MSc Business Studies - Strategy Track Master Thesis – 15.08.2014
Structural differentiation and innovation: the mediating role of an organizational transactive memory system in achieving speed and success of innovation.
Student: Emil Karlsson /Student № 10622934 Supervisor: Pepijn van Neerijnen
TABLE OF CONTENTS
1. INTRODUCTION………...……....……....………4
2. LITERATURE REVIEW………9
2.1 Innovation, Organizational Knowledge and Structural Differentiation………9
2.2 The Nature of Knowledge and Knowledge Barriers …..……….10
2.2.1 Cognitive and Cultural Barriers ….………..11
2.2.2 Physical Barriers ..………12
2.2.3 Coordinating Mechanisms ...………14
2.3 Structural Differentiation and the Speed and Success of Innovation………..16
2.4 Coordination and Integration of Dispersed Organizational Knowledge ……….18
2.5 An Organizational Transactive Memory System ………...21
2.6 Transactive Memory System and Structural Differentiation ………..…23
2.7 Transactive Memory System and the Speed and Success of Innovation ………24
2.8 The Mediating Role of a Transactive Memory System ………..26
3. METHODOLOGY ………...29
3.1 Research Setting ……….….29
3.2 Measurement and Validation of Constructs ………....30
3.3 Computing Variables ………...………...33
3.4 Aggregation ……….33
4. ANALYSIS & RESULTS ………....33
5. DISCUSSION ………...….…………...37
5.1 Theoretical Contributions ………..……….38
5.2 Practical Implications ………..41
5.3 Limitations and Avenues for Future Research ………41
6. CONCLUSION ………43
7. REFERENCES………...………...………49
3 Abstract
In this study, we investigated how structurally differentiated organizations handle problems associated with dispersed organizational knowledge. The objective was to determine whether the increased efficiency of structural differentiation, and coordination and integration of dispersed knowledge assets through an organizational transactive memory system can improve the speed and success of innovation. The method included survey responses that were collected out among CEOs from 408 Dutch firms in 2012, out of which 220 responses were used for this empirical investigation. The results confirmed the hypothesized relationship between structural
differentiation, speed and success of innovation and the mediating role of a transactive memory system. Seemingly, the working of informal social relationships has an important role to play in knowledge intensive industries and dynamic environments in which innovation is imperative. These findings contribute to existing literature on organizational design, knowledge management in general and are complementary to previous undertakings on firm coordination and integration mechanism. Most importantly this undertaking has practical value for managers in that these findings stress the importance of an integrated approach in managing innovation. Conclusively, an organizational transactive memory system was found to mediate the relationship between structural differentiation and the speed and success of innovation. More research is needed to replicate and validate these results through case studies as well as longitudinal investigations.
4 1. INTRODUCTION
Innovation is undoubtedly the mainstay of organizations that seek to safeguard their competitive advantages. The attempts by managers, and the pursuit of scholars, have been to identify
components of particular importance to the process of innovation. Lately, organizational knowledge has become a principal focal point of these investigations. The speed at which new products and services are rendered, time to market and the rate of success these innovations ultimately relish is dependent on the access to internal knowledge and expertise (Barney, 1991; Penrose, 1959). Novelty emerges from altering the resource base as a response to opportunities and threats: a re-combinative ability of internally available knowledge and expertise that
underlies its capability of sustaining the competitive advantage of an organization (Teece, 2007). However, organizational knowledge is naturally dispersed (Tsoukas, 1996) and “sticky” in nature (Szulanski, 1996). These properties make it difficult to locate, access and integrate internally available expertise (Carlile, 2004). In addition, the geographic spread of operations, and more decentralized and specialized organizational structures have created additional intra-organizational barriers to knowledge flows. Although current accounts have confirmed the importance of knowledge flows to innovation (Du Plessis, 2007: Di Benedetto, 1999) our understanding of the mechanisms that can alleviate their negative effects are limited. Moreover, research on organizational design has found that efficiency in the innovative process can be achieved through specialization by structurally differentiating organizational subunits in general (Lawrence and Lorsch, 1969), and explorative and exploitative innovation in particular (Raisch et al., 2009). However, structural isolation is contradictory in its contribution to understanding the innovative process as it institutionalizes existing barriers to knowledge flows through limited interaction between work groups of the multiunit organization (Jansen et al., 2009). Intuitively,
5 organizational design needs to be coupled with additional mechanisms that counteract the
rigidity of knowledge flows, facilitating the efficient working of a structurally differentiated organization. It is time to integrate related concepts from adjacent domains in order to increase our understanding of the contextual and interactive effect of structural differentiation effects on innovation performance. Improved understanding of the entire process through integration can help organizations to enhance the effectiveness and success of new product development and maintain the strategic objective of organizations: sustaining a competitive advantage through continuous innovation.
Recent progression in the field of organizational design requires that specific
coordinating and integrating mechanism are identified in order to improve knowledge flows. The structurally differentiated organization first set out to improve efficiency through specialization, autonomy and departmentalization (Lawrence and Lorsch, 1967). However, structural isolation limits the social interaction between subunits which reinforce the rigid nature of organizational knowledge. That is, the path dependent nature of organizational learning reinforces the narrow understanding between different knowledge vectors (March, 1972). Dougherty (1992) calls this phenomenon an emergence of different thought worlds. The language used to describe and share ideas differ which complicates the communication between members of different subunits (Carlile, 2004). The result is limited knowledge flows through (1) reduced awareness of and access to knowledge input between units, and, (2) a more intricate integration of knowledge output from structurally differentiated units. First, lacking awareness leads to suboptimal use of available knowledge resources, which limits the number of possible resource combinations (Teece, 2007). Accordingly, Kogut and Zander (1996) argue that a mechanism to facilitate coordination of organizational knowledge is essential in order to improve knowledge flows.
6 Second, new knowledge that may materialize within structurally separated units needs to be integrated in a way that promotes value creation (Raisch et al., 2009). That is, a novel idea may be conceived by an individual, but it is the process and collective effort that turns a novelty into an innovation (Van de Ven, 1986). In order to institutionalize and profit from a new idea an organization is required to have an integration mechanism present to facilitate this transition. Although certain formal and informal integration mechanisms have been identified (Jansen, Tempelaar, van den Bosch and Volberda, 2009), the current literature does not provide a clear explanation how these mechanisms are able to bridge the cognitive, cultural, and physical
divides between structurally differentiated units. Thus, there is a need for a greater understanding of why organizational members would be willing to aid colleagues from a different sub-unit in an organization with limited formal social interactions.Research on structural differentiation demonstrates that the objective of improved efficiency is achieved, but the question remains how organizational members are able to share knowledge and information despite their limited
common knowledge background, that is, coordination and integration of dispersed knowledge assets in a structurally differentiated organization.
