Incumbent-disruptor dynamics & disruptive innovation in a digital age : disruptor disrupted : a case study of the Dutch energy sector

(1)

Incumbent-Disruptor Dynamics & Disruptive Innovation in a Digital Age

Disruptor Disrupted: a Case Study of the Dutch Energy Sector.

University of Amsterdam Master Thesis

Master of Business Administration – Digital Business Track

Supervisor: Prof. Dr. H.P. (Hans) Borgman

Name: Lara Plandsoen Student No: 11411295 Email: l.l.plandsoen@gmail.com

(2)

Statement of originality

This document is written by Lara Plandsoen who declares to take full responsibility for the contents of this document.

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.

(3)

Preface

This thesis is the final step in completing the Master Business Administration at the University of Amsterdam. Following the newly created Digital Business track of the program has confirmed and extended my interest in the overlapping field of business and information technology. I can truly say that I have enjoyed all courses in the program and that I feel prepared to take the next step in my career.

Firstly, I would like to thank my supervisor Prof. Dr. Hans Borgman for his critical suggestions, support and insightful comments that have helped me to push my limits and achieve and end result I am genuinely proud of.

Secondly, I want to thank all people that have me helped understand and gain deep insights in the world of incumbents and disruptors, the world of disruptive innovation, and last but not least the Dutch energy sector. To all people that I have interviewed, held discussions with and asked for more information time and again: thank you, all of your comments have been extremely valuable and helpful.

Last but not least I want to thank the most loyal team of support one could wish for when doing such a project: my family and my girlfriend. Their continuous motivation, support and incredible proofreading skills have helped me get through this process while at the same time helping me improve the end result time and again.

(5)

2

Abstract

The purpose of this research is to provide an understanding of managerial factors influencing differences in responses of young-incumbents and mature-incumbents to upcoming waves of disruptive innovation. Based on a literature review three propositions are derived which in turn guide semi-structured interviews with management-level employees of strategy and innovation departments in a multiple case study of the Dutch energy sector. This research proposes a new categorization of the different actors in a situation of disruptive innovation in which a clear distinction is made between mature-incumbents and young-incumbents. The findings of this research indicate there is indeed a need for more variation in the incumbents category as there are clear differences in managerial attitudes and responses to disruptive innovations between the two groups of incumbents. Therefore, this research implicates that young-incumbents should be seen and reckoned with as a stand-alone category between the disruptors and mature-incumbents. Due to the industry-specific nature of the case-study, further research is necessary to test generalizability of the theory. This research is consistent with other studies emphasizing the important role that cognitive managerial factors and experience play in an organization’s response to disruptive innovations. I expand on this literature by proposing a tripartite rather than a bipartite of actors in a situation of disruptive innovation, and suggest managerial factors in which the organizations of the two newly created categories differ that can lead to differences in response to disruptive innovation. The added category implicates a reassessment of practitioners’ current view of entrants versus incumbents. Results of the research can be used by young-incumbents and mature-incumbents alike to assess not only their competitor’s attitude, but also their own attitude and possible responses to an upcoming wave of disruptive innovation in order to strengthen their competitive position.

Keywords: disruptive innovation, incumbents, cognitive managerial factors, dynamic capabilities

(6)

3

1. Introduction – Disrupting the disruptor?

The purpose of this research is to provide an understanding of the differences in responses of young-incumbents and mature-incumbents to upcoming waves of disruptive innovation. The term ‘disruptive innovation’ was first coined by Clayton Christensen in 1995 and has been a popular subject of research ever since. Whether the innovation has been described as radical (Ansari & Krop, 2012), discontinuous (Macher & Richman, 2004) or disruptive (Bower & Christensen, 1995; Christensen, 1997), and whether the two acting parties are called incumbents and entrants (Diekhof, 2015), challengers (Ansari & Krop, 2012), attackers (Foster, 1986) or disruptors (Bower & Christensen, 1995) disrupting the disruptee (Lacourbe, 2013). Although thelabelmight be different, the same concept is being researched. Disruptors unexpectedly appear on a well-established mainstream market after having occupied a niche market for a certain period of time, only to be noticed by incumbents that thought they were safe and sound when it is already too late (Henderson & Clark, 1990). The disruptors have found an entrance and won enough ground to be able to say: we are here to stay (Christensen, 1997).

Early research in the subject area focused on discontinuous technological innovations (Bower & Christensen, 1995; Christensen, 1997). Later, the concept was extended by Christensen and Raynor (2003) to include both technological and business model disruptions in the definition of ‘disruptive innovation’. This more comprehensive term covering both technological and business model innovations will be the definition of the concept of ‘disruptive innovation’ used in this research, as Christensen (2006) as well as Markides (2006) have pointed out that disruptive technology has become the antecedent to disruptive impact of business model innovations. Or, as Chesbrough states, “today, innovation must include business models, rather than just technology and R&D” (2007, p. 12).

Research has covered the further review of the disruptive innovation theory (Danneels, 2004; Schmidt & Druehl, 2008; Yu & Hang, 2010), incumbent-entrant dynamics (Acemoglu & Cao, 2015; Ansari & Krop, 2012), potential responses to disruptive innovation (Macher & Richman, 2004; O’Reilly & Tushman, 2008), and factors influencing incumbents’ responses (Sandström, Magnusson, & Jörnmark, 2009). Yet further theoretical groundwork is necessary to clarify variation in incumbent’s responses

(7)

4

to technological change (Lavie, 2006). An interesting question that has remained unanswered would thus be: what happens next? In the situation described before (assuming the incumbents survive the wave of disruption), these incumbents will experience a true wake-up call. They will likely want to make sure that they do not find themselves in a similar situation again. This wave of disruption might be over and the incumbent may not have seen it coming, however, does this imply they will be better prepared for the next wave or will they fall into the same trap again?

Another question that comes up: what happens to the categorization of the incumbent and disruptor in case a new wave of disruption arrives? According to current literature, the disruptors from the previous wave will now be categorized as incumbents responsible for a considerable market share, revenue, growth and profit (Bradley et al., 2015; Schmidt & Druehl, 2008). After all, new disruptors are on their way. But have those entrants really become incumbents already? Can we tar those new incumbents with the same brush as the older incumbents when discussing possible, necessary or optimal responses towards the newly emerging wave of disruptive innovation? After all, major differences not only in market share and resources, but also in business model, strategy, management, company culture and other factors are evident, as will be illustrated later in this thesis.

