Research Methodology

26

27

complex realities that are problematic to capture through reductionist treatment of few variables (Gephart, 2004). As Popper (cf.1969:115) expressed, positivist research effectively makes “guesses” through hypotheses which are then checked. Interpretive research however, attempts to make discoveries during research. Among the tools available to interpretive researchers, case studies are particularly appropriate when phenomena are difficult to isolate from their context, thus case studies rely on rich data of events in their real-life context (Yin, 2003). Case study research follows induction logic, building and expanding theory rather than merely testing it. Due to its predilection for theory building, case study research is often regarded as a precursor for deductive, quantitative studies, and thus often relegated to sharpening hypotheses which can then be tested through quantitative modeling (Flybjerg, 2006). Indeed, inductive theory building and consequent deductive testing can be regarded as two complementary elements of one research cycle. In addition to being used for exploratory purposes, however, the case study is a method in its own right and can be used to create theoretical constructs and mid-range theory (Eisenhardt and Graeben, 2007). Case studies, for instance are equally suited to perform an explanatory function, in research setting which pursue “why” or “how” questions (Yin, 2003).⁶

When justifying a choice of scientific method, it is mandatory for the researcher to ensure that it is best suited to the question posed. In this instance, the choice of method is strongly predicated both by phenomenological and theoretical ambiguity. The central research question asks how technology-intensive firms can match external knowledge distant from their current technology/product/market context with their existing routines and processes. From a phenomenological point of view, this question is warranted due to the increasing risk of firm failure associated with technology convergence and divergence dynamics.

Incumbent firms are under growing risk of creative destruction from new entrants from outside the industry. In addition, they are under pressure to find new areas of growth as existing markets stagnate.

Therefore, technology-intensive firms need to constantly monitor not only related, but also, distant technology developments. While managers are able to utilize a number of different mechanisms to access such distant knowledge, the critical task remains how such distant and more radical knowledge can be reconciled with the existing routines (Pavitt, 1998; Rosenkopf and Nerkar, 2001) in order for timely adaption or to create new areas for growth. From a phenomenological point of view, therefore, dependence on distant knowledge is continually growing, with little prescription how such distant knowledge can effectively be matched with existing routines and capabilities.

6 It is important to note that there is some degree of confusion when it comes to the distinction between interpretive methods and qualitative research. Qualitative research is sometimes used synonymously with interpretive research.

Gebhart (2004) for instance therefore equates qualitative research with interpretive research characterized by strong social constructionist tendencies. While this study uses qualitative data in the form of interviews, it doesn’t aspire to be explicitly social constructionist. In general, the purpose of this methodology chapter is to convey the “rigor, creativity, and open-mindedness of the research processes while sidestepping confusion and philosophical pitfalls”

(Eisenhardt and Graeben, 2007:30).

28

From a theoretical perspective, research has addressed various issues pertaining to the disaggregation of R&D. Historical studies (Chandler, 1977; Mowery, 1983; Langlois, 2003; Teece, 2003) have charted the changing dynamics of industrial organization. As highlighted in the introduction, a large body of research has in recent years dealt with mechanisms for accessing new knowledge, for instance, in the form of mobile inventors (Arrow, 1962; Almeida et al., 2003; Singh and Agrawal, 2008). As a further mechanism for accessing external knowledge, research on strategic alliances has featured prominently (Contractor and Lorange, 1988; Dyer and Singh, 1998; Hagedoorn et al., 2002; Rosenkopf and Almeida, 2003; Grant and Baden-Fuller, 2004). More research on the sources of innovation deals with lead users, as pioneered by Eric von Hippel (1988; 2006). Furthermore, the presence in geographically concentrated areas of economic activity has been widely discussed (e.g. Marshall, 1920; Porter, 1998 Saxenian, 1994). Recent innovation literature further addresses the new sources of innovation as a result of the growing efficiency in the market for technology, such as start-ups, venture capital firms, internal venturing and so-called innovation intermediaries (Arora, 2002; Chesbrough, 2002; 2006; Harryson, 2007, Tidd and Bessant, 2009). In terms of providing insights to the central research question of this thesis, the above theory has largely ignored the problem. Moreover, the majority of research has relied on large datasets such as patent data or industry level data in empirical analyses. Even some studies which have addressed the present problem (e.g. Rosenkopf and Almeida, 2003) do not provide more detailed insights into the micro-foundations of how firms are able to match distant knowledge with existing capabilities (Knudsen, 2009;

Teece, 2009).