When turning to the knowledge management literature one mechanism is identified that possesses characteristics that may alleviate the impediment posed by structural differentiation on knowledge flows, namely, an organizational transactive memory system. A transactive memory system is an informal mechanism that aids an organization in the “acquisition, storage,
distribution and retrieval of organizational knowledge and information” (Adams & Lamont, 2003, p. 144.). Through informal lateral relationships outside the formal structure that otherwise limits knowledge flows, this mechanism transcend spatially dispersed units and increase the awareness of available expertise and allows for the efficient coordination and integration of
7 dispersed organizational knowledge. By hypothesizing that the presence of a transactive memory system mediates the relationship between structural differentiation and the speed and success of innovation this thesis contributes to the literature on innovation speed and success in two important ways. First, a transactive memory system is conceptualized as firm level coordinating mechanisms (Kogut and Zander, 1996) in that it ensures increased awareness of and access to dispersed knowledge assets (Tsoukas, 1996). An organizational transactive memory system is an informal social mechanism that reduces the “stickiness” of knowledge (Peltokorpi, 2014;
Szulanski, 1996). It replaces the formal social interactions with informal lateral relationships which increase the connectedness between sub-units, facilitating both exploratory and exploitative innovations (Jansen et al., 2006). Intuitively, coordination and integration of dispersed knowledge assets should be facilitated by informal social interactions, replacing the lacking formal social interactions that structural differentiation entail (Hansen, 1999). This assertion is supported by the findings that connectedness through increased inter-unit social interaction improve internal knowledge flows (Tsai, 2002). The result is increased knowledge sharing which facilitate organizational learning (Carlile, 2004), development or organizational capabilities (Grant, 1996) and, knowledge creation and communication through a common language, codes and organizing principles (Kogut and Zander, 1992; Jansen et al., 2009).
Second, a transactive memory system is proposed to function as a decentralized firm level integration mechanism (Lawrence & Lorsch, 1967). This adds to the discussion on integration mechanisms in which previous research that have identified centralized integration mechanisms (Tushman and O’Reilly, 1997; Jansen et al., 2009). The social connectedness on which transactive memory systems are built makes it a flexible, firm level mechanism, by including all organizational members, vertically and horizontally. This confirms the assertion
8 that knowledge is socially constructed (Huber, 1991), and, in a constant state of flux. Social interaction creates a network of relationships that tie dispersed units together through informal linkages (Hansen, 1999). The result is lowered barriers to knowledge flows (Szulanski, 1996) which facilitate the integration of dispersed organizational knowledge and thus the collective achievement of the innovative process (Van de Ven, 1986).
Accordingly, the following section includes the literature review and hypothesis pertaining to the proposed research model. The literature review is followed by the empirical findings based on the data from 220 Dutch firms. Subsequently, we will elaborate on the
findings, discuss its implications for managerial practices as well as avenues for further empirical investigation.
9 2. LITERATURE REVIEW
2.1 Innovation, Organizational Knowledge and Structural Differentiation
Innovation is the source of value creation in an organization and the fundament of the capitalist economy of today (Drucker, 1985). The consensus that the competitive advantage of
organizations depend on the ability to change is seldom challenged. Thus, innovation is a strategic objective, the ability to change in response to competition and changing market
demands, that is, the evolutionary theory of the firm (Nelson and Winter, 1982). To facilitate the effective rendering of new products and services, internal knowledge assets have been proposed to be the condition sine qua non of the innovative process (Penrose, 1959; Barney, 1991; Teece, 2007). To sustain the competitive advantage through innovation we must understand the nature of knowledge, barriers to knowledge flows, and the capabilities that constitute the ability to recombine existing expertise allowing new competences to emerge (Kogut and Zander, 1995).
During the different stages of the innovative process, the importance of access to and integration of internal knowledge assets cannot be understated. Essentially innovation is
concerned with leveraging existing knowledge assets and competences, and in creating new ones (Bartlett and Ghoshal, 1989). Knowledge resides within individuals and the social relationship between them, their interaction is the fundament of the organization as a social community (Kogut and Zander, 1992). In order to leverage existing competences and creating new ones, the firm develops capabilities that are a result of a combination between individual and social knowledge. These capabilities constitute the ability to alter its resource base as a response to threats and opportunities in the marketplace (Teece et al., 1997). Expertise being the main resource of an organization, the capability to alter its resource base is dependent on both the access to and integration of knowledge (Du Plessis, 2007). However, there are barriers to
10 knowledge flows that limit the access to and integration of knowledge, having negative
consequences on the speed and success of new product renderings.
There are three main barriers that complicate the successful sharing, transfer and integration of organizational knowledge. These include, (1) the natural dispersedness of organizational knowledge (Tsoukas, 1996) and its uneven distribution (Tsai, 2002), (2) social aspects such as cognitive limitations and socio-cultural differences between units that emerge in a multiunit organization (3) the structural isolation that is the result of geographical spread of operations and the departmentalization and specialization within the organization that reinforce and institutionalize the “sticky” nature of organizational knowledge (Szulanski, 1996).
These barriers to knowledge flows are captured by the concept of structural
differentiation. Structural differentiation is the division of tasks into specialized business units (Lawrence and Lorsch, 1969). In the context of innovation structural differentiation facilitates specialization in incremental and radical innovative processes, what is commonly referred to as exploitation and exploration (Raisch, 2008). By structurally separating these units the objectives of incremental and radical innovation can be met through ad hoc architectures, competencies and underlying strategies (Jansen et al., 2009). Although structural differentiation facilitates radical and incremental innovation through lowered interference between them (March, 1991), the decentralization, local autonomy and structural isolation also promote the institutionalization of existing barriers to knowledge flows. The circumstances that produce these outcomes become more apparent when the nature of knowledge and barriers to knowledge flows are accentuated. 2.2 The Nature of Knowledge and Knowledge Barriers
The nature of knowledge as an organizational resource possesses characteristics that act as barriers resulting in negative consequences on knowledge flows. Knowledge is partially
self-11 contained within individual members of the firm and partially as social practices and routines that these members perform (Kogut and Zander, 1996). These social practices have three dimensions: social expectations related to a specific role, dispositions and interactive situations (Tsoukas, 1996). Thus, it is constantly being re-created in firm routines (Nelson and Winter, 1982). As such, knowledge emerges path dependently when knowledge is applied to local circumstances (March, 1972). Moreover, what complicates the matter further is that knowledge exists in different forms, being tacit and explicit (Nonaka, 1996). A firm being a knowledge system (Grant, 1996) where organizational knowledge is inherently dispersed implies that it cannot be controlled in the sense of a centralized “control room”. Partial coordination is feasible (Jansen et al., 2009), but due to the limited cognitive ability of the human brain one person alone cannot comprehend all available knowledge within an organization (Tsoukas, 1996). Still,
managers need to facilitate creativity and the novelty that emerge from knowledge assets through coordination and integration. To accomplish successful coordination and integration of
knowledge the different barriers to knowledge flows need to be addressed. 2.2.1 Cognitive and Cultural Barriers
Barriers to knowledge flows have a social aspect in addition to the nature of knowledge, including political, cognitive and cultural differences between individuals and groups. The difficulty of transferring knowledge can be related to the fact that knowledge is “sticky”. Stickiness can be traced back to knowledge as a social construct which creates both cognitive and cultural barriers to knowledge flows. The work by Szulanski (1996) confirms that the absorptive capacity of the recipient, causal ambiguity and the relationship between the sender and receiver are the most important origins of stickiness. However, these boundaries are not stable but are in a constant state of flux as it moves across the continuum that is the innovative
12 process. Essentially there are two properties of these boundaries, difference and dependence (Carlile, 2004). First, the background of groups and individuals differs in terms of cumulative experience and expertise mainly due to the path dependent nature of knowledge (March, 1972). Second, dependence is the relationship between groups and individuals, that is, their power relations (Emerson, 1962). These boundaries do not necessarily have to be problematic, ceteris paribus, but in the presence of innovation which require two parties to interact makes their different “thought worlds” more apparent (Dougherty, 1992).