Current literature provides us with useful theories and handles to assess and analyze organization’s responses to disruptive innovations in the light of the traditional way of categorization of disruptor versus incumbent (Macher & Richman, 2004; Sandström et al., 2009; Gemici & Alpkan, 2015). When discussing suggestions for further research, Sandström et al. state that “the heterogeneity of incumbents has been downplayed by the previous literature and it calls for further investigations to allow for the development of a more nuanced view of how established firms can respond to disruptive innovations” (2009, p. 14). Hence, with respect to the rising frequency and intensity of waves of disruptive innovation nowadays it is important to consider variation in incumbent heterogeneity. This is necessary in order to properly predict, assess, analyze and advice on the appropriate response for a given organization to an emerging wave of disruptive innovation.

The most important reason for this need to investigate heterogeneity of incumbents is that the current division is too simplified. From today’s broadly accepted

(8)

incumbent-5

disruptor dynamics point of view: if a new technology comes up that has the potential to disrupt the market, the disruptors of the previous wave will be categorized with the incumbents as they are now established players in the market. Should we however expect them to respond to this new disruptive trend in the same way as the ‘old’incumbents? As Sandström et al. put it, “in the discourse regarding disruptive innovation, incumbents are often treated as one population vis-à-vis entrants rather than as many populations with different resources, market positions and strategies” (2009, p. 8). Osiyevskyy and Dewald add that “incumbent responses to a technology that has a disruptive potential do not have to be homogenous” (2015, p. 60).

Can the current academic theories on incumbents’ response to disruptive innovations be applied to these ‘young’ incumbents? The youngsters might still see themselves as disruptors, as quite recently they have successfully entered a market long occupied by the old incumbents through deploying innovative business models and disruptive technologies, providing value to customers in an innovative way. They see themselves as inherently different from the ‘old’ incumbents. Perhaps it is time that the academic community starts doing the same by taking this distinction into account?

The distinction between ‘old’ or ‘mature’-incumbents, and the ‘new’ or ‘young’-incumbents will be illustrated in this research by taking the Dutch energy sector as an example. In the category of incumbents reside the long-established firms such as Nuon (1994), Eneco (1995), and Essent (1999). With a combined experience of over 60 years these three companies can be called the mature-incumbents of the Dutch energy industry. In the last few years however, several new firms have made their entrance. Examples are NLE, Qurrent, EnergieFlex and EnergyZero. All are less than twelve years old and all have entered the market with a proposition that had the potential to disrupt the status quo. Either in terms of technological advances or in terms of business models and value propositions.

Up until the liberalization of the Dutch energy market in 2004, all ‘big three’ incumbents were comfortably positioned in a relatively stable market. After the liberalization this started to change. The emergence of the internet facilitated new ways for customers to compare prices and advantages, and the entering organizations offered business models and value propositions that had not been seen before. This forced the ‘big three’ incumbents to adapt in order to stay in the game.

(9)

6

One example of a newly emerging business model introduced by an entrant positions the entrant as an intermediary, efficiently buying and reselling energy rather than producing it through capital intensive assets, resulting in lower prices. Another entrant offered energy in the form of a prepaid payment model. A third one harnessed technology in the form of software development to help customers save energy. Yet another entrant focused on the green, environmental aspect of the energy sold as their main value proposition while another entrant again took it a step further and acts a facilitator of peer-to-peer trading of surplus green energy produced. The change of the competitive environment produced by this plethora of new business models was further enhanced/empowered by rapid developments in technology such as smart homes, the internet of things, the rise in popularity of electric vehicles, rapid developments in battery capacity and peer-to-peer trading of renewable energy and blockchain technology. With the passing of the years, those entrants have developed themselves into full-grown companies. However, the development of technology does not stop. Therefore, both the party of former entrants as well as the ‘big three’ incumbents are now facing new waves of disruptive innovation. Will their responses be similar? Or will they employ different strategies in order to deal with the upcoming changes and developments in the market? In this thesis I propose a new framework of categorization of the different actors in a situation of disruptive innovation, distinguishing the organizations that have just ‘become’ an incumbent: ‘young-incumbents’, from the older ones: ‘mature-incumbents’ that were already considered incumbents during the previous wave of disruptive innovation.

The assumption in this research is that mature-incumbents do possess more of the ‘obvious advantages’, as Teece (1986) dubbed an incumbent’s resources and scope, than young-incumbents possess. Therefore, the focus will lie on the different managerial factors that set the two categories apart.

Levinthal & March (1993) display cognitive biases as characteristics of decision-makers that could create differences in organizational learning and future performance. These cognitive biases could hence play an important role when assessing inter-incumbent variance. Later on in the development of the research stream of disruptive innovation, many leading authors on the subject emphasize the importance of managerial cognition and decision making when analyzing incumbent responses to disruptive innovations as

(10)

7

cognitive biases and experiences from the past provide a framework aiding management decision making in situations of uncertainty (Benner & Tripsas, 2012; King & Tucci, 2002; Osiyevskyy & Dewald, 2015; Teece, 2007; Tripsas & Gavetti, 2000). Tripsas and Gavetti (2000) explored how managerial cognitive representations play a central role in the way an organization responds to disruptive change. Osiyevskyy and Dewald further examine this subject by arguing that “heterogenous incumbent strategies are driven by divergent managerial beliefs and perceptions of the disruptive approach” (2015, p. 60). King & Tucci (2002) focus their research on the role of experience and managerial choice influencing responses to disruptive innovation.

This research will follow suit in the focus of the research stream on cognitive biases and experiences from the past as a framework for managerial decision making. Hence, this research is performed with the strategic agency perspective in mind, implying that “decisions of leaders precede the behavior of the organization” (Osiyevskyy & Dewald, 2015, p. 63).

To summarize, in line with the strategic agency principle thinking, the aim of this research is to assess the necessity of the newly proposed categorization of young- and mature-incumbents and highlight differences in managerial factors that will result in different responses to disruptive innovations. This will answer the research question:

How do young- and mature-incumbents differ from each other with regards to managerial factors at play in a situation of disruptive

innovation?

