Business model innovation and organizational ambidexterity : theoretical foundation or superficial link?

(1)

1

Business Model Innovation and

Organizational Ambidexterity – Theoretical

Foundation or Superficial Link?

University of Amsterdam

Faculty of Economics and Business Master Thesis - Strategy (6314M0250Y)

Student: Manuel Metz (11087242) Submitted to: Dr. Stephan von Delft

(2)

2

Statement of Originality

This document is written by Manuel Metz who declares to take full responsibility for its contents:

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

Signature:

_______________

Manuel Metz, 18.08.2016

(3)

3

2.2.1 In search of a theoretical foundation ... 18 2.3 ORGANIZATIONAL AMBIDEXTERITY ... 19 2.3.1 Three forms of Structural Ambidexterity ... 21 2.3.2 Assessing the usability of the three forms ... 24 2.4 TRANSLATING THE THREE STRUCTURES ... 25 2.4.1 Structural Separation ... 26 2.4.2 Parallel Structures ... 27 2.4.3 Temporal Separation ... 27

2.5 COMBINING ORGANIZATIONAL AMBIDEXTERITY AND BUSINESS MODEL INNOVATION ... 28

3 ANALYTICAL FRAMEWORK ... 31

4 RESEARCH DESIGN ... 38

4.1 RESEARCH CRITERIA ... 39

4.2 LINKING THEORY TO RESEARCH DESIGN ... 40

4.3 QUALITATIVE DATA ANALYSIS STRATEGY ... 41

5 CASE STUDIES ... 42

5.1 CASE STUDY DEVELOPMENT AND RESEARCH PROCESS ... 42

5.2 SAMPLE SELECTION ... 42

(4)

4

5.4 DATA COLLECTION – COMPLEMENTARY DATA GATHERING ... 45

5.5 DATA COLLECTION – CASE STUDY DATABASES ... 45

5.6 LIST OF CASE STUDIES ... 45

6 DATA ANALYSIS ... 47

6.1 COMPANY ALPHA – OUTSOURCING AND LOGISTICS ... 47

6.1.1 Financial Resources ... 47

6.1.2 Human Resources ... 48

6.1.3 Past Organizational Learning ... 49

6.1.4 New Organizational Learning ... 50

6.2 COMPANY GAMMA – NEWSPAPER ... 51

6.3 COMPANY DELTA - AIRPORTS ... 55

6.4 COMPANY EPSILON – MUSEUM – VIDEO GAME CASE ... 59

6.5 COMPANY EPSILON – MUSEUM – ONLINE COURSE CASE ... 63

(5)

5

6.6 COMPANY ZETA – BUSINESS MODEL SEARCH ... 66

6.6.1 Financial Resources ... 67 6.6.2 Human Resources ... 68 6.6.3 Past Organizational Learning ... 69 6.6.4 New Organizational Learning ... 69 6.7 CROSS-CASE ANALYSIS ... 70 6.7.1 Financial Resources ... 72 6.7.2 Human Resources ... 72 6.7.3 Past Organizational Learning ... 72 6.7.4 New Organizational Learning ... 73 7 DISCUSSION ... 74 8 RESEARCH IMPLICATIONS ... 76 9 MANAGERIAL IMPLICATIONS ... 78 10 LIMITATIONS ... 79 11 CONCLUSION ... 80 12 LIST OF REFERENCES ... 81 APPENDIX 1: INTERVIEW QUESTIONS / GUIDE ... 93 APPENDIX 2: LINKING QUESTIONS TO ANALYTICAL FRAMEWORK ... 95 APPENDIX 3: CASE STUDY DEVELOPMENT AND RESEARCH PROCESS ... 97 APPENDIX 4: INTERVIEW STATISTICS ... 98 APPENDIX 5: CASE STUDY OVERVIEW ... 99

(6)

(7)

7

List of Figures

Figure 1: Structural Separation according to Raisch (2008) ... 22

Figure 2: Parallel Structures according to Raisch (2008) ... 23

Figure 3: Temporal Separation based on Raisch (2008) ... 24

Figure 4: Analytical Framework - schematic ... 32

Figure 5: Types of case studies according to Yin (2013) ... 40

(8)

8

List of Tables

Table 1: Selected business model definitions, based on Zott et al. (2011) (emphasis added) .. 11

Table 2: Analytical Framework ... 33

Table 3: Case Study Development Process Overview ... 42

Table 4: List of Case Studies ... 46

Table 5: Cross-case analysis comparison ... 71

Table 6: Interview Questions ... 94

Table 7: Interview Question Mapping ... 96

Table 8: Case Study Development and Research Process Details ... 97

(9)

9

1 Introduction

A lot has been written about business models in the last decades, since the topic arrived in the focus of academic research interest. A concept that followed naturally was that of business model innovation, taking a more dynamic perspective of the concept, while at the same time raising the question of how business models can be changed and innovated, when the competition catches up with the frontrunners.

However, business model innovation is lacking a theoretical foundation to support it. This lack has been identified by multiple researchers (e.g. Markides, 2013; Teece, 2010). However, there are multiple suggestions to which field might provide the theoretical foundation needed for business model innovation. One of these fields is organizational ambidexterity.

Despite these theoretical links and propositions, very little research has been done in testing or validating them (Markides, 2013). This thesis aims to bridge the gap between the concepts of business model innovation and organizational ambidexterity. In particular, it examines to what extent structural ambidexterity, i.e. strategies to organizationally structure and setup explorative activities, is used by firms pursuing business model innovation. By using a qualitative research approach, with semi-structured interviews and complementary documents for the collection of data, this thesis will explore the following research question:

How, if at all, are organizations using the forms of structural ambidexterity to organize their business model innovation activities?

(10)

10

2 Literature Review

2.1 The Origin of Business Models

While the history of the concept of business models can be traced back to Lang (1947), the academic research on it has only gained traction in recent decades. This is predominantly attributed to the rise of the Internet in the 1990s (DaSilva & Trkman, 2014; Magretta, 2002; Muzellec, Ronteau, & Lambkin, 2015; Teece, 2010; Zott, Amit, & Massa, 2011). While the e-Commerce business models of the 1990s were mainly fueled by hope that the companies would come up with highly profitable business models, the general idea of firms operating one or multiple business models (Teece, 2010) struck with researchers and practitioners alike. The growing number of articles in both theoretical and practitioner-oriented publications can be seen as a representation of the growing interest by researchers and practitioners alike (Zott et al., 2011).