In his paper Carlile (2004) describes three types of boundaries, that is, syntactic, semantic and pragmatic boundaries and the processes they involve. The syntactical approach is concerned with transferring knowledge using a common language, where a syntactic boundary arises as a result of the new syntax needed to explain the novelty. The semantic boundary is concerned with the translation involved in the transaction, and the problem of interpreting the transferred
knowledge without the presence of a common language. The semantic boundary is enhanced by the fact that both tacit and explicit knowledge exist (Nonaka, 1994). Last, the pragmatic
boundary is where the knowledge is transformed. That includes the compromises of two parties meeting at the boundary, to adapt their current knowledge in order to understand and integrate the new knowledge. The pragmatic boundary puts the difficult nature of knowledge center stage, being a product of specific local practices that emerge path dependently. As such, knowledge is paradoxical in that it represents both a source and a barrier to innovation (Carlile, 2002). The social nature of organizational knowledge means that it is naturally dispersed, and that the cognitive and cultural aspects of social interactions constitute a barrier that limit the flow of knowledge across structural boundaries.
13 In addition to cognitive and socio-cultural barriers the specialization, decentralization and the geographically dispersed nature of the modern organization is also a cause of rigidity of knowledge flows. First, expertise is vertically dispersed. The conception of an idea may be conceived by an individual and elaborated on by a project team concerned with research and development. As the innovation process unfolds the innovation (new knowledge) need be transferred (e.g. to a manufacturing site, sales and marketing departments) and institutionalized (Van de Ven, 1986). Moreover, the problems of knowledge flows related to the horizontal spread of operations is well known and has been studied extensively (Kogut and Zander, 1995).
Horizontal knowledge transfer is, on the contrary, the transfer of knowledge between one
manufacturing site to another. These physical boundaries are impediments to knowledge flows of which its rigidity limit knowledge transfer and integration, vertically and horizontally, within an organization.
In conclusion, the difficulty of knowledge transfer is derived from one or a combination of several factors. The difficulties encompass the context of the transfer, the characteristics of the knowledge, and the cognitive and cultural abilities of and the relationship between the sender and receiver of the knowledge being transferred (Szulanski, 1996). A structurally differentiated organization lowers the degree of social interaction and prevents the formation of formal lateral relationships that would allow for an organic flow of knowledge and expertise. By doing so the cognitive, socio-cultural, political and communicative differences are reinforced and
institutionalized resulting in negative consequences on knowledge flows. Accordingly, an organization being a dispersed knowledge system establishes a need to limit the asymmetrical distribution of knowledge through coordination (Kogut and Zander, 1996). The need for coordination is enhanced further in the multi-unit organization (Tsai, 2002), which depends on
14 knowledge sharing to increase creativity, and, especially in the structurally differentiated
organization where the isolation of subunits institutionalize the inherent problems of distributed knowledge.
2.2.3 Coordinating Mechanisms
A coordinating mechanism is necessary to alleviate the nature of dispersed organizational knowledge. First, individuals and groups need to improve practical and political capacities to represent and transform knowledge (Carlile, 2004). Alleviate competitive elements of internal resources and external market share between units (Tsai, 2002). Closing cultural and cognitive gaps creating one “thought world”, improving ambiguity and absorptive capacity and thus both the ability to share and to receive and integrate dispersed knowledge assets (Szulanski, 1996). Previous research has highlighted the importance of coordinating mechanisms to promote the development of organizational capabilities (Grant, 1996). Capabilities being a firm’s ability to access, share and integrate organizational knowledge (Kogut and Zander, 1996). As such, the need of coordinating mechanisms is highlighted by numerous researchers to reduce the sticky nature of knowledge and barriers to knowledge flows, in order to increase access to dispersed organizational knowledge.
Despite the consensus, little evidence exists of coordinating mechanisms in the multi-unit organization. However, there are some clues on what they might entail. The social network perspective holds that knowledge sharing and coordination depend on intra-organizational networks (Hansen, 1999). That is, a unit may access knowledge through a network in which subunits are connected through formal and informal linkages. Similarly, Tsai (2002) found that increased inter-unit social interaction improve internal knowledge flows. Moreover, these
15 knowledge sharing. This resonates with the idea that no single individual or centralized function could possibly comprehend all knowledge available within the firm (Tsoukas, 1996).
Furthermore, informal lateral relationships may alleviate the barriers to knowledge flows and especially the dimensions of cognitive and cultural differences. These differences emerge as a result from lacking formal social interactions in the structurally differentiated firm. Accordingly, informal social relationships are the main lead to the characteristics of an ad hoc coordinating mechanism of dispersed organizational knowledge.
Social interaction in the multiunit organization emerges as a central theme in which innovation performance can be improved. It rests upon the inherent logic that knowledge is socially constructed, and that organizational learning involves a complex social process in addition (Huber, 1991). Through interaction between units a common language and
understanding is developed with an increased absorptive capacity as a result (Carlile, 2004; Lichtenthaler, 2009). The access to knowledge is important but if the knowledge cannot be integrated it is useless. Thus, it is the ability to both access and integrate knowledge that is a source of a competitive advantage (Tsai, 2001).
Thus far the argumentation has focused on the need for less hierarchical constraints and more social interaction to encourage internal knowledge flows (Tsai, 2002), improving
organizational capabilities (Kogut and Zander, 1996). That is, the access to knowledge is an important ingredient and input of the innovative process. Moreover, the issues that arise from distributed knowledge are enhanced in a structurally differentiated organization. The question is, why do firms structurally differentiate their operations in the first place? Research on
organizational design enlightens us why the target is not the source of the rigidity of knowledge flows. That is, structural differentiation being another important contender of increased speed
16 and success of innovation. Instead, coordinating mechanisms are needed to allow for the need for structural isolation, and simultaneously improve knowledge flows through coordination.