Hence, I do not aim to provide a comprehensive overview of all possible factors involved in the dynamics and differences between mature and young-incumbents in response to disruptive innovations. If the results of this research show no significant differences in managerial factors regarding responses to disruptive innovations, there is considerably less benefit in further pursuing research on young-incumbents and mature-incumbents as two separate categories in the framework. In that case, the original incumbent-disruptor model will continue to apply. However, if there is reason to assume the current framework is insufficient, the inquiry of separate categories of mature- and young incumbents and their different responses towards disruptive innovations is well justified.

(11)

8

Propositions are offered based on an overarching assumption that mature-incumbents possess resources (and capabilities) that differ from those of young-incumbents, and that both groups have to adhere to the same set of governmental rules and regulations. The rest of this thesis is organized as follows. The next section reviews the existing literature from which several propositions are derived with regards to the new categorization of mature- and young-incumbents in the light of a new wave of disruption. The section thereafter provides an account of the research methods applied and analysis performed. The subsequent section illustrates and tests the constructed propositions by means of a multiple case-study of the Dutch energy sector. The final section contains an analysis of the case study and a discussion of managerial and theoretical implications of the newly proposed framework of categorization of mature- and young-incumbents.

(12)

9

2. Theory & Propositions

Based on the extant academic literature I offer a set of propositions outlining the different managerial factors that are expected to differentiate the young- and mature-incumbents from each other in a situation of disruptive innovation. The central proposition is that mature-incumbents and young-incumbents possess fundamentally different managerial characteristics and that these differences will be reflected in the attitude towards disruptive innovation.

Overconfidence

Incumbents tend to focus more on incremental improvements of products whereas a disrupting party’s focus lies more on exploring new options and techniques (Henderson & Clark, 1990). In the light of the new categorization demonstrated in this research, young-incumbents might find themselves still resonating more with Henderson & Clark’s description of disruptors than with that of the incumbents. This could be the case as not too long ago, they themselves were the agile and innovative attacker of the market rather than a part of the party defending themselves against a wave of disruptive innovation. Entrepreneurship studies show that previous positive outcomes in dealing with risky situations will amplify the willingness to change the existing business model and thus increase the leader’s openness towards risk-taking through exploration and innovation (e.g. Curseu & Louwers, 2008) as innovation is considered a managerial risk due to its uncertain results (Latham & Braun, 2008). We can extend this thinking in the light of a previous wave of disruption where young-incumbents have experienced positive outcomes in dealing with a risky situation by successfully disrupting the market.

At the start of the research period I have held orienting discussions and interviews with field-experts that have worked with management from both from the startup world and incumbent companies. Throughout these conversations I found agreement with the aforementioned studies and my statement of expectation that leaders of younger organizations indeed seem to be very confident, tend to be willing to take more risks and are more open to innovation provided they have recently experienced successes with their organization.

(13)

10

These expected high levels of confidence in young-incumbent’s attitude could be explained through a metaphor of David and Goliath that young-incumbents may identify with. Imagine a young-incumbent looking at its own situation a few years ago: during its period as David the disruptor, all odds were against him. Not one, but several mature-incumbent Goliaths resided comfortably in their strong market position. No one would have thought David would stand even a chance. Granted, in most industries today, not all Goliaths have been slain. Bit by bit however, the underdog, David the disrupter, has been chipping away at Goliaths market share. He is able to do this by using his disruptive technology or business model as a slingshot against the shining armor and sword of scale, scope and resources that Goliath relies upon. The reasoning in this research is that if David were to encounter another Goliath, or get into another situation in which he has to take on the same Goliath, he will be more and more confident in his ability to succeed as, against all odds and expectations, he did so previous time. However, who is to say that following the first encounter, all Goliaths have not learned from this experience, and they will not be so easily surprised anymore? They might even have their own equivalent of a slingshot ready?

Thus, it is easy to imagine that young-incumbents reason a way similar to the above described metaphor, resulting in the expected relatively high levels of confidence. The question that arises here is: what happens however when this confidence turns into overconfidence? Galasso and Simcoe show that “overconfident CEOs are more likely to take their firms in a new technological direction” (2011, p. 1469), however, Trevelyan (2008) found that “a business leader’s overconfidence is harmful when making decisions in response to setbacks” (2008, p. 986), and Simon and Houghton (2003) discovered that overconfident managers were more likely to be certain about future success of product introductions that objectively measured were relatively risky and less likely to succeed. Risk-taking might therefore be an indication of overconfidence, whereas a risk-averse attitude is indicating the absence of overconfidence in a manager’s attitude.

Larrick, Burson & Soll (2007) performed two quantitative studies that concluded there is a positive relationship between better-than-average perceptions and overconfidence. As there is a fine line between confidence and overconfidence, it is not easy to measure whether an organization’s management acts in a way that is either confident or overconfident. Therefore, the remainder of this thesis is based on the assumption that an

(14)

11

organization that displays high levels of confidence in addition to a relative low level of more down to earth and modest attitudes, or even high levels of confidence in addition to attitudes of looking down on competitors and having a negative image of their capabilities, can be seen as displaying a ‘better-than-average’ belief and attitude. Following this reasoning, it can be assumed that these organizations therefore have an increased likelihood of forming overconfident attitudes.

Thus, referring back to the David and Goliath metaphor, successes gained from an underdog position create a possible overconfidence in the young-incumbent’s own power and ability to beat the giants again and again. In this research I argue that this recent success with managerial risk taking (their successful disruptor-phase in which they took on the giants) will create an overconfidence in young-incumbents’ ability to make successful moves in future risky situations.

After having looked at risk-taking and innovation as a sign of overconfidence, another subject currently popular in the area of literature on innovation is the concept of open innovation. According to Chesbrough (2004), open innovation is increasingly becoming the norm, requiring organizations to change the way they manage innovation: “external sources become more prominent, while external channels to market also offer greater promise” (2004, p. 23). Because of the paradigm shift from closed innovation towards open innovation, Chesbrough argues that the closed innovation approach is no longer sustainable. However, even though the academic literature supports the need for open innovation, not all organizations are embracing this approach: some still prefer the Do-It-Yourself-method (DYI-method). Therefore I argue that an incumbent’s DYI attitude would possibly be an indication of overconfidence, rather than justifiable confidence. A positive stance towards the use of open innovation on the other hand would indicate that an incumbent is not overconfident of its own abilities in this area, thus indicating an absence of overconfidence.