As also outlined in the introductory chapter, another major research stream addressing the question of how organizations learn is represented by the organizational learning literature. This literature stream, however, is highly fragmented. As Vera and Crossan (2003) highlight, all facets of this research field display considerable incongruence in research tradition, focus, philosophy, method and unit of analysis. As they also point out, organizational learning requires the analysis of multiple levels of analysis, thus requiring appropriate research methods, which, so far, have not been used sufficiently (Foss et al., 2007;

Doz et al., 2008).

A key contribution of the organizational learning literature, though, concerns the question of how novelty settles into a dominant design and results in increasingly rigid routines, path-dependence, and inertia (e.g.

Prahalad and Bettis, 1986; Levitt and March, 1988; Leonard-Barton, 1992). This aspect of organizational learning stresses that, as organizations mature, the focus of innovation changes from product to process innovation and incremental product improvement (Abernathy and Utterback, 1978). As a dominant design emerges for the “technology/product/market” combination (Gilsing and Nooteboom, 2006:3), managers rely increasingly on heuristics, or a “dominant logic” (Bettis and Prahalad, 1995). Learning, i.e.

the combination of novel knowledge with existing routines leading to a change in routines, is seriously

29

hampered under these circumstances – an organization’s routines have become “core rigidities” (Leonard-Barton, 1992). Moreover, the cumulative nature of incremental learning results in path-dependence which prevents organizations from searching for new knowledge from outside of their existing context (Cohen and Levinthal, 1990; Rosenkopf and Nerkar, 2001). This “local search bias” also extends to the mechanisms for accessing distant knowledge discussed earlier (Rosenkopf and Almeida, 2003). In this respect, the notion of path-dependence and organizational inertia presents a rival theory to the empirically observed “opening up” of innovation processes. The effects of organizational inertia militate against accessing and utilizing external knowledge which is distant from the current technology/product/market context. A central theoretical conundrum therefore, consists of the questions whether, and how this inertia can be overcome. Finding answers to this question could provide an explanation of the limits to choosing markets over hierarchies for organizing innovation. It would also shed more insights into the limits of certain mechanisms to instill learning and change.

The choice of the case study method, thus is predicated both on phenomenological as well as theoretical challenges. Technology-intensive organizations are increasingly forced to find and apply knowledge which lies outside of the scope of their current operations. Moreover, many managerial publications currently advocate the use of external sources of technological input for R&D. Yet, the challenges associated with matching such distant knowledge with the existing system and routines are largely ignored. Managers have been attempting to source outside sources of innovation and impulses with little concern for their organizational idiosyncrasies, creating R&D outposts or web-based innovation platforms. Often, such strategies have little impact. At the same time, organizations time and time again fail to react in a timely manner to discontinuous changes, often resulting in creative destruction. There is, thus, imminent need for inductive research to provide in-depth answers whether and most important how organizations can manage such matching. There is compelling evidence that organizations are becoming increasingly dependent on knowledge distant from their existing technological and organizational trajectory. Attempts by firms to integrate such distant knowledge are less informed by theory, but seem to be based on trial and error. The costs both for companies but also for the economy at large could be substantial, given the high risk of firm failure as well as the wrong-guided investment decisions undertaken my managers. As widely discussed in evolutionary perspectives on strategy, most firm failure has at its root the inability to react to changes exogenous of the immediate organizational and technological context consisting of new knowledge. Detailed accounts of such episodes of “creative destruction” are legion. Yet, theory addressing the fundamental challenge of how organizations can not only identify but match new pieces of knowledge with their existing capabilities is scant. Practitioners, therefore, have at best heuristics and trial-and error learning to guide their decision making.

30