2.3 Structural Differentiation and the Speed and Success of Innovation
In order for an organization to achieve its strategic objective of continuous change through innovation, both radical and incremental innovations are required. This is achieved through the processes of exploitation and exploration (Raisch, 2008). However, these processes have inconsistent and contradictory demands (March, 1991). The efficient workings of explorative and exploitative innovation require that a balance between these units is strengthened by organizational design. Accordingly, structural differentiation creates a more focused effort in both processes through (1) direction, (2) specialization and (3) lowered interference through structural isolation. First, it gives the sub-unit one strategic direction and increased motivation through the given autonomy to pursue either exploitation or exploration (Jansen et al., 2009). Second, simplification through departmentalization, task ownership and repetition, the pooling of experiences and specialization increase learning and the efficiency in which a particular task is performed (Levinthal & March, 1993). Third, the structural separation allows for flexibility to adapt to local requirements, and, ensures the preservation of a singular objective and
specialization through decreased interference between them (Volberda, 1996; Benner and Tushman, 2003). That is, exploitation is often a dominant managerial logic that may concentrate its focus on leveraging current capabilities which has a negative effect on the ability to sense and seize emerging opportunities. Thus, through separation, structural differentiation can bridge the inherent contradictions and paradoxical challenges of exploration and exploitation and
accomplish the strategic objective of continuous change (Burgelman, 1984; Van de Ven, 1986; O'Reilly and Tushman, 2008).
17 Although scholars disagree on the degree of separation, there is a general consensus, and factual evidence to support the acclaimed functioning of structural differentiation. Empirical support has been gathered which demonstrates that structural differentiation aids in new venturing (Burgers, Jansen, van den Bosch and Volberda, 2009). With creativity at the root of innovation in organizations, structural differentiation leads to higher creativity through a combination of perceived task ownership and freedom (Amabile, Conti, Coon, Lazenby and Herron, 1996). Benner and Tushman (2003) stressed that although process management
activities can improve exploitation, it will have a negative effect on explorative activities if not separated. Despite the fact that some scholars argue that exploitative and explorative units need to be completely separated (Christensen, 1998), the majority of scholars promotes an approach where the conflicting activities of exploitation and exploration can be balanced by a loose separation of these units (Burgelman, 1984; Jansen et al., 2009).
The improved performance achieved by structurally differentiating organizational units is a result of the combinative power of a myriad of positive outcomes. First, by lowering the
intervention between units by segmentation, specialized units can coexist without interference (Golden and Ma, 2003). Without interference they can focus on the respective contextual, functional and directional demands of explorative and exploitative innovations (Marsch, 1991; Raisch et al., 2009). As such “pragmatic boundaries” emerge that protect creativity in explorative units from being affected by the dominant managerial inclination of exploitation in larger
organizations (Carlile, 2004; Benner and Tushman, 2003). Accordingly, this structural flexibility allows for a pooling of resources and experiences to meet the sole objective of the respective unit (Volberda, 1996).
18 Second, improved learning, deeper understanding and routinization enhance
specialization and grant the utilization on the emerging expertise to be applied locally to internal problems and external opportunities that these units encounter (Raisch, 2008). The resulting specialization result in a more fast-paced problem solving, a more efficient innovative process which improves time to market through the speed in which new products can be rendered (Lawrence and Lorsch, 1969). The evolving expertise allows for a more timely and accurate response to opportunities and threats through evolving task specific routines, through both exploration or exploitation (Raisch, 2008). Their fine-tuned senses allow for the opportunity to be seized at an appropriate moment in time by reconfiguring the resource based (Teece, 2007). Together with improved effectiveness of product rendering ultimately increase the rate of the success that new products and services generate, of which speed and timing both have been identified as key success factors in new product launches (Di Benedetto, 1999). Accordingly the first hypothesis concerns the main effect of the proposed model, where structural differentiation is expected to have a positive effect on the speed and success of innovation. Thus,
Hypothesis 1: Structural differentiation has a positive effect on the performance outcome: speed and success of innovation.
2.4 Coordination and Integration of Dispersed Organizational Knowledge: Achieving Speed and Success of Innovation
Although structural differentiation solves the paradox of the conflicting demands of explorative and exploitative innovation processes, the previously addressed problem of knowledge barriers remains. In addition, at least three new challenges arise as a result of isolating differentiated units, that is, lowering the knowledge flows through ceased formal social interactions: (1) the lowered awareness, access to and coordination of internal resources and capabilities on which the
19 innovation process depends (Westerman, McFarlan and Iansiti, 2006), (2) the integration of knowledge output of spatially dispersed units (Siggelkow and Levinthal 2003, Westerman et al. 2006) and (3) facilitating the process of innovation as a unified effort (Van de Ven, 1986).
First, in a structurally differentiated organization the level of interrelations and shared routines are reduced through the limited formal social interaction that structural isolation implies. Intuitively structural differentiation should lower the awareness of available knowledge and thus limit the innovative capacity. That is, connectedness within units and throughout an organization has been demonstrated as a particularly important antecedent of both exploratory and
exploitative innovation (Jansen et al., 2006; Tsai, 2002). These social practices need be replaced with a coordinating mechanism that can substitute their functioning to deliver efficient and successful innovations.
As demonstrated earlier, informal social relations can facilitate such coordination. Informal social relations are a voluntary mode of coordination based on informal personal
linkages (Tsai, 2002). Empirical evidence has shown that it improves knowledge sharing, but not yet been linked to improve innovational performance. However, the new informal relationship establishes a new connection that transcend structurally differentiated units, and increases the access to and awareness of available knowledge and capabilities through substitution with increased performance in exploitative and explorative units as a result (Jansen et al., 2006). Increased connectedness through informal social relationships lowers the barriers to knowledge flows (Tsai and Goshal, 1998), which are accentuated by structural differentiation. That is, the informal nature of these relationships traverse the organizational structures which may mitigate socio-cultural, cognitive and communicative divides between spatially dispersed units. In conclusion, the missing link is a mechanism that increases the awareness, access to and
20 coordination of organizational knowledge by creating a connectedness through informal social linkages between spatially dispersed units.
Challenge number two, structural differentiation poses a challenge through its implications on the management of knowledge integration from structurally separated and differentiated units. Scholars that promote the differentiation approach argue that output from explorative and exploitative units need to be recombined, achieving synthesis, in order for value to materialize (O'Reilly and Tushman, 2007; Raisch, Birkinshaw, Probst and Tushman, 2009). Accordingly, we can couple structural differentiation with integration mechanisms that can access and incorporate the produce of spatially dispersed sub-units (Jansen et al., 2009).