To summarize, a young-incumbent’s recent successes in entering and disrupting the market has a positive influence on their current attitude towards innovation, however, this innovation might very well come in the form of closed DYI-innovation. Risk-taking can be advantageous in that it can reap fruitful rewards, however, it can also have negative consequences on the organization when responding to a setback in the business-environment. An example of a setback in the light of this research could be a new

(15)

12

disruptor gaining ground in the field, or mature-incumbents ‘hitting back’ against a prior disruption. Apart from a risk-taking and DYI attitude, high levels of confidence combined with a relative low level of level of down to earth and modest attitudes, or even high levels of confidence combined with looking down on competitors and having a negative view of their capabilities thus indicate an increased possibility of overconfidence in managerial attitude.

I argue that young-incumbent may display forms of overconfidence that influence their ability to respond to upcoming disruptive threats in an appropriate way.

Stated as a proposition:

Young-incumbents are more likely to express overconfidence in their attitude towards disruptive innovation.

If the results of the research support this proposition, the theory of social contagion can be brought forward to further strengthen grounds for support by displaying a manner in which mature-incumbents, as opposed to young-incumbents, are less likely to express overconfidence. On the contrary, it could provide for an indication of the opposite: a lack of confidence in one’s own ability to innovate or keep up with the innovations in the business environment which is displayed by the incumbent displaying FoMO, or, Fear of Missing Out. Disruptive change brings about high levels of uncertainty. Choosing to adopt, initiate or invest in a certain innovation encompasses risk-taking. One of the factors influencing this choice can be described as ‘social contagion’. This phenomenon can be seen as the handling of uncertainty through looking at others in order to create a socially acceptable interpretation of an uncertain situation: “social contagion arises from people proximate in social structure using one another to manage the uncertainty of innovation” (Burt, 1987, p. 1288). This is a common occurrence in everyday life, however, as Keynes already suggested in 1936, businesses tend to do the same. When an incumbent only has limited information available about the consequences of its decision on how to handle a disruptive threat or opportunity, social contagion can be a driving factor (Tripsas & Gavetti, 2000). They may feel like the choices of other organizations convey or are based on information more valuable than their own for decision making (Banerjee, 1992).

(16)

13

Social contagion theory can therefore be connected to a situation where businesses mimic competitors’ activities, based on the previously mentioned concept of ‘Fear of Missing Out’ (FoMO): ‘a pervasive apprehension that others might be having rewarding experiences from which one is absent’ (Przybylski, Murayama, DeHaan, & Gladwell, 2013, p. 1841). When we translate this to a business context, we might very well expect to see the example of an incumbent attributing part of their budget to research or invest in a technological development they are not necessarily (planning to become) specialized in, however, as competitors seem to make advancements in that area, they try to keep up in fear of missing out on the latest technology-hyped bandwagon. This response to a certain development in the competitive environment would, as with social contagion, stem from uncertainty and a lack of confidence in one’s own ability to assess and make strategic choices, and the impression that others have information more valuable for decision-making than the incumbent itself has. The original concept of FoMO is often portrayed in a negative light. I argue however that, if FoMO does indeed appear in a business context, it does not necessarily have negative consequences. This is because this form of motivation might just be the nudge an incumbent needs to stay in the game of exploration. FoMO might therefore be a stimulant for an incumbent to overcome path dependency, organizational rigidity and an incumbent’s tendency to favor exploitation over exploration.

(17)

14

Transformational experience

Comparing young-incumbents to mature-incumbents when both are in need of organizational change in a response to shifts in the business environment, young-incumbents may be seen as the agile and flexible party that will easily maneuver upcoming turbulence in the market. Opposite of this one might expect to see the mature-incumbent that, due to its age, size and institutionalized capabilities, faces ‘mature-incumbent inertia’, due to having its core capabilities turned into core rigidities (Leonard-Barton, 1992). This may cause for the incumbent to be considered less flexible and adaptable to sudden changes in the market. However, King & Tucci (2002) describe two kinds of experience a firm can gain: static experience creates a routine that reduces a firm’s dynamic capability, and transformational experience which generates routines that increase dynamic capability. They argue that ‘transformational experience can influence dynamic capabilities either by reducing the buildup of organizational inertia or by creating routines that support organizational change’ (2002, p. 173).

When applying these theories to the setting of this research, the situation in which the incumbent has gained the transformational experience is the previous wave of disruptive innovation, which can also be seen as a wake-up call to the incumbent caused by the disrupting party. In terms of ‘incumbent inertia’, this can be considered a ‘reset of the organization’s inertial clock’ (Amburgey, Kelly, & Barnett, 1993). Considering the current categorization of young- and incumbents in the light of experience, mature-incumbents are found as the sole party that has previous experience with overcoming disruption. The young-incumbents were the disruptors themselves during the previous wave of disruption. Following this reasoning, mature-incumbents would be the sole party to possess this form of transformational experience in terms of dealing with disruption. This experience can increase the incumbent’s sensitivity and preparedness for change. As young-incumbents were standing at the other side of the fence during this previous period of uproar, they would thus lack this experience.

Aside from direct experience in dealing with a situation of a market being disrupted, the experience King & Tucci (2002) discuss can also be found in other, more general forms of experience. Examples are previous experience with product introductions and new market entries resulting in reduced search costs when assessing future opportunities. In addition to that they also learn from their own mistakes and misperceptions and thus

(18)

15

adjust their expectations in future situations. Teece et al. note that “the capacity to reconfigure and transform is itself a learned organizational skill. The more frequently practiced, the easier accomplished” (1997, p. 521). King & Tucci add that “institutionalizing the innovation process creates new business opportunities” (2002, p. 173). One result of the institutionalization of the innovation process is the creation of innovation-related routines and innovation-related key performance indicators (KPIs) that, according to Parmenter (2015) are a set of measures focusing on the most critical aspects for an organization’s success.

Nelson & Winter (1982) touch upon the subject of routines by arguing that with experience, routines are gained that help an organization to make sense of the world. They define organizational routines as “all regular and predictable behavioral patterns within firms” (1982, p. 14). Specifically focusing on large and complex firms, they argue that information on which decisions are made is possessed by various members of the organization. For the quality of the information, they emphasize the importance of linking individual memories by shared experiences in the past. These experiences establish the important communication system that underlies routine performance. These routines are modified and improved with time and organizational experience. In the book, Nelson and Winter further acknowledge Schumpeter’s notion of an opposition between routinization and innovation, but argue that “routinized arrangements for producing innovations and solutions to problems take a variety of forms” (1982, p. 132). They also “propose to assimilate to our concept of routine all of the patterning of organizational activity that the observance of heuristics produces, including the patterning of particular ways of attempting to innovate (…) but emphasize that viewing innovative activity as routine in this sense does not entail treating it’s results as predictable” (1982, p. 133). In short, organizational behavioral patterns accumulated through transformational experience can help firms cope with complex situations such as situations of disruptive innovation. Hence, well-defined routines regarding the path of innovation and the resulting KPIs can be found in a firm that has accumulated transformational experience. Stated as a proposition:

Mature-incumbents are more likely to express transformational experience in their attitude towards disruptive innovation.

