areas” (Barthel et al., 2020, p. 4). RAABE ET AL.(2020a), on the other hand, understand DIUs as

DI facilitators with a primarily internal focus. However, this internal focus conflicts with their empirical observation of incubators with an external orientation. Furthermore, their study regarding the focal activities of DIUs, concludes that 28 of the observed 38 cases of DIUs, either diffuse DIs also externally, explore new digital technology trends, or participate in existing DI ecosystems. Thus, such activities can be understood as being at least not purely internal (Raabe et al., 2020a). This mixed focus on internal and external development of DIs aligns with INFRONT (2019)andBARTHEL ET AL.(2020), highlighting

different design types that can have the goal to nurture new market-oriented businesses. Other grey literature describes the term DIU more generally as organizational units facilitating DI for the PaCo (etventure, 2020a). Depending on the source, the term comprises digital labs, corporate accelerators, corporate incubators as well as corporate company builders (Infront, 2017; etventure, 2020a). In popular business magazines, the term digital lab prevails to describe DIUs. When looking at the examples of the units discussed, these however often also include units, being rather characterized as incubators or accelerators (Kyriasoglou, 2020; Infront, 2017; Kreimeier, 2019). The focal commonality for all types of DIUs lies within the separation of the DI task from the core

business. This shows that, especially within business magazines, there is currently still a lack of nuance regarding a definition and a consensual typology, and many related terms are used synonymously. This observation expands to academic literature.HOLOTIUK AND

BEIMBORN (2018)especially highlight the centralization of the innovation task to foster

DI within DIUs, but however, call the units digital labs as well. This overlap of objectives and varying terminologies supports the need for further research about DIUs.

2.5.4 DIUs – Characteristics and Design Dimensions

The recency of the research subject DIUs also results in knowledge gaps regarding their characteristics. FUCHS ET AL.(2019) approach a foundational understanding of DIUs by

developing a taxonomy for digital units. Understanding them as facilitating the PaCo’s digital transformation efforts with innovation activities, their taxonomy provides a first conceptual framework to describe DIUs (Fuchs et al., 2019). Building on this foundation, BARTHEL ET AL.(2020)adapt the taxonomy for DIUs by classifying characteristics and

design options along eight dimensions (s. Figure 1). These dimensions are (1) the innovation orientation, (2) market focus of innovation, (3) scope of the innovation process, (4) how the DIU is embedded into the organization, (5) the degree to which it externally cooperates and (6) the degree to which it internally collaborates. Furthermore, the authors describe (7) how projects are selected and (8) the location of the projects as the last dimensions. The adapted taxonomy is comparable to the design dimensions of corporate incubators that LAU (2019) describes along twenty different constitutive

criteria. These criteria are grouped into the four major dimensions (1) strategic orientation, (2) cooperation with the PaCo, (3) cooperation with external innovation partners, and (4) location and facilities. However, it needs to be clarified that there is a difference between DIUs as a collective term compared to corporate incubators, being conceptualized as one specific type of DIUs (Infront, 2018).

RAABE ET AL.(2021)specifically investigate the area of activity of DIUs, while in RAABE ET AL.(2020b)the authors mainly develop two types of DIUs along the (1) focus of the

units, (2) how they operate, and (3) based on the different innovation stages that the DIUs are engaged in. These stages of the innovation process comprise the discovery, also known as ideation phase, the development or incubation, which includes prototyping, and finally scaling, also known as commercialization (Infront, 2019; Barthel et al., 2020).

Figure 1 Taxonomy of DIUs

Note. Adapted from “Embedding digital innovations in organizations: A typology for digital innovation units”, by Barthel et al., 2020, p. 8. Copyright 2020 by GITO Verlag.

FUCHS ET AL.(2019)reframe the practical interpretation of autonomy as a crucial DIU

characteristic, by pointing out its limitation through organizational guiding principles. This contrasts with the understanding of autonomy as fundamental to the idea of digital units as embodiments of structural ambidexterity (O'Reilly III & Tushman, 2013; Christensen & Overdorf, 2000).