Integration in this sense is “the process of achieving unity of effort among various subsystems in the accomplishment of the organization's tasks” (Lawrence and Lorsch, 1967: 4). The
fundamental issue is not a decision that concerns solely organizational structure but whether there are complementary processes available that can facilitate integration of these separate exploratory and exploitative subunits in a way that promotes value creation (Teece, 2007; O'Reilly and Tushman, 2008).
Third, innovation is a process, a collective achievement where the speed of its development and the rate of its success is partially determined by production, sales and marketing departments in addition to research and development (Van de Ven, 1986; O’Reilly and Tushman, 2004). Certain coordinating and integrating mechanisms such as senior executives vision (Tushman and O’Reilly, 1997) and tight senior team integration have both been identified (Benner and Tushman, 2003) and verified (Jansen, van den Bosch and Volberda, 2006).
However, the nature of dispersed knowledge makes it difficult to gain oversight (Becker, 2001), which is implied by a coordinating and integrating mechanism located in the elevated levels of
21 the organizational hierarchy. Moreover, scholars remain open to the existence of different
mechanisms, including formal and informal integration (Gilbert, 2006). The “how” of
coordinating dispersed knowledge in vast organizations is more complex to only include a few centralized coordinating and integrating mechanisms (Tsoukas, 1996). A distinct mechanism is necessary that is cross-functional, that transcends dispersed organizational units and that has the ability to coordinate without being an institutionalized “control room”. In a structurally
differentiated organization, with departmentalized explorative and exploitative units, an additional mechanism is required that can facilitate the innovative process as a unified effort.
In order for a structurally differentiated firm to enjoy an improved speed and success of innovation it thus must have one or more mechanisms that ensure (1) the awareness of and access to expertise through coordination over time and across dispersed units and, (2) a
successful integration of the new knowledge that was the result of the activities performed within the separated units and (3), the facilitation of a unified effort that ensures the completion of the innovative chain and the launch of the new product or service. By resorting to the knowledge management literature mechanisms that possess these characteristics can be identified. 2.5 An Organizational Transactive Memory System
A transactive memory system (from now on referred to as TMS) aids an organization in the “acquisition, storage, distribution and retrieval of organizational knowledge and information” (Adams and Lamont, 2003, p. 144. ). TMS is a mechanism built on team-level tenets (Wegner, 1986), as an interconnected transactive memory of individuals within an organization
(Peltokorpi, 2008). It has been used by scholars to study cognitive processes in knowledge intensive teams where it allows for a maintained awareness of available expertise as members of an organization specialize in a certain task (Lewis, 2003). The transactive memory of an
22 individual is a result of the interaction between two components, (1) the internal memory
(expertise of the individual) and (2) the external memory (what individuals know collectively about the expertise of others within the group) (Peltokorpi, 2012). It emerges through increased social interaction and team member familiarity, as the frequency of interaction and
communication improves between members of an organization (Moreland and Argote,
2009). The social interaction results in social networks through direct face-to-face interactions and e-mails as well as indirect interaction when knowledge is shared in databases where it can be located and retrieved (Moreland, 1999). Although the antecedents of a TMS can be traced back to dyads and workgroups (Lewis & Herndon, 2011), it has recently been extended to
organizational level where it aids in a collective processing of dispersed knowledge (Jackson, 2012; Peltokorpi, 2012; Peltokorpi, 2014).
The empirical evidence of TMS in improving innovation performance are inconclusive and largely conceptual (Moreland and Argote, 2003; Peltokorpi, 2014). It has been promoted as a useful tool in dynamic environments, where it spurs creativity by increased utilization of an organization’s dispersed knowledge assets with an improved innovation performance as a result (Lewis and Herndon, 2011; Argot and Ren, 2012). Through improved efficiency, scope and flexibility of knowledge integration and recombination, an organization can better utilize
intellectual capital available to the organization (Kogut and Zander, 1992). Recent research have found that TMS facilitates knowledge sharing through informal lateral relations that emerge from inter- and intra-unit interactions (Peltokorpi, 2014). However, because TMS is improved by keeping the same teams over time it is reported to have a negative impact on idea generation due to increased homogeneity of the team composition creating rigidity in the formal processes and routines of the workgroup (Ren and Argote, 2011). In addition, a negative effect was found
23 between TMS and speed-to-market (Akgun, Byrne, Keskin and Lynn, 2006). Ren and Argote (2011) also report that research on TMS in isolation is inconclusive and that no research to date has established a relationship between TMS and innovation performance on the organizational level.
2.6 Transactive Memory Systems and Structural Differentiation
The inconclusive findings of the TMS functioning and the paradoxical workings of structural differentiation may be explained by a previously undetected relationship between them. As noted earlier, structural differentiation implies a formal isolation of subunits in order to allow for specialization to increase efficiency of explorative and exploitative efforts (Jansen et al., 2009). The isolation involves lowered connectedness through reduced formal social interaction between them which suggests a reduced ability to innovate (Jansen et al., 2006: Tsai, 2002). However, isolation through structural differentiation only implies a limited formal interaction between members of differentiated units. Knowledge being an important contender of innovative efforts, formal interaction may be substituted by a similar functioning. As Peltokorpi (2014) emphasizes, TMS emerges as a social network built on direct and indirect informal social interactions.
Knowledge being a fundamental component of the innovative process, members of spatially dispersed sub-units may rely on face-to face interaction outside the formal process of everyday routines, email and shared databases to access and retrieve required expertise (Lewis, 2003). As organizational units differentiate, and the members of these units specialize in a complex task, they rely increasingly on informal coordination mechanisms (Jackson, 2012), collective processing of dispersed knowledge assets (Peltokorpi, 2012), that allow individual members encode, store and retrieve knowledge from different but compatible domains (Peltokorpi, 2014). As such, interrupted formal linkages caused by structural differentiation may result in the
24 formation of informal social linkages that replace them. Accordingly, TMS is proposed to be caused by structural differentiation when formal social linkages are replaced with its informal counterpart.
2.7 Transactive Memory System and the Speed and Success of Innovation
An organizational TMS improves speed and success of innovation by increased access to and awareness of internally available knowledge and competences. It is a result of three functions that rest upon the logic of increased informal social relationships between members of an organization: (1) as an informal coordinating mechanism (Brauner and Becker, 2006), (2) reducing “stickiness” of organizational knowledge (Szulanski, 1996) by lowering barriers and utilizing pragmatic boundaries to knowledge flows (Carlile, 2002) and improving the
connectedness between members of an organization (Jansen et al., 2006: Tsai, 2002). Initially the presence of an organizational TMS, conceptualized as an informal coordinating mechanism, the speed and success of new product development by providing a shared awareness of organizational expertise (Jackson, 2012; Peltokorpi, 2014). Coordination is achieved through lowered drivers behind problems arising from dispersed knowledge: (1) large numbers of knowledge vectors in vast organizations, (2) asymmetrical distribution of knowledge resources and (3) structural uncertainty through limited ability of decision makers to specify all relevant outcomes (Becker, 2001). As such, TMS utilize a large number of organizational members into an asset rather than complication and relieve the pressure on the more centralized coordinating functions of decision makers (Tsoukas, 1996). Accordingly, TMS increases the willingness of members of an organization to share knowledge, as centralized and formal coordinating functions make members sensitive to share information (Moreland, 1999). Moreover, it complements formal coordination which by itself is insufficient in larger
25 organizations due to the costs associated with it (Van de Ven, Delbecq and Koenig, 1976).