(19)

16

Program persistence bias

Nelson and Winter (1982) emphasize that over time, organizations tend to develop path-dependent routines that help them cope with the existing competitive environment. The other side of the coin is however that this can constrain them in adapting to a changing environment. Teece (2007) states that incumbents are often ill-structured for innovation: bureaucracy unquestionably slows decision making as permission has to be sought in several places and committees, sometimes even outside the organization unit the idea stems from. When committees are involved, a ‘program persistence bias’ will cause for innovative ideas to be considered threatening to advocates of other programs and ideas.

Program persistence refers to the funding of programs beyond what can be sustained on the merits, and follows from the presence or influence of

program advocates in the resource allocation process. This proclivity almost automatically has the countervailing effect on reducing funds available to new programs, which are unlikely to be well represented in

the decision making process (Teece, 2007, p. 1327).

Program persistence can be illustrated within the concept of ambidexterity. The basis of the concept of ambidexterity is that the ability of a firm to simultaneously perform both exploration and exploitation rather than favoring one over the other enables a firm to adapt to changes in the market over time. Ambidexterity is a much-discussed way to solve Christenson’s innovator’s dilemma (e.g. O’Reilly & Tushman, 2004). As March puts it: “the basic problem confronting an organization is to engage in sufficient exploitation to ensure its current viability and, at the same time, devote enough energy to exploration to ensure its future viability” (1991, p. 105), and “established organizations will always specialize in exploitation, in becoming more efficient in using what they already know. Such organizations will become dominant in the short run, but obsolescent and fail in the long run” (2003, p. 9).

An obvious example of program persistence bias as a result of imbalance in ambidexterity where exploitation is favored over exploration is when an organization’s budget is distributed in favor of exploration: managers of established product lines act as program advocates, working against the ‘threat’ of new programs or ideas. Those program advocates are backed up by investment decisions that are more comfortably made when future cashflow is relatively certain compared to the uncertain payoffs regarding innovative ideas (Teece, 2007).

(20)

17

Hence, when an organization puts more emphasis on improving current products and services than on exploring new opportunities, this might be a display of program persistence bias. On the other hand, organizations with a balance between the two, or even favoring exploration over exploitation are not expected to display this bias as they will be more flexible and innovation oriented.

Disruptive innovation can often force an incumbent to shed certain assets and capabilities. Program persistence in the form of an anti-innovation bias might therefore cause incumbents to be more likely to develop technologies that can utilize existing complementary assets. Through too narrow a focus on exploitation, firms limit their own chances to spot potentially disruptive innovations (Teece, 2007).

Internal politics can also be seen as a form of an anti-innovation attitude, favoring current programs over unknown new ideas as multiple studies have shown support for the perception that organizational politics impede the implementation of innovative ideas (e.g. Frost & Egri, 1991; Parker, 1995).

The anti-cannibalization bias is another form of program persistence bias. As Steve Jobs’ famous quote goes: “If you don’t cannibalize yourself, someone else will”. According to Houck (2014), successful companies routinely need to cannibalize business in order to survive, however, evidence suggests that only a minority of organizations has the stomach for such significant changes. Teece refers to anti-cannibalization as “a particular manifestation of incentive and structural problems that can thwart innovation in established enterprises” (2007, p. 1335).

To summarize, it would not be surprising to see incumbents limit themselves in terms of innovation and responses to disruptive changes in the market due to program persistence biases such as to anti-innovation as well as anti-cannibalization bias.

Stated as a proposition:

Mature-incumbents are more likely to express program persistence bias in their attitude towards disruptive innovation.

(21)

18

3. Research

Previous chapters have discussed the newly proposed framework of actors in a situation of disruptive innovation and the corresponding propositions that have been derived from literature on incumbent-disruptor dynamics and theories on disruptive innovations and dynamic capabilities. This chapter describes the manner in which research has been conducted in order to test the propositions. Firstly an overview of the research setting and a short history of the industry in which the case study has been performed are presented. Subsequently, the methodology will be described and an account of the data collection and analysis process is given. This will provide the basis for an elaborate discussion of the analysis of the results in the subsequent chapter 4.

Research Setting

Energy sector & disruption

The energy sector was chosen as the industry to illustrate the new categorization of incumbents in the light of disruptive innovations because it has been experiencing disruptive innovations in various areas in the last decade. Next to the development of price comparison platforms and new ways of communication due to the popularization of the internet, examples of disruptions in the energy sector are found in areas such as: smart homes; smart meters and data analysis; the decentralization of power systems; renewables; distributed generation and energy storage; and all resulting new business models fitting these new technologies (Pérez-Arriaga & Knittel, 2016). In addition to that, the waves of disruption have not stopped coming. According to the 2015 PwC global power and utilities disruption index survey, 97 percent of respondents (global electric power executives) indicate they expect a medium or high amount of disruption in their main home market by 2020. The report also states that on a global level, the market disruption index rises by 42 percent between 2015 and 2020 and that Europe remains the most disrupted region in 2020 (PwC, 2015).

(22)

19

The Dutch energy sector

This case study will focus on the Netherlands. This country, like many others, has seen several waves of disruption in the energy sector in the last decadesand finds itself on the brink of encountering several more (Staupe et al., 2016). The Dutch energy market is especially suited as subject for the case study because it went through a process of liberalization of the market as late as 2004. Up until liberalization, the Dutch energy sector could be characterized by a homogeneous oligopolistic market structure (Lise, Linderhof, & Kuik, 2003). After decades of government-protected market environment, the market opened up and created opportunities for new entrants. This clean cut was an important reason for selecting this sector in this specific country as it provides for ideal means of illustration of a situation where the proposed extended framework of categorization would be applicable.

In order to better understand the research-stage, a short history and summary of the research setting follows.