2.5.5 DIUs – Design Typologies

Building on configurational options, initial research has developed different typologies for DIUs, based on the dimensions and constitutive criteria. BARTHEL ET AL.(2020) define

three types of DIUs as ideal set-ups to reach the goal of the PaCos. They differentiate between the types “internal facilitator”, “external enhancer” and “external creator”. While the internal facilitator is distinguished by its close ties to the organization focusing on developing internal projects of the core organization, the two other types have a more external orientation. The external enhancer cooperates with external partners and develops ideas more autonomously within existing business areas. The external creator fosters business ideas and develops them autonomously. Common to all types is their coupling with the core organization to cooperate with them whenever necessary (Barthel et al., 2020). RAABE ET AL. (2020a) distinguish between the DIU types “center of

excellence” and “supportive coaching & screening” (C&S) units. C&S types focus on

Category Purely Internal Primarily Internal Primarily External Purely External Idea Generation Idea Selection Innovation Implemen-tation Innovation Commer-cialization External Cooperation Internal Ties Project Selection Location Standard Working Mode Sporadic Liaison Autonomously Offsite Line Organization Seperate Legal Entity None Core Organization Integrated Order Onsite Irregularly Regular Liaison Mixed Mixed Dimension How? Setting & Design Pr oj ec ts Characteristics Balanced

Existing Business Areas Novel Business Areas Innovation Development Executive Department What? Objective & Scope Co lla bo -ra tio n Embedding Scope of Innovation Process Market Focus of Innovation Innovation Orientation

ideation for exploring DIs. Center of excellence types are responsible for the entire innovation process until diffusing innovation within the respective business unit (BU) (Raabe et al., 2020a). Consultancies, however, distinguish rather between better known constructs such as e.g., incubators, labs, or accelerators (etventure, 2020a; Ohr, 2020). RAABE ET AL.(2021)conclude across all different types of DIUs mainly two objectives

for the units, being the selection of DI for solving a PaCo’s business problems and transforming the business prompted by DI. So far, there is no consensus on the differentiation of different DIU types. The external creator and corporate incubators are similar regarding the objective of nurturing business ideas to address new markets. Digital labs are comparable with C&S units regarding the focus on ideation (Infront, 2018; Raabe et al., 2020a; Holotiuk & Beimborn, 2018; Barthel et al., 2020). However, beyond these generic types,BECKER AND GASSMANN (2006)further differentiate between four major

types of corporate incubators, mainly distinguished by the objectives (i.e., profit vs. non-profit) and the affiliation with the core business of the PaCo. LAU (2019)builds on this

distinction and proposes six different types of corporate incubators. Here, e.g., the “idea lab” is comparable to C&S units. On the other hand, LAU (2019)also proposes corporate

accelerators as a type, which e.g., INFRONT (2017) differentiates from incubators instead

of understanding it as a sub-category. The examples show that the different DIU types proposed by researchers so far, do not entirely align with each other. These discrepancies and the overarching objectives of DIUs by RAABE ET AL.(2021)finally show that the

differentiation is mainly based on whether DIs target internal processes or external business and on the distinction along different innovation phases. The missing adoption of these constructs beyond the introduced articles highlights the recency of the research phenomenon DIUs and the need to understand more about how they evolve in practice.

2.5.6 Conclusion and Own Definition

Resulting from the described discrepancies, we could not find a single definition that is legitimized by either its dispersion or its exhaustiveness. Accordingly, we combine the explanation of INFRONT (2017), BARTHEL ET AL.(2020), and RAABE ET AL.(2020a), by

defining DIUs as

separated organizational entities of incumbent firms that focus on the facilitation of digital innovations on behalf of and for their parent company.

Within the context of this study, DIUs comprise amongst others corporate accelerators, corporate incubators, and digital labs that try to nurture DIs also utilizing market-oriented business model innovations. These DIs can be either sustained, disruptive, radical, or incremental innovations. In answering the question of why DIUs close, we focus on the organizational instrument and deliberately do not differentiate between the different types or configurations of DIUs. Such a differentiation decreases the possibility of finding patterns for DIU closure being applicable beyond the context of distinct DIUs. Through this inductive approach, the study contributes to the understanding of DIUs and how established companies utilize this instrument.