Shared awareness improves mobility of dispersed knowledge assets and limits its asymmetrical distribution of through a coordinating function (Kogut and Zander, 1996). TMS exhibit several of the characteristics of the strategies that Becker (2001) identified that firms use to overcome the obstacles associated with dispersed knowledge, being knowledge substitution, coordination and increase available information through different information channels. Organizational knowledge becomes more accessible under the influence of TMS which improves both
exploratory and exploitative efforts that depend on the access to knowledge components (Teece, 2007).
Second, through informal lateral relationships TMS overcomes cognitive and cultural barriers to knowledge flows (Tsai, 2002). That is, the social interactions make knowledge less “sticky” by establishing a relationship between the sender and the recipient, creating a common language minimizing ambiguity and improving absorptive capacity (Szulanski, 1996). The social aspect of the TMS functioning allows for tacit knowledge to be transferred, translated and
transformed. That is, the lowered syntactic, semantic and pragmatic boundaries are lowered through the social interaction and informal lateral relationships that an organizational TMS entails. As innovation often emerge at the pragmatic boundary when knowledge is transformed in-practice (Carlile, 2002), TMS allows for novelty to materialize within units and as a result of the interaction between units through cross-fertilization. Accordingly, TMS reduces barriers to knowledge flows and encourages innovations to emerge through spontaneous re-combinations of existing knowledge and competences.
The increased connectedness between members of an organization (Jansen et al., 2006) combined with improved mobility of knowledge assets as well as reduced barriers to knowledge
26 allows for a more speedily and accurate responses to opportunity and threats. In turn, knowledge flows improve the access to knowledge, improving capabilities and organizational learning with increased creativity as a result. Increased awareness of available expertise, acquisition and improved distribution, the organization can respond in a more timely fashion to both
opportunities and threats by reconfiguring its resource base. To exploit “existing internal and external competencies to address changing environments'' (Teece et al., 1997, p. 510) it requires that organizational members are aware of which competences are currently available within the organization. An organizational TMS can expand the awareness of what available organizational knowledge and thus the access to it. An organizational TMS may facilitate timing through its memory functioning where the storage of tacit knowledge can be retrieved at an appropriate moment in time (Adams and Lamont, 2003). Naturally, lacking awareness of accessible components would slow down this process and thus the ability to respond to an opportunity would be lost. As knowledge assets are the main input in any innovative process the presence of a TMS is likely to improve the effectiveness of both explorative and exploitative units.
2.8 Structural Differentiation and the Speed and Success of Innovation: the Mediating Role of a Transactive Memory System
The logic behind studying TMS in and its mediating effect between structural differentiation and the speed and success of innovation are multifold. That is, integrating TMS on a firm level assists the structurally differentiated organization in achieving more speedily and successful innovations through three distinct functions: (1) facilitating task specialization and differentiation without interrupting knowledge flows (Peltokorpi, 2014), (2) facilitating integration of new competences that emerge as a result of structurally differentiated explorative and exploitative units and (3), through coordinating and integrating functions TMS taps into the value of
27 departmentalization (specialization) and simultaneous promotion of the innovation process as a unified effort (Van de Ven, 1986).
TMS facilitates to achieve the objective of structural differentiation, improved
effectiveness of subunits through specialization and direction, without limiting knowledge flows. In the case of achieving speedily and successful innovations these subunits concern exploration and exploitation (Raisch et al., 2009). These units require that internally available expertise remains accessible while simultaneously allowing for specialization in complex tasks. TMS grants the efficiency of expertise role differentiation to be achieved while promoting shared expertise awareness simultaneously (Lewis, 2003; Peltokorpi, 2014). Accordingly, TMS
facilitates informal coordination where organizational members aware of available expertise can access it through both lateral interactions and ICT (Moreland, 1999). Moreover, TMS allows structural differentiation to meet the differing contextual and architectural demands of
explorative and exploitative capabilities (Marsh, 1991). As such, institutionalization of barriers to knowledge flows is avoided and the objective of structural differentiation is achieved. In case of innovation this implies increased efficiency of explorative and exploitative efforts, improving speed and success of innovation.
Second, TMS facilitates the integration of the new knowledge that was the result of the activities performed within structurally differentiated units. More precisely this is the result of the factors that distinguish TMS “(1) differentiated knowledge, (2) transactive encoding, storage, and retrieval processes; and (3), the dynamic nature of TMS functioning” (Lewis and Herndon, 2011, p. 1256). TMS is a more flexible and including mechanism than the formal coordination of team executive vision and senior team integration, mainly due to the fact that it is impossible to gain overview of all available organizational knowledge. TMS transcends hierarchical levels and
28 includes vertical and horizontal dimensions that a top-down perspective arguably could not comprehend (Tsoukas, 1996). TMS facilitates both vertical and horizontal transfer of new knowledge, where it can be implemented and replicated facilitating institutionalization which is required for the value of new innovations to materialize (Raisch et al., 2009).
Third, TMS ensures speedily and successful innovations through the support of the innovative process as a unified effort. That is, a transactive memory is being initiated on an individual basis out of which TMS emerges as a collective capability. As such, it is equally dispersed as the knowledge assets it is proposed to coordinate and integrate. In combination TMS facilitates knowledge as an input to the innovative process through coordination, and, ultimately integration on the new competences that emerge through the re-combinative
capabilities of structurally differentiated sub-units. A well-functioning TMS minimizes power relations (Emerson, 1962), the competitive and self-interested behavior that emerges in
structurally differentiated organizations as sub-units compete for internal resources and external market shares (Bower, 1970). That is, resolving tensions and achieving synergies between spatially dispersed sub-units (Teece, 2007) and thus facilitating the innovative process as a unified effort (Van de Ven, 1986)
Conclusively, TMS is proposed to be a mechanism that underlies the observed relationship between structural differentiation and improved innovative performance. The co-specialized qualities of TMS in a structurally differentiated firm allows for the conceptualization of its functioning as a firm-level informal coordination (Peltokorpi, 2014) and integration
mechanism (Raisch et al., 2009). Its mediating role in a structurally differentiated firm is composed of several factors. The informal social relations TMS can balance the need for structural differentiation and simultaneously facilitate a balanced allocation of internal
29 knowledge resources. Moreover, as innovation is a strategic objective, TMS enable structurally differentiated units in achieving coherence in meeting demands of both radical and incremental processes. The lowered tensions and the creation of a unified effort, allows for the synergetic effects of spatially dispersed units to be realized. If TMS facilitates synthesis through
coordination and integration through increased awareness of and access to organizational knowledge it may be a solution to the antithesis of dispersed and differentiated organizational units (Lawrence & Lorsch, 1967). The preposition is that the workings of a developed TMS will positively mediate the relationship between structural differentiation and innovation
performance, with improved speed and success of innovation as a result. Thus,
Hypothesis 2: Transactive memory systems positively mediate the relationship between structural differentiation and the speed and success of innovation.