Despite the short academic history of the concept, two main ideas have emerged, as characterizing business models: the creation of value and the capturing of value within organizations (Amit & Zott, 2001; Chesbrough, 2007; Teece, 2010). In simple terms, the two ideas describe the need for organizations to create value (while interacting with customers or other organizations), deliver it, and use capture mechanisms to turn value into profits (Teece, 2010). The two main ideas have been incorporated into various definitions developed by researchers. Yet, in spite of agreement on general principles, there is no unifying and universally accepted definition. In their overview article, Zott et al. (2011) conclude that “despite the overall surge in the literature on business models, scholars do not agree on what a business model is” (p. 1020). Additionally, the research on business models and their nature seems to develop in various silos (Zott et al., 2011), further showing the fragmentation of the business model research community.

(11)

11 Nonetheless, the ideas of value creation and value capturing are represented in many of the definitions discussed in the article, across the various the silos:

Table 1: Selected business model definitions, based on Zott et al. (2011) (emphasis added)

Author(s) (Year) Definition

Timmers (1998) The business model as “an architecture of the product, service, and information flows, including a description of the various business actors and their roles; a description of the potential benefits for the various

business actors; a description of the sources of revenues” (p. 2).

Amit & Zott (2001); Zott & Amit (2010)

Business models as a representation of “the content, structure, and governance of transactions designed so as to create value through the

exploitation of business opportunities” (Amit & Zott, 2001, p. 511).

Subsequently, this definition was evolved into stating that a business model is “a system of interdependent activities that transcends the focal firm and spans its boundaries” (Zott & Amit, 2010, p. 216).

Chesbrough & Rosenbloom (2002)

A business model as “the heuristic logic that connects technical potential with the realization of economic value” (p. 529).

Johnson, Christensen, & Kagermann (2008)

Business models are made from “four interlocking elements, that, taken together, create and deliver value” (p. 52).

Teece (2010) “A business model articulates the logic, the data and other evidence that

support a value proposition for the customer, and a viable structure of revenues and costs for the enterprise delivering that value” (p. 179).

All definitions in the table incorporate the two main ideas of value creation and value capturing. The definition developed by Amit & Zott (2001) for example, states that the business model represents the transactions, their structure and the governance used to control them, in order to

(12)

12 “create value through the exploitation of business opportunities” (p. 495). Another definition, developed by Chesbrough & Rosenbloom (2002) defines the business model as the logic used to connect technological innovation and the “realization of economic value” (p. 529). In Johnson et al.’s (2008)’s definition, the two ideas are represented in the four elements of their definition and how they describe the creation of value through transactions (the value proposition) and the capturing of this value (the profit formula, key resources, and key processes).

Next to these two main ideas, researchers have described another important aspect: The use of resources in creating and capturing value. A strong link to the resource-based view as a source of value creation is made in Amit & Zott’s (2001) article. The authors discuss that resources, if they fulfill the criteria proposed by Barney (1991), can be beneficial to the creation of value. This link is also made in DaSilva & Trkman (2014) and their article, in which they clarify the term from a theoretical perspective and, more importantly, distinguish the business model concept from other managerial concepts.

The link to resources and their usefulness in creating and capturing value is also given a prominent spot in the definition by Johnson et al. (2008). The authors see “key resources” as one of the four elements of any business model. They argue that the key resources of an organization are necessary to “create the value for the customer and the company” (p. 53). Furthermore, the importance of all kinds of key resources is highlighted, both the ones involved in creating value for the customer and the ones used for capturing this value, i.e. turning the revenue into profits.

As this definition of the business model is also the one that will be used throughout the remainder of this thesis, a brief overview will be given. Johnson et al. (2008) define business models by developing four elements: Customer Value Proposition, Profit Formula, Key

(13)

13

Customer Value Proposition refers to the ability of the company to propose a certain value,

which the customer recognizes as important. In the words of the authors, it is getting “an important job done” (p. 52), by solving a “fundamental problem in a given situation that needs a solution” (p. 52).

The next element, the Profit Formula, is a representation of how the company prices its solution internally and externally. Meaning that the revenue of the business model, as well as the costs incurred to create the value need to be monitored (Johnson et al., 2008, p. 53).

The third element, Key Resources, refers to the use of resources (e.g. people or technology) in providing the value to customers. Additionally, the interaction between the resources is of importance, as only some resources provide competitive differentiation and others are too generic to deliver this (Johnson et al., 2008).

The fourth and last element, Key Processes, is more about the execution, or the ability of the company to create the customer value in a repeatable and scalable manner. Additionally, a company’s rules, metrics, and norms are also grouped under this term (Johnson et al., 2008). The main idea is that business models are especially valid if the processes that lead to the creation and capturing of value can be repeated and put into certain other processes.

There are three reasons this definition was chosen for use in this thesis:

1) It covers both the value creation and the value capturing aspect of the business model, using its four elements

2) It emphasizes the use of resources as an important part of the business model

3) It establishes a clearly structured business model definition, more so than the more abstract definitions by Timmers (1998), Amit & Zott (2001), or Magretta (2002). This is supplemented by the provision of specific descriptions and examples for each of their four elements, allowing a better understanding of their definition

(14)

14

2.2 Business Model Innovation (and its importance for businesses)

Whereas the definitions and research on business models create the impression that an established and well-functioning business model does not require changes, and is thus static in nature, the business model innovation literature takes a more dynamic approach (Chesbrough, 2007, 2010; Doz & Kosonen, 2010; Lindgardt, Reeves, Stalk, & Deimler, 2009; Teece, 2010). Doz and Kosonen (2010) emphasize that “strategic discontinuities and disruptions” (p. 370) create the need for changes to an existing business model. Teece (2010) adds to this, when he describes the necessity for business model learning, i.e. the adjusting of an existing business model in light of a new, rival one.

The perspective of constant change complements this perspective of changes due to discontinuities and disruptions (e.g. competing business models or major leaps in technology). Researchers argue that a business model requires constant refinement and improvement, based on experiences in the market place (Chesbrough, 2010; Demil & Lecocq, 2010; Mitchell & Coles, 2003, 2004; Teece, 2010). Demil & Lecocq (2010) summarize this by stating that business models are “permanently in a state of transitory equilibrium” (p. 240), as new experiences need to be incorporated.

Both continuous and discontinuous change perspectives have expressed that it is more beneficial for organizations to change and innovate a business model proactively, before external pressures impose such a change (Hamel & Valikangas, 2003).

When comparing the research on business model innovation with the research on business models, two similarities stand out:

(15)

15 2) The research is also developing in silos, such as organizational learning (e.g. Sosna, Trevinyo-Rodríguez, & Velamuri, 2010), innovation management (e.g. Bucherer, Eisert, & Gassmann, 2012) or strategic management (e.g. Doz & Kosonen, 2010) Despite these aspects, a number of key ideas have been developed and researched over the course of the business model innovation history:

Based in the organizational learning domain, Sosna et al. (2010) describe the importance of

trial-and-error based learning for business model innovation in their Naturhouse case study.