Founding of the big three

Up until the 1980s, energy production and distribution was largely regionally organized and owned by Dutch municipalities. In the early 1980s the Dutch government got more involved and through mergers the amount of suppliers was decreased in order to increase efficiency, after which the municipalities remained involved through being shareholders. Through new mergers in the 1990s the current ‘big three’ energy companies of the Netherlands were created: Nuon (1994), Eneco (1995) and Essent (1999) (Verbong & Geels, 2007).

Liberalization of the market

Following movements in the European markets, the Dutch government started implementing liberalization policies in 2001 with the goal of increasing efficiency of production and distribution through increasing competition in the sector (Lise et al., 2003). Upon completion of full market liberalization in 2004, every customer in the Netherlands was allowed to choose their supplier (International Energy Agency, 2004). After the liberalization, various newcomers entered the market, amongst which were NLE, Qurrent, Energieflex and EnergyZero.

(23)

20

Unbundling of the market

After the liberalization, the following large regulatory change in the Dutch energy sector concerned the unbundling of all energy companies. Unbundling in the energy market is the separation of internal accounts for each of the transmission and distribution activities of distribution networks and commercial companies, up until that point belonging to the same holding. The second European Electricity Directive (2003) required legal unbundling, however, the Dutch government also demanded ownership unbundling of all distribution companies from the parts of the holdings with a commercial interest in the energy sector before 2010, in order to prevent competitive problems (de Nooij & Baarsma, 2009). The resulting reorganizations of the sector led to international acquisitions of two out of the ‘big three’ energy providers in the Netherlands: Nuon (acquired by Swedish state-owned utility Vattenfall AB) and Essent (acquired by the German power and natural gas utility RWE AG) (Newton, 2009). All shares of Eneco group are still owned by 53 Dutch municipalities. Recently, Essent and RWE have undergone another restructuring of the organization. As a result, Essent now belongs to the Innogy-branch of RWE (Essent.nl, 2016).

Research methodology

As the proposed framework did not yet exist and previous work on the incumbent-disruptor dynamics has indicated the need for more research of heterogeneity and intervariance within the group of incumbents, a qualitative approach is chosen for this research. The research conducted is mostly inductive and explorative in nature as the goal is to provide and assess the necessity of a new framework of categorization that highlights differences in managerial factors between young- and mature-incumbents in the light of disruptive innovations (Eisenhardt, 1989).

Following the literature research, several working propositions have been formed that have further guided data collection in the form of an inductive multiple case study within the Dutch energy sector. The multiple case study approach has been chosen in order to locate overarching patterns between the cases as well as disconfirming evidence for higher quality theory-building (Eisenhardt, 1989; Yin, 2009).

(24)

21

Data Collection

The cases have been selected through a purposeful intensity sampling technique (Patton, 1990). The cases have been selected as they are particularly suitable to illustrate the newly proposed framework (Eisenhardt & Graebner, 2007): five out of the total of ten semi-structured in-depth interviews that have been held with strategic and innovation-level management have been performed in young-incumbent companies (case-group 1), and the remaining five in mature-incumbent companies (case-group 2). Within each of the two cases, multiple interviews have been performed in order to create a balanced amount of viewpoints both inter as well as intra-group, which should reduce bias (Saunders, Lewis, & Thornhill, 2012). These interviews have been selected through a process of homogeneous sampling, selecting people of similar positions in order to describe the particular cases in depth on specifically focused issues (Patton, 1990), in this case the issue of responses to disruptive innovations. The ten interviews add up to a total of over 11 hours of recordings. All interviews have been transcribed and subsequently coded and analyzed. For an overview of selected organizations and interviewees per case-group, see appendix A. The interview protocol is located in appendix B. All transcriptions and analysis data are available upon request.

As mentioned in the introduction, this research performs a case-study in the Dutch energy sector which has been liberalized in 2004. As discussed, this cut provides for an ideal situation when categorizing the two parties of young- and mature-incumbents into two separate groups, both facing new waves of disruptive innovation. Therefore, the cases have been carefully selected adhering to the following selection criteria: the mature-incumbents needed to be established in the Dutch energy sector for a minimum of 15 years. As such they had already established themselves before the liberalization of the Dutch energy sector in 2004 and thus they have experienced the previous wave of disruption of the market following the liberalization from the incumbent side. Furthermore, young-incumbents need to be established in the market no longer than twelve years, ensuring that they entered the market after the liberalization and occupied the position of disruptor in the previous disruptive wave. All selected respondents hold a position on strategic or innovation management, or advisory level in the company. In the early stages of this research I have visited several conferences and networking events held on themes linked to disruptive innovation and the energy sector. At these

(25)

22

events I have interviewed and discussed with several experts on the subject in order to create a clear overview of the topic in the real world. This has aided to ascertain the real-world applicability of the concepts conceived through my academic literature research. As described in the introduction, the conclusions drawn from patterns and differences across the interviewees’ answers within and between the two case-groups will answer the research question: How do young- and mature-incumbents differ from each other with regards to managerial factors at play in a situation of disruptive innovation?

In order to increase rigor and validity, data triangulation in the form of extensive desk research of reliable news-sources, annual reports and company websites has been done to verify interviewee’s statements regarding innovation expenses and strategies. Now follows a description of how the analysis was conducted, how research data was interpreted, analyzed and reported by using qualitative analysis tool NVivo 11.

Description of the coding process

For the coding process, a combination of deductive and inductive coding approaches has been used. A priori specification of constructs have been selected in order to create a starting point for the theory building research (Eisenhardt, 1989): from the literature, the working propositions and several expert-interviews that were held, several main starter-codes have been deducted. These main code categories were adjusted and elaborated upon by iterative rounds of test-coding in which was searched for key themes, patterns and relationships that emerged from the transcripts. Through these test-rounds the starter codes were confirmed as being the appropriate superordinate code categories. In addition to the confirmed started codes, more specific subordinate categories of codes emerged. This process finally resulted in the final ‘confirm’ code groups under which interviewees’ expressions confirming one of the three propositions are categorized. In order to reduce researcher biases and increase rigor and validity, throughout the entire research process I purposefully looked for counter-evidence that would suggest the need to refute a proposition. This is illustrated by the coding structure: after the ‘confirm’ code-categories were created using the propositions as guidelines, ‘counter-evidence’ code categories were created. These counter-evidence code categories are aimed at coding all instances that indicate a disconfirmation of the presence of the subject

(26)

23

of the various propositions. The same continuous process of iterative test-rounds have eventually resulted in the final ‘disconfirm’ code groups. This process can be compared to the ‘pattern matching’ procedure for theory testing as described by (Hyde, 2000) in which a score-sheet solution (Campbell, 1975) is used as an initial interpretation of the results.