3. METHODOLOGY 3.1 Research Setting
The sample was obtained from a survey called Innovation Benchmark that was carried out in 2012 among a total of 408 Dutch firms. Initially the survey was designed to improve our understanding of the differences in performance with regard to general organizational
characteristics including strategy, investments and knowledge management. The survey included the assessment of the four variables pertaining to this research as well as certain useful control variables. When these variables had been selected the output of 220 surveys with complete data remained. The respondents were CEO/directors and had a mean age of 46.4 years (standard deviation (s.d.) = 8.6). The size of the firms had an average of 312 employees (s.d. = 1965.7) and were operating in a broad range of different industries including manufacturing, construction, wholesale, transportation, financial services, other professional services, and other industries.
30 3.2 Measurement and Validation of Constructs
Dependent variable: Speed and success of innovation. The dependent variable of speed and success of innovation (α = 0.79) is a second order construct based on two factors with a four item scale (see Appendix A): Success of products and/or services (α = 0.84) and Speed of
development of products and services (α = 0.77). The questions concern the speed in which new products /services were developed and launched and the performance measure of the rate of success the new product/service enjoyed. The answers to the questions regarding both speed of product/service development and its success were asked to be given in comparison to the relative performance of its largest competitor.
Previous to the factor analysis and computation of the variable factorability was examined. First, the Kaiser-Meyer-Olkin measure of sampling adequacy was .79, above the recommended value of .6. Second, the four item scale correlated with at least β = .5 (p = .000) in factor one, and the four items of factor two correlated with at least β = .3 (p = .000). Lastly, communalities were above .3 for all eight items which confirms that they shared some common variance with other items. Given the combination of the above indicators a factor analysis was regarded suitable including all items pertaining to this variable.
The exploratory factor analysis confirmed the two factor structure with significant correlations, and, meeting the criteria of factor loadings above .4 (based on N = 200) and cross-loadings below.3 (see Appendix B). After this confirmation the two factors could be computed into a single variable.
Independent variable: Structural differentiation. Following the study conducted by Jansen et al. (2009) a six item scale (Appendix A) was used to represent structural differentiation (α = 0.64). Initially these items determine the extent to which an organization is divided into units, spatially
31 separated and with unique characteristics that mirror the needs of the specific function of that unit and the demands of its environment. In addition, the differences between the units that surface through this scale include mindset and time orientation (Lawrence and Lorsch, 1967, Golden and Ma, 2003, Jansen et al., 2009).
Kaiser-Meyer-Olkin measure of sampling adequacy of .69 was determined. Second, the six items correlated with at least β = .05 in factor one. Item one and five were the only two items that did not correlate significantly. With communalities above .3 an exploratory factor analysis was performed which confirmed the significant correlations among the six items, and, meeting the criteria of factor loadings above .4 (based on N = 200) and cross-loadings below.3 (see Appendix B). Accordingly, the items were computed into the variable that represent structural differentiation. Mediating variable: Transactive memory system (TMS). TMS is a second order construct that measures the collective system of knowledge and the way the members of an organization encode and retrieve stored knowledge from different and separated units (alpha = .88). Moreover, it is based on three factors that researchers have found to indicate the existence of such a system (Argote and Ren, 2012; Lewis, 2003; Moreland et al., 1998). The factors include The localization of expertise within the organization (α = .83)(five items measuring the degree of an individually developed knowledge and memory specialization), The quality of expertise ( α = .6)(four items tapping into the credibility and trust individuals have for each other’s expertise) and Knowledge coordination ( α= .87)(four items indicating the ease at which group members can collaborate to complete a task)(see Appendix A).
The variable factorability is considered good with the Kaiser-Meyer-Olkin measure of sampling adequacy of .86. Second, the four items correlated with at least β = .58 (p = .000) in factor one, and the four items of factor two correlated with at least β = -.013 (p = .000) and factor
32 three with at least β = .54 (p = .42) . Last, communalities were above .54 for all 13 items.
However, reliability improved significantly when item number three was removed from factor one.
The following exploratory factor analysis revealed the anticipated factor structure with significant correlations (β =.54, .67 and .53, p = .001), factor loadings above .65 and cross-loadings below.3 (see Appendix B). After this confirmation the three factors could be computed into a single variable.
Control Variables. To remain open to alternative explanations control variables were used in this study. First, firm size was included since larger organizations are more likely to be structurally differentiated compared to smaller firms due to the financial means involved. Similarly, TMS as a collective function is dependent on that the firm is big enough to include a number of work groups. Accordingly the number of full time employees has been used to control for firm size. Second, R&D investment was included to control for differences in investments between firms and industries. In addition, we control for return on investment and sales in order to have
additional performance measures complementary to speed and success of innovation. Fourth, the variable cross functional collaboration (α = .83) was included to control for other integrating mechanisms and thus improve the trustworthiness of the results. Cross functional collaboration is composed of six items (see Appendix A) and met the general criteria of factorability (see
Appendix B). Lastly, to control for contextual differences industry dummies were used as exploratory and exploitative efforts might differ between industries (Jansen et al., 2009), and thus the extent to which structural differentiation is applied. The 220 firms were divided into a total of seven industries including in accordance with standard Industry Classification Codes:
33 manufacturing, construction, wholesale, transportation, financial services, professional services, and other.
3.3 Computing Variables
After determining the correlation between the items, confirmed the factor structures and reliability with an alpha above .7 (except structural differentiation with an alpha of .64), the factors were computed into variables.
3.4 Aggregation
Since the variables are individual level constructs, and with the hypotheses concerning firm level relationships, the variables needed to be aggregated from individual to firm level. However, because the number of respondents among the 220 firms used in the computations only had one respondent, an aggregation would not make a difference with regard to the analyses. As such the individual level equals that of the firm level constructs and will be used and interpreted
accordingly.
4. ANALYSIS & RESULTS
Two tables are included where Table 1 shows correlations and the descriptive statistics
pertaining to all variables and Table 2 includes the output of the performed regression analysis for speed and success of innovation.
Model 1 includes the chosen control variables. Model 2 concerns the effect of structural
differentiation on speed and success of innovation. Model 3 demonstrates the functioning of the proposed model in its entirety, including the proposed mediating variable transactive memory system.