2.6 Organizational Failure

Failure and closure of DIUs are often equated with each other. Furthermore, DIUs are separate organizational entities with their own strategic goals. Accordingly, understanding why organizations generally fail can contribute to answering our RQ.

2.6.1 Failure of Established Companies

While the term organizational failure (OrFa) is not clearly differentiated, it can be characterized mainly by its negative consequences for the organization, such as its closure (Mellahi & Wilkinson, 2004). Its causes are even more multifaceted. Within research, two major streams prevail. On the one hand, the deterministic perspective highlights exogenous factors. On the other hand, the voluntarist approach emphasizes how managerial decisions or actions by the company cause its failure (Mellahi & Wilkinson, 2004; Khelil, 2016). TUSHMAN AND ANDERSON (1986) emphasize that innovative changes

can cause the failure of existing businesses by destroying the relevance of previously important capabilities of established firms (Tushman & Anderson, 1986). This explanation for OrFa is based on the notion that it is often caused by external discontinuities (Schumpeter, 1942). This daring thesis of SCHUMPETER (1942)

demonstrates that periodical failure results amongst others from major technological changes or economic crises making existing businesses redundant. Despite such external factors that organizations seem to be powerless against, other scholars see managerial (in)actions as causing OrFa. Hence, acknowledging external circumstances, one prevailing cause for the failure of established organizations is the inertia of reacting to disruptive changes (Tripsas, 2009). HAMBRICK AND MASON (1984) constitute that

resulting in their success or failure. This deterministic view assumes a more decisive role of the company’s actors regarding OrFa. HERACLEOUS AND WERRES (2016) especially

outline that OrFa is caused by strategically wrong decisions over time and is hence a result of a process rather than of single events. HABERSANG ET AL. (2018) paint a more

differentiated picture by highlighting that OrFa often occurs in conjunction with a multitude of reasons, enabled by strategically wrong decisions and external factors contributing to this failure process. Other studies have attempted to derive more specific causes for OrFa and have found a lack of cashflow, low employee satisfaction, poor managerial communication, and inertia to changing circumstances as underlying drivers of OrFa (Lizano et al., 2014; Longenecker et al., 1999; Bhandari & Iyer, 2013). While these findings sound obvious, it could be argued, that these factors partly represent rather symptoms than causes. Hence, the nuanced perspectives of HABERSANG ET AL. (2018)

and MELLAHI AND WILKINSON (2004) contribute more to understanding the dynamic

interrelationships of external and internal causes leading to OrFa.

2.6.2 Failure of Startups

Startups fail for different reasons than established organizations. Startups are young companies “designed to create a new product or service under conditions of extreme uncertainty” (Ries, 2011, p. 27). They are similar to DIUs in their working modes and their goals of quickly delivering results regarding innovation-oriented business (Christensen & Overdorf, 2000; Wildhirt & Bub, 2018; Infront, 2017). Accordingly, understanding their failure can help for understanding the closure of DIUs. A widely distributed statistic states that 90% of startups fail (Patel, 2015; Krommenhoek, 2018; Carrigan, 2020). Due to their uncertain landscape of activity, startups are associated with high risks, as their projected business development is often based on assumptions regarding desirability or technological feasibility that have yet to be validated (Ries, 2011). RIES (2011)argues that often the lack of management strategies or too unstructured

processes are the underlying reasons for startup failure. Recent academic and practical studies report that the main reasons for failure of startups are (1) that the products and the respective business model do not fit market needs, (2) that startups run out of funding, and (3) the wrong team composition (CB Insights, 2019; Cantamessa et al., 2018). FEINLEIB (2012)furtherpoints out that often the entrepreneurs fail to execute a long-term

strategy and lack financial resources to continue their business beyond the initial stages. Especially not meeting market requirements and the inability to execute a suitable

strategy are also observed by GIARDINO ET AL. (2014), who researched early-stage

software startups. One major difference of DIUs and startups is the resource of funding. Startups rely on external investors and the founders’ resources. DIUs are funded by their PaCo that initiated the setup of the DIU (Infront, 2017). Accordingly, a missing product-market fit and an insufficient strategy execution could also result in DIU failure. However, due to the secured funding from the PaCo, we assume that a lack of financial resources might not be an applicable reason for why DIUs fail.