Specifically, they report on how it took the company more than five years to develop their actual business model from an initial business model and how the learnings made throughout this journey were required to develop the actual one.

While the trial-and-error based approach already implies the use of feedback from customers in

shaping and fine-tuning a business model, Loïc, Lecocq, & Angot (2010) develop a framework

in which the customer is seen as a resource and thus as having an influence on the business model. This moves the discussion forward, by strengthening the role of the customer, making them an active participant and a helpful resource in the process of business model innovation. The process of transitioning between two business models is also an important aspect. Shifting from one business model to another, as it has been labelled by Massa & Tucci (2013), creates significant conflict and threat potential for the organization (Markides & Charitou, 2004). Overcoming these challenges is a major issue for organizations, which is of interest to both academia and practitioners alike. For academia the processes and tasks, required to successfully transition from one business model to another, need to be clarified and understood. Practitioners have a different perspective, being in charge of the transition, and raising the question of what pitfalls to avoid.

(16)

16 Given these challenges and risks, another question that has been asked is whether or not the benefits of business model innovation outweigh the risks for organizations? Markides & Charitou (2004) answer this question by stating that business model innovation can assist companies in differentiating themselves from other companies. An argument that has also been reported in IBM’s 2006 and 2008 CEO studies (IBM, 2006, 2008). The two studies found that the ability to differentiate a company’s services and products in the eyes of the customer is one of the main reasons CEOs believe in business model innovation. The importance of business model innovation and its ability to provide differentiation to the organization, have been found by multiple researchers (e.g. Amit & Zott, 2010; Chesbrough, 2010).

For researchers, the idea that business model innovation provides differentiation and potentially longer-lasting competitive advantages is attractive. This has its roots in the history of the strategic management field. As researchers in this field are interested in understanding how firms gain and potentially sustain a competitive advantage over their competitors (Barney, 1991; Porter, 1985), business model innovation has become one of the topics that seems to deliver this. Research (e.g. by Amit & Zott, 2012; Casadesus-Masanell & Zhu, 2013; Chesbrough, 2007; George & Bock, 2011) has shown a positive link between business model innovation and firm survival or growth, which can be seen as the representation of a competitive advantage over a firm’s competitors. Evidence to this relevancy is found in the publication of multiple special issues of journals revolving around the topic of business model innovation. For example the special issue of the International Journal of Innovation Management (edited by Spieth, Tidd, Matzler, Schneckenberg, & Vanhaverbeke, 2013) or Long Range Planning (edited by Baden-Fuller, Demil, Lecocq, & MacMillan, 2010).

From the theoretical definition point of view, Johnson et al. (2008), based on their research, argue that a new business model is required when “significant changes are needed to all four elements” (p. 57) of the current or existing business model. However, as this definition is quite

(17)

17 strict, the authors acknowledge that the judgement of the management team of an organization plays an important role in deciding whether a new business model is required or not (Johnson et al., 2008). This in mind, they further state that a significant change in two or three elements might be sufficient to constitute the need for a new business model or a change to the existing one, depending on the situation (Johnson et al., 2008).

Another, less abstract definition for business model innovation, is the one developed by Amit & Zott (2010). They define it as the “process of designing a new, or modifying” (p. 2) the (existing) business model.

A third definition is developed by Casadesus-Masanell & Zhu (2013), who define it as the “search for new logics of the firm, new ways to create and capture value“ (p. 464). According to the authors, the emphasis needs to be on discovering new ways the firm can generate revenues, while simultaneously defining “value propositions for customers, suppliers, and partners” (p. 464)

Interestingly, none of the definitions seem to incorporate the external perspective, or the reason why organizations start to search for alternative ways to configure their business model or to develop a new one. The idea of organizations weighing the potential benefits and risks beforehand is not incorporated into the definitions, yet it seems like an important step from a practical perspective given the risks described above.

The definition that will be used for the remainder of this thesis is the one provided by Johnson et al. (2008), as it allows a certain flexibility in discussing and evaluating business model innovation.

To summarize the discussions around business model innovation, Amit & Zott (2012) present three important arguments for why the discussion around business model innovation is important, to academics and practitioners alike. Firstly, business model innovation is

(18)

18 underutilized and represents a “source of future value” (p. 43). Secondly, it might be more challenging for competitors to “imitate or replicate” an entire business model (p. 43), than just a product innovation. Thirdly, managers of organizations must be aware of their competitors’ efforts to engage in business model innovation as well, emphasizing again the dynamism and continuity of business model innovation (Amit & Zott, 2012).

2.2.1 In search of a theoretical foundation

While the research on business models and business model innovation seems logical and self-explanatory, multiple researchers have pointed out the fact that it lacks a solid foundation in theory (Khanagha, Volberda, & Oshri, 2014; Markides, 2013; Teece, 2010). However, organizational ambidexterity might provide the theoretical foundation needed for business model innovation. It has a key similarity to business model innovation in one aspect:

In organizational ambidexterity, firms try to balance exploration (i.e. the “pursuit of new knowledge, of things that might come to be known” (Levinthal & March, 1993, p. 105)) and exploitation (i.e. the use of what the firm already knows) (He & Wong, 2004; Levinthal & March, 1993).

In business model innovation, firms have to balance the activities that serve the innovation of their business model(s) and the use of their existing business model(s). The latter is important, as the costs for innovation activities need to be financed by the existing business model(s). According to Markides (2013), this business model innovation problem (e.g. managing a current and a future business model at the same time) can be framed as an ambidexterity challenge. The idea that the two topics might be a fruitful combination and share certain commonalities has been articulated by several authors (e.g. Chesbrough, 2010; Markides, 2013; Spieth, Schneckenberg, & Ricart, 2014). Despite these theoretical links, very little research has

(19)

19 been done in testing or validating them (Markides, 2013). This thesis will therefor perform an analysis of this proposed link and will find evidence in favor or against it.

2.3 Organizational Ambidexterity

As outlined in the introduction and the previous section, this thesis will analyze the proposed link between business model innovation and organizational ambidexterity. Having discussed the concepts of business model and business model innovation, the following section will define and give an overview of organizational ambidexterity.

Organizational Ambidexterity is often described as the ability of organizations to balance exploration and exploitation. Exploration refers to the “pursuit of new knowledge, of things that might come to be known” (Levinthal & March, 1993, p. 105) and is characterized by “search, discovery, experimentation, risk taking and innovation” (He & Wong, 2004, p. 481), while exploitation describes increasing the utilization of what the organization already knows (He & Wong, 2004; Levinthal & March, 1993).