A final test-round was performed in order to refine the coding framework. In every following part of the process I have kept in mind the continuous and iterative characteristic of the type of research and added or adjusted coding categories when necessary.

The total coding process eventually resulted in twenty-two subordinate categories that aggregated into six superordinate categories. An overview of the total group of superordinate and subordinate categories can be found in appendix C. Definitions of every code category are located in appendix D.

As a result from the continuous and iterative coding process described, interviews Y1-Y5 (young-incumbents) have yielded a total of 128 codes, whilst interviews M1-M5 (mature-incumbents) have yielded a total of 135 codes, as visible in figure 1. This near equal amount of codes extracted from both categories provides for more equal grounds for comparison.

Figure 1 - Total amount of instances coded per category

Mature-Incumbents: 135 x coded (51%) Young-Incumbents: 128 x coded (49%)

(27)

24

4. Case Study

In order to ensure high quality analysis of this multiple case study and to make it easier for the reader to interpret the results, use has been made of several ways of displaying and analyzing the data. As described in the previous section, the amount of coded instances will provide initial grounds for comparison. By itself however, this is not nearly a solid ground for comparison and analysis of the cases as clearly some expressions extracted from the interviews carry more weight than others in confirming or refuting the propositions. Hence, I emphasize that comparison of amount of mentions with regards to the confirmation or disconfirmation of a certain proposition are by no means the main grounds of comparison used. They are merely indications and illustrations of the possible large differences between the two case-groups. The actual contentual differences will subsequently be analyzed, discussed and illustrated by quotes. Quotes will be indicated as coming from either Yx (a young-incumbent interviewee) or Mx (a mature-incumbent interviewee).

At a glance

In order to identify early patterns from the great volume of data produced through the coded interviews, a cluster analysis has been performed at the outset of the analysis process. An early first impression of the results are visualized in a 3D graph in figure 2. The cluster analysis has been used as an exploratory tool in order to explore the similarity in coded expressions between all interviews. An table containing the results of the cluster analysis can be found in appendix E.

Figure 2 - 3D visualization of explorative cluster analysis.

All sources are clustered by coding similarity

(28)

25 15 46 34 0 29 11 63 29 3 7 ₂ 26 0 10 20 30 40 50 60 70 1 : OVC

(Confirm) (Disconfirm)2 : OVC (Confirm)3 : T.E. (Disconfirm)4 : T.E. (Confirm)5 : P.P.B. (Disconfirm)6 : P.P.B.

Comparison of main code categories

Mature-Incumbents Young-Incumbents

As visible in the diagram, two clear clusters emerge. This indicates a clear separation in correlation of coded expressions between the two case-groups in which the interviews were held. The differences between the groups of young-incumbent companies (Y1-Y5) and mature-incumbent companies (M1-M5) indicate careful signs of initial support for the central proposition that mature-incumbents and young-incumbents possess fundamentally different managerial characteristics and that these differences will be reflected in the attitude and responses to disruptive innovations.

In order to find and understand the underlying reasons for this separation, a discussion of the general outcomes of the analyzed interviews will now follow. Using the three propositions as a central thread, the two case-groups of young-incumbents and mature-incumbents will subsequently be discussed, compared and analyzed.

Matrix coding queries have been performed in order to visualize the amount of codes each code-category has accumulated throughout the coding process. Figure 3 is an example of the visualization of the total amount of codes in the superordinate code categories, compared between the two case-groups: mature-incumbents versus young-incumbents.

As becomes clear in this general overview of amount of codes assigned per superordinate coding category, initial comparisons show relatively higher counts for young-incumbents

Figure 3 -Comparison of main code categories

Code Definition

OVC Overconfidence

T.E. Transformational Experience P.P.B. Program Persistence Bias

(29)

26

in the category of expressions confirming the presence of overconfidence (OVC) whereas the mature-incumbents have higher counts when discussion subjects that confirm the presence of expressions confirming the presence of transformational experience (T.E.) and program persistence bias (P.P.B.). Hence, this can be seen as a sign of initial support for all three propositions.

However, in order to gain a more solid understanding of the meaning of and reasons behind these differences and initial indication of support for the propositions, a deeper analysis is warranted. This will lead to the final decision of confirming or refuting the propositions. In the following subchapters the results are analyzed and discussed per proposition, followed by overall conclusions, implications for the academic and business world, and suggestions for further research.

(30)

27 Mature-Incumbents 19% (15x) Young-Incumbents 81% (63x)

Overconfidence (confirm)

Figure 4 - Comparison of total amount of instances expressions of young- vs mature- incumbents were coded in superordinate categories confirming and disconfirming the presence of overconfidence (proposition 1)

Proposition 1 – Overconfidence

P1: Young-incumbents are more likely to express overconfidence in their attitude towards disruptive innovation

As can be seen in figure 4 displayed above, an overwhelming amount of expressions by interviewees of the young-incumbent case-group have been coded that provide support for the confirmation of the overconfidence proposition. Mature-incumbents expressed more disconfirming statements regarding the presence of overconfident attitudes, indicating further support for proposition 1. In order to gain a more profound understanding of the reasons behind this suggested outcome, an analysis of both the subordinate code categories of the confirming and disconfirming superordinate code categories of overconfidence now follows.

When interpreting figures 5 and 6 (displayed below), it becomes clear that although the contribution of the subordinate categories of expressions indicating a risk-averse attitude and expressions indicating a risk-taking attitude is relatively small, the results do indicate that young incumbents were the only party expressing a risk-taking attitude whilst mature-incumbents were the sole party expressing a risk-averse attitude. For example, Yx notes: “we are willing to make difficult decisions (…) we have the courage to do it and maybe we will fail but at least we will learn a lot from it, and we will make it work somehow in the end”.