2.6.3 Failure of DIUs

The failure or closure of DIUs are comparably new topics underrepresented in research. One prominent case was the early closure of the idea lab of German publisher Axel Springer resulting from its strategic realignment (Wirminghaus, 2014). Recently the closure of Daimler’s Lab1886 gained public attention as a majority was sold to a new PaCo, where the unit was strategically repositioned to realize its potential within new organizational boundaries, while business press declared its failure (Daimler, 2020; Kyriasoglou, 2020). While innovation experts report a rate of 80% to 90% of failing DIUs, they lack empirical evidence substantiating this assumption (Capgemini, 2015b). Hence, the reported failure of DIUs poses the question of whether these instruments for facilitating DIs on behalf of their PaCo are not as effective after all (Unger, 2018; etventure, 2020b). Beyond this question, DIUs often lack specific goals, making their success and failure difficult to measure (Infront, 2018; etventure, 2020a; Raabe et al., 2021). The only existing empirics for closed DIUs are provided by RAABE ET AL. (2020b),

investigating challenges of DIUs. They point out that the two closed DIUs to their knowledge “were closed because their leads left the unit” (Raabe et al., 2020b, p. 5). They further elaborate that this caused the top management of the PaCo to lose trust and close the DIU (Raabe et al., 2020b). Consequently, a common understanding of business press, scholars, and practitioners is that the closure of DIUs coincides with their failure. However, this prevalent assumption lacks the clear distinction between failure and closure (Raabe et al., 2020b; Unger, 2018). This shortcoming is a crucial knowledge gap because if DIUs closed only because they failed, the reported failure of DIUs would question their impact. OrFa literature however distinguishes more nuanced between closure and failure (Mellahi & Wilkinson, 2004). TUMBAS ET AL. (2018) point out that DIU closure can also

result from their design-inherent impermanence or internal competition abrogating them. This more nuanced view in combination with the distinction from OrFa literature

challenges whether DIUs are only closed when failing to achieve their goals, posing the question what role failure plays in this regard. Despite these knowledge gaps, the challenges of DIUs are already well described (Raabe et al., 2020b; etventure, 2020a). Frequent challenges for DIUs are dissonances with the PaCo regarding support and constructive collaboration, financial bottlenecks imposed by corporate cuts plans and the assignment of part-time employees within DIUs (Infront, 2019; etventure, 2020a; Raabe et al., 2020b). Finally, RAABE ET AL.(2020b)name unclear DIU objectives as a major

obstacle and conclude that the lack of backing by executive management is the most common challenge for DIUs. This multitude of challenges with a focus on the relationship with the PaCo and the missing clarity of the DIU’s goals align with the obstacles outlined by INFRONT (2020)as well as ETVENTURE (2020a). PARC e.g., is reported to have failed

due to the missing ability to diffuse its ideas within Xerox’s BUs (Poguntke, 2019). This observation relates to the “Not-Invented-Here-Syndrome” (NIHS), a challenge between DIU and PaCo, describing the aversion of BUs to innovations originated elsewhere (Chiesa, 2000; Raabe et al., 2020b). While scholars agree on the challenges DIUs face, whether these challenges lead to DIU failure or closure, or whether there are entirely different reasons for these events, remains to be empirically verified.

2.7 Theoretical Framework

DIUs are a comparably new research field incorporating structural ambidexterity, with a focus on facilitating DIs on behalf of the PaCo. Accordingly, to understand the paradox of why DIUs are being closed, while other companies continue implementing new ones, it is essential to gather new knowledge going beyond individual cases of closed DIUs. This requirement becomes especially apparent by the fact that on the one hand, a multitude of different DIU typologies exists, while on other hand, the term itself is not yet comprehensively defined and the distinct objectives of DIUs vary across the examples (Raabe et al., 2021; Lau, 2019; Infront, 2019). Beyond these limitations within the literature, it can be summarized that new knowledge needs to be gathered for the closure of DIUs, as it has been assumed so far that the closure of DIUs results from their failure and that the challenges faced by DIUs are the causes of this failure. With their similar characteristics, startups are challenged by the viability of their business model and their funding, while DIUs are also exposed to challenging relationships with existing structures and the culture of the PaCo that tend to be contributing factors to the failure of established

organizations (Tripsas, 2009; CB Insights, 2019). Aligning with OrFa literature, it is crucial to understand whether external reasons determine a DIU closure (i.e., deterministic view), or whether managerial decisions and the performance of the DIU cause its closure (i.e., voluntarist view) (Mellahi & Wilkinson, 2004; Habersang et al., 2018). Only utilizing these concepts can ensure to comprehensively understand why DIUs are being closed.