Over the last decades, research has suggested that failure to achieve such a balance potentially holds serious consequences for organizations (Benner & Tushman, 2003; Burgelman, 1991; Tushman & O’Reilly, 1996). The common agreement among researchers is that such imbalance will lead to either failure in the long-term (excessive focus on exploitation) or in the short-term (excessive focus on exploration) (Burgelman, 1991; Levinthal & March, 1993). Some researchers have argued that one root cause can be found in the dynamism and uncertainty of the environment for organizations which requires more than a simple exploitation strategy from them (O’Reilly & Tushman, 2013).

Numerous researchers find that organizational ambidexterity has a positive influence on e.g. innovation, sales growth, subjective performance ratings, or firm survival (a list of relevant

(20)

20 articles and additional factors can be found in O’Reilly & Tushman, 2013, p. 325), as organizations manage to maintain a balance between focusing on the present and the future. In this context, organizational ambidexterity is an umbrella term for the general tension to balance explorative and exploitative activities within organizations (Duncan, 1976; March, 1991). However, beyond identifying the risks, potential benefits and the general tension that ensues, researchers further pursued this area and developed different kinds of organizational ambidexterity. In their article, O’Reilly & Tushman (2013) list three major strains of ambidexterity research: Structural ambidexterity, sequential ambidexterity, and contextual

ambidexterity1.

The core idea of structural ambidexterity is to separate explorative and exploitative activities into different subunits of the firm, each with a dedicated purpose. This approach was first suggested by Duncan (1976), who proposed dual structures to manage the tensions between both kinds of activities. This separation of structures would take place at the “business unit or corporate level” (Raisch & Birkinshaw, 2008, p. 389).

Sequential ambidexterity refers to the changing of structures or processes when a company is

facing changes in its environment (O’Reilly & Tushman, 2013). This process of adjusting structures or processes has also been labelled temporal shifting, considering the idea that following only one side of the exploration-exploitation-continuum will lead to the above mentioned risks for the company.

A quite recent approach has been presented in the form of contextual ambidexterity (Birkinshaw & Gibson, 2004). This explores the idea that employees make their own, individual choices between conducting explorative or exploitative activities, based on their context (Birkinshaw

1_{More details on structural and contextual ambidexterity are available in Gibson & Birkinshaw}

(2004). Further details on sequential ambidexterity are presented in Raisch, Birkinshaw, Probst, & Tushman (2009)

(21)

21 & Gibson, 2004). Additionally, employees are rather generalists than specialists and the nature of their roles remains flexible (Birkinshaw & Gibson, 2004).

As these three approaches to organizational ambidexterity are fundamentally different and broad in nature, covering all of them, including their decades of research, would be unrealistic for this thesis. So, structural ambidexterity will be explored further, as its concepts and propositions can be applied to innovation activities, also allowing a measuring in research. Additionally, structural ambidexterity has been further developed and detailed by Raisch (2008). He developed three organizational structures, which can be used to achieve a balance between the long-term and short-term focus of an organization. His models are introduced in the next part in order to assess their usability for this thesis and the subsequent analysis of the link between business model innovation and ambidexterity.

2.3.1 Three forms of Structural Ambidexterity

The question of how organizations should balance exploration and exploitation, or an old and a new business model, has been raised by researchers in the organizational ambidexterity domain as well as in the business model innovation domain (Markides & Charitou, 2004; Massa & Tucci, 2013; Smith, Binns, & Tushman, 2010; Smith & Tushman, 2005). Leaning towards the structural ambidexterity domain, Raisch (2008) developed three organizational structures that he sees as fit to allow organizations to achieve growth, while maintaining a focus on current strengths. These models are labeled: Structural separation, parallel structures, and temporal

separation. These models will later be used as a baseline for how organizations structure

business model innovation activities.

2.3.1.1 Structural Separation

In this approach, explorative and exploitative activities are separated through dedicated structures within the company. Each unit focusses on short- and long-term goals respectively

(22)

22 (Raisch, 2008). However, this separation comes at the cost of an increased need for integration. So-called integration mechanisms, such as a culture of openness, a shared vision, or a common general manager (more mechanisms are presented in Markides, 2013), are needed to ensure coordination and integration between the activities in both parts of the organization (Markides, 2013). One common way to ensure integration is the top management team (TMT), mostly through the CEO or other members of the TMT being responsible for the integrative efforts (Markides, 2013).

Figure 1: Structural Separation according to Raisch (2008)

These efforts are especially relevant when considering that the different activities create fundamentally different requirements in different aspects of the unit. Organizational hierarchies, strategies, compensation systems, and many other factors will be different for explorative and exploitative units (Raisch & Birkinshaw, 2008). A potential risk for organizations pursuing this approach, is the importance of their leadership in supporting and managing this approach. Smith & Tushman (2005) report on the processes needed to effectively manage structural ambidexterity and the difficulties faced by organizations, especially with regard to the behavior exhibited by the top management team.

2.3.1.2 Parallel Structures

Comparing it to structural separation, parallel structures can be seen as a compromise in order to soften the weaknesses of structural separation. By allowing explorative and exploitative

(23)

23 activities at the same time, while using a partially separated structure, many disadvantages and risks can be mitigated (Raisch, Probst, & Gomez, 2007). Also, employees are more flexible as they are able to move between two kinds of structures, depending on the activity at hand. A more formal primary structure (e.g. an alignment to a department) and a secondary network-based structures that support more innovative activities (Raisch, 2008). As the separation is less structural here, one consequence is a reduced need for integration mechanisms, see chapter 2.3.1.1 Structural Separation, and thus reduced efforts required within the organization.

Figure 2: Parallel Structures according to Raisch (2008)

Additionally, parallel structures can help in reducing the risk to what Tushman & O’Reilly (1996) call “organizational integrity”, by allowing the organization to be more coherent in its structure. This is beneficial, as an improved integration (compared to e.g. structural separation) of units can allow the better exchange or utilization of knowledge (Miller, Fern, & Cardinal, 2007)

2.3.1.3 Temporal Separation

Whereas the other two forms emphasize different focal points for different units and different integration mechanisms, temporal separation describes a different approach. The goal is to align the entire organization towards either exploration or exploitation, depending on the situation the organization is facing (Raisch, 2008; Raisch et al., 2007). When the situation subsequently

(24)

24 changes, the alignment changes. An example could be an organization that has successfully commercialized a new technology and is now looking for ways to optimize its operations and profitability. Once this optimization is completed and the profitability of the new technology has slowed, the organization could move its entire structure towards exploration, again, focusing on customer demands and new technologies or business models (Raisch et al., 2007).