Mature-Incumbents 61% (46x) Young-Incumbents 39% (29x)

Overconfidence (disconfirm)

(31)

28

Figure 5 - Visualization displaying the amount of subordinate codes found that confirm the presence of overconfidence (proposition 1)

Figure 6 - Visualization displaying the amount of subordinate codes found that disconfirm the presence of overconfidence (proposition 1) 8 0 7 0 28 5 28 2 0 5 10 15 20 25 30

Confidence DYI Negative View of

Competitor RiskTaking

Overconfidence - Confirm

Mature-Incumbents Young-Incumbents

Whereas Mx states: “so now we have more of an investor’s mentality, more looking for certainty (…) that has to do with the culture as well. For example the [smart-thermostat] was a relatively safe investment as at that point it concerned an iterative innovation”.

22 14 6 4 0 19 2 6 0 0 0 5 10 15 20 25

Down to Earth Open to External

input Positive View ofCompetitors Risk-Averse FoMO

Overconfidence - Disconfirm

(32)

29

As risk-taking is an indication of overconfidence (Simon & Houghton, 2003) as has been discussed in the theoretic section of this thesis, the examples above provide for the first careful steps towards confirmation of the overconfidence proposition. In the example, Yx does not only specifically mention that they have the courage to take risks, but also expresses strong confidence that they will succeed in the end whereas Mx uses the term ‘investor’s mentality to indicate his organization’s more careful attitude towards investments with relatively uncertain outcomes.

Throughout the analysis of the interviews, it became apparent that young-incumbents are the ones taking on a ‘do it yourself’ (DYI) attitude which has not once come forward in interviews with mature-incumbents. Yx puts it very clearly: “we haven’t done that yet [data analysis services], we are just going to learn it all ourselves”.

Mature-incumbents expressed more often than young-incumbents their openness to external input of information and skills. For example, Mx: “our participation at the blockchain hackathon as partly driven by the hope to be surprised by new ideas and input, we purposefully did not give a lot of direction to the teams”

As discussed in the theory section of this thesis, a DYI attitude could indicate overconfidence as the closed innovation approach is no longer sustainable (Chesbrough, 2004). Therefore, the young-incumbents’ DYI attitude could indicate an overconfident attitude, rather than justifiable confidence. Conversely, openness to external input and information can possibly indicate the absence of overconfidence. A specific example here comes in the form of open innovation which is mentioned by mature-incumbents as an asset several times.

A clear example is provided by Mx: “I think now we have arrived to the next level which is open innovation. Making people understand that it is also important to include your surroundings: your partner landscape and customers, have it open and transparent. Creating ideas is not the problem, the power to execute those this is the problem. Excluding your ecosystem of customers and partners is stupid. This new way of doing innovation: people understand it now. It has not yet been fully implemented but that is a matter of time”. Young-incumbents of course do employ partnerships wherever necessary in order to increase their scope or because of technological constraints, however, the impression given in the interviews is that it is avoided wherever possible. Yx gives an example: “they

(33)

30

[technological partner-company] also do part of the software. That is something we actually prefer doing ourselves but it is so much mixed with the software that that is complicated. We will however in the future do things where hardware is included and then we will do it ourselves”

The subordinate categories that carried the most weight leading to the eventual confirmation of the overconfidence proposition are the categories of ‘confidence’ and ‘down to earth’. The confidence category includes expressions indicating general pride and confidence of the interviewee in his or her organization’s capabilities and those being better or stronger than those of competitors. On the other side of the fence are expressions that indicate a more down to earth attitude in which expressions of self-reflection and the ability to see and admit one’s own organization’s weaknesses and faults are counted.

Mx illustrates a down to earth attitude: “most corporates are always depending on partners: technology partners, agencies, etc. Few companies have those capabilities in-house (…) for an organization like this it is a big change: IT, marketing, product development, project management, etcetera. They all need all kinds of things but do not have the capabilities for it. They want to have them [technological capabilities] but can’t seem to tie them to us. This is also because of our locations: not many people want to be working in those places”.

Of course, an organization can be confident whilst also realizing and admitting its own weaknesses and faults. This is not necessarily a sign of overconfidence. This situation was also encountered during the analysis of the interviews: both case-groups expressed statements for both the confirming and disconfirming categories of the confidence proposition. The assumption here is however that if one case-group displays an overwhelming amount of either confidence or rather an overwhelming amount of self-reflection, this might indicate this group’s tendency to either overestimate their own strengths and abilities and underestimate their competition’s, or vice versa. This reasoning is in line with the ‘better-than-average’ effect as demonstrated by Larrick et al. (2007). As discussed in the theory section, a display of high levels of confidence in addition to a relative low level of more down to earth, modest attitudes or even attitudes of looking down on competitors and having a negative image of their capabilities can be seen as displaying a ‘better-than-average’ belief and attitude.

(34)

31

As visible in the results, mature-incumbent interviewees convey considerably more indications of being down to earth than being confident. Young-Incumbents on the other hand also articulated a fair amount of instances that suggested down to earth views. However, an even larger amount of expressions of confidence has been found in the latter group. This leads to the conclusion that, even though both case-groups conveyed expressions pointing at both sides of the spectrum, mature-incumbents can be placed more on the down to earth side whereas for young-incumbents confidence seems to have the overhand. This confidence in combination with the relatively negative view of competitors they express (discussed in the next section) does lead to the statement that young-incumbents seem to have a tendency of having an overconfident attitude.

Although the interviewees from both case-groups accounted for an equal amount of mentions expressing positive views of their competitors, young-incumbents accounted for an overwhelming part of the mentions that expressed to see competition in a negative light.

Yx presents an example: “personally, I find that a very strange strategy [acquiring technological startups]. They kind of admit that they are not going to make it on their own and therefore they are taking over some small parties. On the long term that does not seem right to me. That is why I do not have any trouble with that, they should definitely do whatever they want and then we will do our thing. (…) The moment they have actually implemented a new business model for example, another startup will already be there with something new. They will always be behind. Innovation is not in their genes”

As mentioned before, the young-incumbents’ tendency to see competition in a negative light combined with a possible tendency towards an overconfident attitude indicates more support regarding the overconfidence proposition.

Another interesting point of discussion regarding the overconfidence proposition is the lack of support found for the concept of FoMO (Fear of Missing Out). After various rounds of coding had been performed, an unexpected result came forward. FoMO, the subordinate code category for which I expected to encounter evidence during the interviews that would indicate an absence of overconfidence in an incumbent manager’s attitude: the concept remained completely unsupported.

Incumbent-disruptor dynamics & disruptive innovation in a digital age : disruptor disrupted : a case study of the Dutch energy sector