3 Research Method

______________________________________________________________________ The purpose of this chapter is to present the methodological background of our study considering our underlying philosophical orientation. It illustrates our data collection and analysis procedure and how theory was derived from the data. In addition, it outlines how research quality was assured, also considering the ethical implications. The overall aim of the study was to explore why DIUs are being closed and hence the methodology, methods, and techniques were chosen to get a diversified perspective on the topic so that theories could emerge from data.

______________________________________________________________________

3.1 Research Philosophy

The research philosophy, hence, our individual perception of reality (ontology), how we think knowledge is created (epistemology), and how our own values and ethics affect the research process (axiology), built the foundation of the research design. It determined the way we collected and interpreted data and hence answered our RQ. Based on our general philosophical stance we detected appropriate research methods and techniques, being

grounded theory (GT), semi-structured interviews, and grounded analysis (GA)

(Easterby-Smith et al., 2018; Charmaz, 2006).

Regarding the ontology, we adopted a relativism approach, implying that we perceive the truth as an adaptive construct that depends on various circumstances and that reality cannot solely be based on objective data, but personal perspectives and interpretations must be considered (Collins, 1983). To understand the overall phenomenon of DIU closure, we built our theory upon varying opinions and experiences of former heads and employees of closed DIUs as well as partly on our own presuppositions and the standpoints of distinct experts. This approach is based on our relativist ontology and ensures the consideration of multiple viewpoints regarding DIUs and the acknowledgment of each perspective as truth. To provide answers to our RQ, without biasing our observations we do not value one viewpoint over another. If we e.g., were to investigate the success of DIUs measured by absolute metrics, such as return on investment (ROI), our ontology would be realistic and as internal realists, we could have researched how a DIU failure becomes visible by a multitude of factors. On the other hand, a nominalist ontology would have been derived from a RQ looking at e.g., the

internal dynamics of a DIU (Easterby-Smith et al., 2018). However, to grasp why DIUs are being closed, multiple perspectives and their individual perception of reality and truth were crucial. This is mainly because different roles in the context of DIUs differently interpret the DIU closure and the reasons for this event because they have different background information and differently perceive e.g., the work within a DIU and its significance for the PaCo. Hence, our relativist ontology was fundamental to create knowledge by incorporating multiple realities and perspectives as truths from the respective viewpoint.

Consequently, we assume that knowledge is constructed through various agents socially interacting and making sense of their experiences, aligning with a social constructionist epistemology (Berger & Luckmann, 1966). Thus, our aim was not to objectively gather facts and derive theory from the pure data, but also interpreting statements and expressions, and contextualizing the statements of each interviewee (INT) with the respective case. We furthermore acknowledged each perspective and how the INTs made sense of their experience as reality and holistically observed the closure of DIUs. From these approaches, we eventually developed concepts and theories of why DIUs close. By combining different sampling strategies to collect data, we were able to capture diverse viewpoints on our research phenomenon, increasing our research findings’ accuracy and credibility (Easterby-Smith et al., 2018; Patton, 2014).

Regarding our axiology, we believe that an entirely objective observation, detached from our own value system is impossible, but that it is important to be aware of this bias. This

value-bound orientation aligns with our research method and technique and the

respective high degree of interaction as well as the conscious inclusion of our own assumptions and experiences that aimed to guide a conversation rather than a predefined questioning. Thus, we play an integrative role within the research process and our interpretations are essential for theory-building, leading to our general philosophical position of interpretivists (Saunders et al., 2019).