Figure 3: Temporal Separation based on Raisch (2008)

As research by Raisch & Birkinshaw (2008) indicates, a focus on explorative or exploitative activities requires different organizational structures, contexts, and strategies. Hence, organizations have to adapt these aspects, depending on the kind of focus it has chosen at a period in time. Research by Venkatraman, Lee, & Iyer (2007) supports this idea and finds evidence that cycling between explorative and exploitative organizational structures can have a positive effect on performance. Examples for this kind of behavior are given by Raisch (2008), with e.g. BMW group undergoing a group-wide integration and shared-services approach in 2000.

2.3.2 Assessing the usability of the three forms

Taking a more detailed look at Raisch’s (2008) three structures, two main factors can be identified to require further detail. First, the close resemblance of temporal separation to

sequential ambidexterity and, second, the unit of analysis applicable for the three forms.

Exploration

Exploitation

Time

Organizational Structure favoring exploitation Organizational Structure favoring exploration

(25)

25 The description and main characteristics of temporal separation, as described by Raisch (2008), bear at least a close resemblance to the concept of sequential ambidexterity. One could go as far as argue that they are the same. While excluding the concept of temporal separation (for being too similar) would be the easy thing to do, this thesis will follow a different route. While excluding temporal separation would be the easy thing to do, this thesis will follow a different route. In order to keep the idea of organizations changing their structure over time a part of this thesis, this aspect will be incorporated in the research design and the subsequent analysis.

Moreover, and generally speaking, all three forms of ambidexterity focus exclusively on the structure of organizations as a whole. Whereas the transition from one business model to another poses a significant challenge for an entire organization, smaller parts of said organization can easier cope with these challenges and risks. This is also the reason why business model innovation is often conducted in individual business units or (smaller scale) projects, as has been reported by multiple researchers (Bucherer et al., 2012; Doz & Kosonen, 2010; Girotra & Netessine, 2014; Johnson et al., 2008; Karimi & Walter, 2015; Sosna et al., 2010). Hence, a translation of the three structures and their characteristics is required, in order to apply them to projects.

2.4 Translating the three structures

As explained before, researchers have found that business model innovation often takes place in project-based environments. The structures Raisch (2008) developed however, are set on the organizational level. It is therefore necessary, to translate the structures from the organizational to the project level, in order to be able to apply them to business model innovation. This is also based on research by Turner, Swart, Maylor, & Antonacopoulou (2016) who state that projects “necessarily combine planning and control (exploitation) with some degree of novelty (exploration)” (p. 199), offering good context to research ambidexterity and the actions that

(26)

26 enable it. This focus on projects can also be found in other research (e.g. Eriksson, 2013; He & Wong, 2004; O’Reilly & Tushman, 2013).

The first step in breaking down the three models from an organizational perspective towards a project perspective, is to briefly recap their essence and main characteristics. This will be followed by a translation of how the concept would look like, when applied to a project environment:

2.4.1 Structural Separation

Summarizing, the concept of structural separation on the organizational level is mainly about separating certain functions, e.g. the innovation department, from other departments within the same organization (Duncan, 1976; Raisch & Birkinshaw, 2008).

Translating this separation to a project level, this structural separation means setting up a dedicated innovation project, equipped with dedicated financial and human resources, resulting in significant amounts of freedom (Raisch et al., 2007). In this context, reporting structures can be set up so the CEO or another member of the TMT present the project’s exclusive link to the organization, receiving progress reports from time to time (Raisch et al., 2007). However, researchers have reported that this structural separation on the project level poses a challenge e.g. with regard to the interaction and leveraging of existing knowledge (McCarthy & Gordon, 2011; Mirow, Hoelzle, & Gemuenden, 2008)

Additionally, such a project could also be disconnected in a physical sense. This could mean that e.g. the offices of the project could be in a different building, city, or country. All with the purpose of separating the project from the (physical) environment of the mother organization (Raisch et al., 2007). An intended consequence of this could be reduce the exchange between members of the mother organization and members of the project.

(27)

27

2.4.2 Parallel Structures

Contrary to structural separation, parallel structures on the organizational level are about creating a second layer of exchange and information flows (Raisch, 2008). One example of such layers are innovation units. In a parallel structure, the innovation unit is linked to other departments in order to facilitate the exchange of knowledge or the sharing of relevant information e.g. from the client-facing sales team (Raisch, 2008).

In a similar fashion, on the project level, connections between the project and the existing organizational structure should be created. Considering the example of reporting structures, such connections can be established through a steering committee in which different departments are represented. Another approach is to form cross-functional teams from various departments, not as full time members, but as part time representatives of their respective department.

2.4.3 Temporal Separation

As outline previously, temporal separation is about aligning the entire organization with either exploration or exploitation, depending on the business context the firm is facing (Raisch, 2008; Raisch et al., 2007). This allows firms to create a strong emphasis on current business needs and a focus on solving them. This focus enables “radical changes to the firm’s existing operating processes” (Raisch, 2008, p. 487) and represents a benefit for the firm in times of uncertain or changing environments (Raisch, 2008).

While the benefits on the organizational level appear clear and consistent, they are much less so for the project level.

(28)

28 Whereas organizations are able to shift their structure over time and adjust to changing environments (Devins & Kähr, 2010; Gibson & Birkinshaw, 2004; Raisch, 2008), projects are usually initiated with a distinctive idea or goal in mind2_{. In a project setting, the time perspective}

can be seen as being represented by two ideas: The willingness to change the project given new circumstances or evidence and the anticipation of success, i.e. the expectation that e.g. a small scale business model innovation study is expanded, once proven successful.

Given these close similarities and characteristics, temporal separation will not be included as a separate form during the analysis, but rather as a part of the analysis. This will be done, by analyzing whether or not organizations anticipate future changes and success in their project setups, e.g. by implementing departments early on, whose value is only realized in the case of project success. The relevant data for this part of the analysis will be collected by adding targeted question into the data collection process.

2.5 Combining Organizational Ambidexterity and Business Model

Innovation

After having established the concepts of business model innovation and organizational ambidexterity, the next chapter will point out the similarities between the two, in order to identify the research gap in the following chapter.

Identical for both topics is the emphasis on balancing short-term and long-term prospects of the firm. In the organizational ambidexterity literature, the emphasis is on the parallel exploitation of existing and the exploration of unknown knowledge (He & Wong, 2004; Levinthal & March, 1993). Similarly, business model innovation focusses on using the existing business model (i.e. the short-term perspective), while, in parallel, developing or implementing a new business

2_{The Project Management Institute (1996) defines it as a “temporary endeavor undertaken to}

create a unique product or service", highlighting that a certain outcome is expected or anticipated.

(29)

29 model (i.e. the long-term perspective) (Casadesus-Masanell & Tarzijan, 2012; Doz & Kosonen, 2010; George & Bock, 2011; Velu & Stiles, 2013). This view is supported by the recent development of the idea that firms may run a portfolio of business models, adding new ones and removing older ones over time (Osterwalder, Pigneur, & Tucci, 2005; Sabatier, Mangematin, & Rousselle, 2010).

Another similarity is the notion that failure to follow both sides of the argument can lead to potentially negative outcomes for the firm. In the organizational ambidexterity literature this is referred to as learning traps or myopia of learning (Levinthal & March, 1993). In the business model innovation literature this is often linked to negative firm performance or a firm’s failure in the market (e.g. by Amit & Zott, 2012; Casadesus-Masanell & Zhu, 2013; Chesbrough, 2007; George & Bock, 2011).

In the last years, these similarities have been pointed at more directly. Chesbrough & Rosenbloom (2002) stated that ambidextrous processes could be beneficial to firms trying to manage innovation and that the concept of business models could be used in this context. Chesbrough (2010) later re-iterates this conceptual link when he emphasizes the importance of organizational capabilities to perceive the environment and shift resources to new business models as required. Markides (2013) makes a stronger link, when he describes the lack of a theoretical foundation for the business model innovation research and proposes ambidexterity as a “natural theoretical anchor” (p. 315). One aspect that the business model innovation literature could incorporate from ambidexterity research are the concepts for how to organize exploration and exploitation within a firm (Markides, 2013), i.e. managing the tensions between the two endeavors. For firms trying to organize and coordinate their activities in business model innovation, there is little theory that could be used (Markides, 2013). This thesis therefor raises the question whether or not firms follow the results of the structural ambidexterity research, namely the two forms, in organizing and structuring their business model innovation efforts.

(30)

30 Also of interest, are the factors that influence a firm’ choice for a structure, when conducting business model innovation. Based on this, the following research question will be answered in this thesis, being a refined version of the question raised in the introduction:

RQ: How, if at all, are organizations using the forms of structural ambidexterity to organize their business model innovation activities?

(31)

31

3 Analytical Framework

The research framework used in this thesis is based on a literature review of business model innovation as well as project success and key factors identified therein. As outlined previously, the use of resources is frequently highlighted by business model innovation researchers as a key factor in these activities (Amit & Zott, 2001, 2012; Chesbrough, 2007; Johnson et al., 2008; Zott & Amit, 2010). As has been described, business model innovation activities are often conducted as projects or smaller initiatives (Bucherer et al., 2012; Doz & Kosonen, 2010; Girotra & Netessine, 2014; Johnson et al., 2008; Karimi & Walter, 2015). To combine the two fields, it is therefore appropriate to look toward the project management or project success literature for relevant resource classifications. In their seminal work, Belassi and Tukel (1996) review a large number of studies on project success and corresponding success factors. In their reviewed studies, as well as in the framework they propose, resources play an important role (Belassi & Tukel, 1996). The authors emphasize that sufficient financial resources and the provision of sufficient and skilled human resources are critical factors for project success (Belassi & Tukel, 1996). This emphasis is based on several reviewed articles and project success frameworks, in which the endowment of resources and provision of skilled human resources is highlighted (Baker, Murphy, & Fisher, 1983; Cleland & King, 1985; Martin, 1976). Two additional dimensions for the framework are based on the dynamism of business model innovation. This dynamism (as explained before) highlights the need for organizations to be aware of their environment and competitors, which could potentially replicate their business model; meaning that continuous business model innovation (Mitchell & Coles, 2003, 2004) is required. This aspect is highlighted in several publications through case study analysis (Bucherer et al., 2012; Sosna et al., 2010). One key criterion is the knowledge derived from previous business model iterations. This aspect of learning from previous experience or projects also has been identified as a project success factor by multiple researchers (Baker et al., 1983;

(32)

32 Cooke-Davies, 2002; Dvir, Lipovetsky, Shenhar, & Tishler, 1998; Sauer, 1993; Yap, Soh, & Raman, 1992). From a broader perspective, learning from previous experience in a business model innovation context is both an important input factor and a key result. Only when organizations specifically account for potential knowledge, can they apply this knowledge to subsequent projects or initiatives. This is for example relevant to business models that undergo multiple iterations, as has been reported by Sosna et al. (2010). Therefore, this research framework includes learning as both an input and output for business model innovation activities.

Combining these four elements into a unified model, the following framework is applied in this thesis:

Figure 4: Analytical Framework - schematic

Given the use of the forms of structural separation, each of the four elements can take three possible values:

1) Use of structural separation 2) Use of parallel structures

3) Use of mixed structures (combinations of structural separation and parallel structures) 4) Use of other, non-structurally separated forms

Business Model Innovation Activities (Projects) Financial Resources Human Resources Past Organizational Learning (“Organizational Knowledge”) New Organizational Learning Unit of analysis Input / Output Factors

(33)

33 Combining Figure 4 and the translation of the forms of structural separation described in Chapter 2.4, Table 2 represents the model in detail. The “use of mixed structures” is omitted in this table since it solely represents a mix of the two forms of structural ambidexterity.

Dimension Structural Separation Parallel Structures Non-structurally _{separated forms}

Financial Resources - Centralized budgeting and reporting - Decentralized budgeting and reporting - External financing - Self-financing - No-dedicated financing Human Resources - Centrally organized and managed - Dedicated human resources - Non-centrally organized and sourced - Shared human resources - External people Past Organizational Learning (“Organizational Knowledge”) - Through TMT - Rare exchange - Indirect - Through business units - Continuously - Directly from knowledge carriers - No leveraging of existing knowledge or capabilities New Organizational Learning - Through TMT - Rare exchange of learnings - Continuously - Less planned - Directly between people - No exchange

Table 2: Analytical Framework

The framework’s goal is to capture various aspects relevant to the organization of business model innovation activities. While the framework focuses on the choice of structural ambidexterity and its models, it also anticipates potentially different results. This is achieved using a third category - “non-structurally separated forms.” Three possible scenarios come to mind for this third category.