3.2 Research Approach

In general, it can be distinguished between qualitative, quantitative, and mixed methods approaches to research a particular topic and hence decide on the respective methods for collecting, analyzing, and interpreting data under consideration of the underlying research

philosophy (Creswell, 2014). Based on our philosophical stance and our RQ, we follow a qualitative research approach. This is according to CHARMAZ (2006) advantageous

over quantitative research regarding its continuous and flexible nature allowing new research paths to be discovered during the data collection process. Correspondingly, we aimed to inductively explore emerging patterns from the data we collected, and in turn, develop a theory that is grounded in this data (Easterby-Smith et al., 2018). Our inductive and exploratory research approach results from the purpose of developing a deep understanding of the discrepancy between the increase in newly opened DIUs and DIUs that are being closed and from the absence of significant academic theories about why DIUs close. An inductive approach is most suitable because the contrasting deduction would require previous assumptions that can be approved or invalidated. Finally, abductive research would have been suitable, if fragmented and testable data about DIU closure existed, that we could have substantiated with new data to develop a theory. However, the novelty of enquiring into the closure of DIUs overall makes an inductive approach in combination with exploring the topic with a qualitative research most suitable (Saunders et al., 2019).

In line with our philosophical stance, we believe that our interrelation with the research participants, the reflection of our own interpretations and presuppositions are essential for theory-building. Furthermore, we are convinced that preliminarily grasping existing literature fosters the knowledge-creation and a deeper understanding of the research problem and purpose. Accordingly, we adopted a constructionist GT approach to our study, following CHARMAZ (2006). Prior to data collection, we gathered our own

assumptions and experiences in a preliminary workshop with three DIU experts and developed our interview topic guide (ITG) based on this procedure. The purposeful consideration of our presuppositions was essential for conducting a high quality and trustworthy research as disregarding them would otherwise have led to a distortion of the outcomes (Charmaz, 2006). Building on our research approach, we incorporated the knowledge and experiences from people with different viewpoints on the topic. This enabled us to conduct a truly in-depth and relevant research. To stay open to existing concepts and approaches, we acknowledged predominant literature to frame our understanding of DIUs (Charmaz, 2006). Moreover, we conducted few but intensive and detailed semi-structured interviews that stimulated a conversation through open-ended

questions to provide an accurate and rich data set. Our interactive approach is characteristic for constructionist GT as being flexible but simultaneously focused to gain insights (Alemu et al., 2015). Accordingly, we iteratively slightly rephrased our ITG after the first two interviews and constantly compared early- and later-stage data throughout the data collection and analysis process. Amongst others, we used memo-writing as a method to engage with our gathered data to ensure a constant comparison. This aligns with the GT principles of positivists GLASER AND STRAUSS (1967), who, however,

otherwise propose that the researcher is an objective observer and that ideas should solely evolve from the data and that a literature review should be conducted after analyzing data to avoid bias (Glaser & Strauss, 1967; Alemu et al., 2017; Glaser, 1978; Charmaz, 2006). To analyze our data, we applied the constructionist GA approach. Hence, the analysis already started during collecting data. As we believe knowledge is co-created by the interaction of the researchers and the INTs, the creation of data and its analysis were highly interrelated (Charmaz, 2006). Thereupon, in constant examination of the data and their systematic analysis, we have tried to discover unambiguous categories and, in turn, to derive theories from them. Contrasting other qualitative analysis methods, e.g., content analysis, GA does not test existing ideas and try to classify data according to prevailing structures but holistically explores new structures (Easterby-Smith et al., 2018).

3.3 Research Design

To answer our RQ holistically, we based our research on multiple viewpoints and experiences through semi-structured in-depth interviews with former agents of closed DIUs and with DIU experts.

Since it was not clear in advance who exactly our research participants would be, several preparations were necessary. Accordingly, we developed a three-stage study design to first identify and recruit suitable participants and then obtain meaningful responses to our RQ. The first two steps contributed to the development of our sampling strategy and enabled exploring underlying themes which in turn built the foundation for our ITG. Subsequently, to get a multi-perspective on our research, in step three of our study design, we interviewed people with different experiences and professional backgrounds. The purpose of this multi-layered structure was to leverage the continuous character of qualitative data collection along the entire process, hence also prior and post the actual