First, organizations could be using alternative forms of structural ambidexterity not covered in the existing literature or not yet fully developed. Second, while this thesis focuses on structural

(34)

34 ambidexterity, companies could be using contextual or sequential ambidexterity, which would fall outside of the focus of this thesis. Third, it could also be evidence that the link between organizational ambidexterity and business model innovation is not as pronounced as anticipated. For these reasons, the analysis and discussion sections of the thesis will also evaluate and discuss whether evidence against the use of structural ambidexterity was found in the data. In conclusion, the four dimensions can take the following values:

1) Use of structural separation 2) Use of parallel structures

3) Use of mixed structures (combinations of structural separation and parallel structures) 4) Use of other, non-structurally separated forms:

a. Other forms, not discovered or developed by research b. Other forms of ambidexterity used

c. Link not as profound as anticipated

To provide further clarity about the model, the dimensions and their theoretical origins are explained in the following paragraphs:

1) Financial Resources: This dimension evaluates how business model innovation activities are financed. In a structurally separated model, for example, centralized budgeting for the business model innovation activity exists and reporting processes should be in place. The financial resources are provided by a central part of the organization, e.g., the CEO or the TMT. On the other hand, in a parallel structure model, financial resources are provided by different business units, with decentralized reporting processes in place. In the scenario that no evidence can be found, it could be the case that the business model innovation activities are, for example, externally financed (e.g., through capital markets), self-financed (e.g., through the selling of goods or services), or no financial resources are required and, therefore, no need for a budget exists.

(35)

35 Different authors have described these topics; for example, Gertner, Scharfstein, and Stein (1994) describe the possibility of internal or external capital markets and Morton I. Kamien (1978) covers the self-financing aspect of R&D projects, which are assumed to be similar to business model innovation projects. The existence of projects with no dedicated budget is reported for multiple companies in a journal article by Gray (2001) 2) Human Resources: In this dimension, the delegation of personnel for business model

innovation activities is evaluated. In the structural separation form, human resources are managed and organized centrally (e.g. through a central project management office). Dedicated human resources are provided to the projects to further strengthen their separation from everyday business and operations. In the parallel structures form, human resources are decentralized (i.e., drawn from different business units) and, to a lesser extent, managed from a central point within the organization. Overall, human resources can be considered shared rather than dedicated. In the third perspective, i.e., a non-structurally separated model, human resources are sourced from external partners (e.g., consultants) and not from within the organization. Researchers have identified three factors over the last decades. For example, the sharing of resources has been identified as a project risk in the research of Meier, Zimmermann, and Nicolau (2015). Moreover, the importance of the human resource function for the success of projects has been discussed by several authors (e.g., Belout (1998); de Wit (1988); Henard & Szymanski (2001)), with some authors listing dedicated human resources as a project success factor

3) Past Organizational Learning (“Organizational Knowledge”): Many organizations have an existing set of knowledge or skills that could potentially be beneficial to business model innovation activities. The importance of using these existing assets has been described by many authors (e.g., Schindler & Eppler (2003); Sosna et al. (2010); Wiig (1997)). Potential benefits are reported by Johnson et al. (2008) in their Dow Corning

(36)

36 case study, where the authors state that the company “did not want to give up the incumbency advantage that deep knowledge of the industry and of their own products gave them. The challenge was to tap into the expertise without importing the old-rules mind-set” (p. 59). In a structural separation model, the contribution mainly occurs through the TMT and in an indirect fashion. In addition, the leveraging of knowledge would be rare as organizations would attempt to abandon the old-rules mind-set that comes along with using existing knowledge. In a parallel structure form, the contribution of existing knowledge is more frequent as different people from different parts of the organization can contribute different knowledge sets. Further, access to the knowledge owners is more direct, through the direct involvement of the business units. In a non-structurally separated form, there would be no leveraging of past learning and no structures in place to ensure that this knowledge is used by projects or initiatives 4) New Organizational Learnings: Since the potential results of business model innovation

activities are not only a new or changed business model for the company, other learning could occur as well. For example, Sosna et al. (2010) demonstrate that several iterations may be necessary before a firm gets its business model right. This dimension tries to analyze whether companies anticipate new knowledge and actively try to incorporate and share it across the organization. The importance of knowledge exchange within an organization has been discussed by multiple authors over the last decades (e.g., Chen & Huang, 2009; Lee & Choi, 2003; Stonehouse & Pemberton, 1999), especially with a focus on how organizations can learn from past project success or failure and apply this knowledge to improving future projects (Pinto & Thoms, 1999). In a structurally separated model, the exchange or capture of knowledge would mainly happen through the TMT and the reporting processes in place. The flow of information is mainly standardized through the top layer of the organization. This would also imply a rare frequency of exchange. In a parallel structures model, the exchange of information and

(37)

37 knowledge is continuous due to people being a part of their organization and the project structure. Hence, it is assumed that the exchange would be less organized or formalized as these forms of exchange could happen naturally when people from different backgrounds or business units meet. In a non-structurally separated model, the exchange of knowledge would be of less importance as the people involved would not be separated through structures and processes

(38)

38

4 Research design

As stated in the research question, this thesis will examine the theoretical link between ambidexterity and business model innovation. The goal is to find evidence in favor or against this association. This will be achieved by looking at the essence of the three forms as proposed by Raisch (2008) and evidence gathered from interviews and document analysis.

A case study approach was selected for several reasons. First, business model innovation is not an idiosyncratic activity; that is, the context in which it takes place is important when the existing business model is being transformed. A case study allows and recommends the collection and description of context to effectively analyze and frame the results (Eisenhardt, 1989; Patton, 1999). Moreover, through contextual analysis we can explore why organizations conduct business model innovation the way they do.

Second, as this is an explanatory thesis, predicting the results is difficult. Hence, it would be speculative at best, to try to anticipate all possible answers in a survey. Thus, an interview format is more appropriate for collecting data using targeted and open-ended questions. This allows the interviewer to probe deeper into emerging topics during the course of the interview (Bryman & Bell, 2015; Saunders, Lewis, & Thornhill, 2015).

Third and last, business model innovation is an activity that often involves an organization’s senior management (e.g., the cases in Bucherer et al., 2012; Johnson et al., 2008). Therefore, involving CEOs or other members of the TMT is critical for the collection of meaningful data. While it would be easy to send out a survey to a number of CEOs, the risk is that the organizational background and context may be misunderstood. Using case studies based on semi-structured interviews allows for individual discussions prior to the interviews to provide interviewees with more information about the study.

(39)

39

4.1 Research Criteria

From a research methodology perspective, the validity (internal and external) and reliability of results are highly relevant.

Validity (internal and external) and reliability can be assured in several ways. Gibbert, Ruigrok, and Wicki (2008) recommend three approaches to ensure internal validity: 1) a clear research framework (see Chapter 3); 2) pattern matching (i.e., matching the research framework with the results via the interview questions); and 3) theory triangulation, which is achieved through multiple case studies across different organizations and industries while using different sources of evidence (e.g., interviews, annual reports, e-mail conversations, and press releases).

Although external validity, even with multiple case studies, is commonly low in qualitative research (Yin, 2013), Eisenhardt (1989) argues that using at least four different cases and a cross-case analysis can significantly enhance generalizability. Both approaches are used in this thesis. Yin (2013) adds that conducting multiple cases inside an organization is another option to improve the external validity of a study, such as the case study of Epsilon, where two projects were analyzed. Lastly, Gibbert et al. (2008) recommend authors provide a clear rationale for case selection and “ample details on the case study context.” All these aspects have been considered in the study design, thus significantly improving this paper’s credibility. Furthermore, reliability is ensured through using both a case study protocol (in the form of the steps undertaken to collect the cases) and a case study database, which are presented in Chapters 5.1 and 5.5.

Finally, the use of diverse cases further increases the validity of this paper; i.e., both embedded (e.g., company Epsilon) and holistic (e.g., company Delta). This is in line with the classification of Yin (2013), as shown in Figure 5.

(40)

40

Figure 5: Types of case studies according to Yin (2013)

4.2 Linking theory to research design

To maintain the chain of evidence, the analytical framework developed in Chapter 3 is linked to the questions used in the interview process. This is achieved by matching the questions with the dimensions of the analytical framework. This matching is shown in Appendix 2: Linking Questions to Analytical Framework

The interview questions start with broad and general contextual questions, thus allowing the interviewees to describe the business model innovation project with their own words and to become comfortable in the interview setting. The second set of questions target the four dimensions of the analytical framework explore the details of their projects.

Context Case Context Case Context Case Context Case Context Case Context Case Context Case Context Case Context Case Context Case Embedded Unit of Analysis I Embedded Unit of Analysis II Embedded Unit of Analysis I Embedded Unit of Analysis II Embedded Unit of Analysis I Embedded Unit of Analysis II Embedded Unit of Analysis I Embedded Unit of Analysis II Embedded Unit of Analysis I Embedded Unit of Analysis II single-case designs multiple-case designs holistic (single-unit of analysis) embedded (multiple units of analysis)

(41)

41

4.3 Qualitative Data Analysis Strategy

For the analysis of qualitative data, Yin (2013) presents four general strategies. One of these was chosen as it fits the research question and approach for this thesis:

Relying on theoretical propositions: The analytical framework, as described in Chapter 3, and the proposed link between business model innovation and ambidexterity, provide a sufficient theoretical framework. This framework not only allows analysis of the data, but also allows evaluation and application to the research question and the primary aim.

Data from the interviews and complementary documents are transposed through a coding process that assigns a keyword or short statements to specific responses, thus capturing the essence of that response (Saldaña, 2016). Although codes can be developed before or during the analysis (Bourque, 2004), in alignment with the deductive nature of this thesis, the coding was developed prior to the analysis using the theoretical framework from Chapter 3 as a foundation (Bourque, 2004). This links the theoretical framework to the interview guide and the data to be analyzed.

The codes are entered into nVivo, which manages the data, conducts analyses and database queries, and provides reports for the evaluation of the data.

(42)

42

5 Case Studies

5.1 Case Study Development and Research Process

Collecting data in a structured manner requires a clear and transparent case study (i.e., data gathering process). Table 3 summarizes the process used in all cases incorporated in this study to allow comparability between cases. Appendix 3 shows a table with more details for each stage. Stage 1) Identification 2) Initial Discussions 3) Interviewing 4) Transcription

5) Feedback and Approval

6) Complementary Data Gathering Table 3: Case Study Development Process Overview

To allow comparability between cases, this process was applied to all case data included in this thesis.

5.2 Sample Selection

Given the focus on business model innovation, companies included in the sample should reflect such processes. Hence, relevant organizations either have completed the process of developing or implementing a new business model or are currently in the midst of this process. This criterion is chosen since both situations allow an examination of the proposed context, i.e., the structures used to conduct business model innovation. This context is of particular relevance as business model innovation is unlikely to happen in the short run. Rather, companies will make

(43)

43 careful steps in the form of projects when trying to validate and then implement an altered or new business model.

To identify suitable companies for this case study, two sampling strategies were used:

1) Convenience sampling: This sampling strategy allowed the author to make use of his personal and professional network to identify potential companies (Saunders et al., 2015). Acquaintances, colleagues, family, and friends were asked to identify whether their company or companies they are aware of, had conducted activities that led to a change in an existing business model or the development of a completely new business model.

2) Snowball sampling: During the data collection process, interviewees were asked to think of other, potentially interesting, interview candidates or projects (Saunders et al., 2015). This method proved successful as the interviewees were working within existing organizations, with access to colleagues and projects. Additionally, the author’s personal and professional networks were asked to help identify suitable companies. This also proved helpful in generating leads for further case studies.

5.3 Data Collection – The interview process

Data were collected from semi-structured interviews comprised according to the description in the previous section. This approach is chosen because, compared with structured interviews, semi-structured interviews allow for flexibility in terms of order of questions (Bryman & Bell, 2015; Saunders et al., 2015). Moreover, semi-structured interviews offer the opportunity of topic probing, i.e., following up on topics that emerge during the interviews and are seen as relevant or interesting (Bryman & Bell, 2015).

(44)

44 An interview guide was developed prior to the interviews3 that contains three sets of questions. The first set is designed to investigate the nature of the business model innovation activity and the origin of the idea or concept. The second set is aimed at capturing the actual project structure used in conducting the business model innovation activities. The third set is concerned with the actual implementation and potential difficulties that the case companies faced.

The generic guide was then customized so the business model innovation could be assigned with the actual project name. The generic word “organization” was also replaced with the respective company name to improve identification with the question by the interviewee. IN addition, company-specific terms, such as the term “publisher” in the case of the newspaper, were also included in the customized versions.

For transcribing purposes, each interview was recorded with the consent of the interviewee. This allows an analysis and evaluation as close to the spoken word as possible. All interviews were conducted in German or English and transcribed in their original language.

While the interview process (i.e., step 3 in Table 3) is the most important part of the data collection, step 2 (initial discussion) is also significant in determining the relevant input for step 3. Not only does it provide the opportunity to validate the author’s assumptions with regard to the nature of business model innovation, , it also allows a rapport to be established with the interviewee, improves the interview quality (Bryman & Bell, 2015), and validates the significance of new or changed business models. It also allows the exclusion of organizations in which the form of innovation is more product- than business-related. In the case of a medium-sized brewery, for example, this initial discussion led to the exclusion of the case as it was solely a product innovation and had no impact on the business model of the brewery.

Business model innovation and organizational ambidexterity : theoretical foundation or superficial link?