35 Table 2 Results of Hierarchical Regression Analyses:
Structural Differentiation, Coordination, Integration, and The Speed and Success of Innovation
Model 1 Model 2 Model 3 Control variables
Cross functional teams 0.29*** 0.25*** 0.12
Firm size -0.17 -0.02 0.04 Sales 0.07 0.09 0.11 Return on investment 0.03 0.01 -0.04 R&D investment 0.22 0.19* 0.14* Manufacturing 0.34**** 0.34**** 0.34**** Construction -0.76 -0.1 -0.05 Wholesale -0.04 -0.07 -0.00 Transportation -0.04 -0.08 -0.00 Financial services -0.03 -0.06 0.04
Other professional services -0.00 -0.04 -0.24**
Other -0.36*** -0.40 -0.32
Independent variable
Structural differentiation 0.17* 0.14
Mediating variables
Transactive memory system 0.42****
Adjusted R² 0.36 0.30 0.39
Δ adjusted R² 0.36**** 0.39**** 0.48****
Note. Standardized regression coefficients are reported. *p < 0.10, **p < 0.05, ***p < 0.01, **** p < 0.001.
In order to test the proposed model, including the relationships between structural
differentiation and speed and success of innovation, the four step procedure of Baron and Kenny (1986) was applied.
Initially the relationship between structural differentiation and speed and success of innovation is examined in isolation. As demonstrated in Model 2, the proposed relationship was found to be positive and significant (β = 0.17, p < 0.10). Second, speed and success of innovation need to be significantly related to the mediating variable. The results (Model 3) confirmed that transactive memory system was significantly related to speed and success of innovation (β = 0.42, < 0.001). Third, structural differentiation needs to significantly predict transactive memory system,
36 accordingly a regression analysis was performed with the mediating variable as a dependent variable and structural differentiation as an independent variable. The regression demonstrated that structural differentiation is significantly associated with transactive memory system (β = 0.27, p < 0.01). Four, when the mediating variable is introduced in the regression model the relationship between structural differentiation and speed and success of innovation need to become insignificant. When running the full regression model an insignificant relationship was established (β = 0.14, p > 0.10). Thus, this indicates that a transactive memory system and transformative capacity fully mediates the relationship between structural differentiation and speed and success of innovation.
Although the four-step approach from Baron and Kenny (1986) is a common approach among scholars their procedure is not without critique. First, the significance of the indirect effect is not tested. Second, it may fail to detect certain true mediation effects, or, type II errors (MacKinnon, Fairchild, & Fritz, 2007). Alternatively, the indirect effect should be tested for significance. Consequently an additional statistical test will be used to address the indirect effect. Following Preacher and Hayes (2004) PROCESS mediation will be applied for this purpose.
To verify these findings, additional regression analyses were performed in order to assess the different components of the proposed mediation model. First, it was found that structural differentiation was positively associated with speed and success of innovation (B = .16, t (206) = 1.94, p = 0.06). Moreover, structural differentiation was found to relate positively to transactive memory system (B = .19, t (206) =2.21, p = 0.03). Lastly, the mediator transactive memory system was found to be positively related to speed and success of innovation ( B =.29, t (206) = 2.76, p = .007). Having confirmed the significance of both the a-path and b-path, a mediation analysis was performed using bootstrapping method with bias-corrected confidence estimates
37 (MacKinnon, Lockwood & Williams, 2004). A confidence interval of 95% of the indirect effect was obtained with 5000 bootstrap examples (Preacher and Hayes, 2008). The results of the mediation analysis confirmed the mediating role of a transactive memory systems in the
relationship between structural differentiation and speed and success of innovation (B = .06: C1 = .005 to .14: z = 1.66, p = .1 ). Moreover, the direct effect of structural differentiation on speed and success of innovation became insignificant (B =.09, t (206) = 1.25, p = .21) when
controlling for transactive memory systems, thus suggesting full mediation.
With regard to the two hypotheses the correlation, regression and mediation analyses have provided some interesting results. The output confirmed Hypothesis 1, the proposition that a structurally differentiated firm can enhance performance outcomes of speed and success of innovation. Similarly empirical support was found for Hypothesis 2, the mediating effect of transactive memory systems on the relationship between structural differentiation and the speed and success of innovation. Moreover, the findings proved that an organizational transactive memory systems indeed promotes effectiveness and efficiency of the innovation processes with improved speed and success of innovation as a result.
5. DISCUSSION
This thesis has sought to answer two questions: (a) how do structurally differentiated
organizations handle problems associated with dispersed organizational knowledge? (b) Does increased efficiency of structural differentiation, and improved coordination and integration of dispersed knowledge assets through TMS improve the speed and success of innovation?
The findings suggest that structural differentiation, by itself, does not solve issues associated with the dispersed nature of organizational knowledge. On the contrary it
38 result of the creation of sub-groups of which separation lowers shared routines and thus formal social interactions. The objective of structural differentiation being efficiency gains by
specialization through departmentalization it relies on additional functions, formal and informal coordinating mechanisms to ensure knowledge flows. The importance of the access to
knowledge forces coordination of dispersed knowledge assets to occur through informal channels, giving rise to informal coordinating mechanisms such as TMS.
Accordingly, an interaction effect was established between structural differentiation and TMS. TMS was found to coordinate dispersed knowledge assets through increased awareness and access of internally available knowledge and expertise. As such it allows for the increased efficiency of structurally differentiated units that involve explorative and exploitative
innovations through specialization and continued access to necessary competences. The result is a more speedily rendering of new products and services, meeting the differing demands of both current and future market.
5.1 Theoretical Contributions
Initially, the findings confirmed the assertion that structural differentiation has a positive influence on the speed and success of innovation. As such it lends support to theories on organizational design that argue that efficiency gains can be achieved through structural
differentiation of organizational units (Lawrence and Lorsch, 1969) by meeting their inconsistent of demands on an ad hoc basis (Marsch, 1991). Similarly, it provides evidence that structural differentiation aids in improved innovation performance by providing and maintain different strategic directions and flexible functioning of subunits that enables an organization to meet the demands of both current and future markets (Golden and Ma, 2003; Volberda, 1996). Moreover, the emerging “pragmatic boundaries” (Carlile, 2004) promote explorative efforts by protecting it
39 from being influenced by exploitative activities (Benner and Tushman 2003; Jansen et al., 2009). The combined effect of the improved efficiency of structurally differentiated units provide an organization with the capability to a more fast paced and timely alterations of its resource base providing speedily and successful innovations.
Proof was found that support the hypothesized relationship that TMS mediates the relationship between structural differentiation and the speed and success of innovation. The coupling of structural differentiation and TMS adds to theories on TMS which have remained largely conceptual (Moreland and Argote, 2003; Ren and Argote, 2011) and lacking on an organizational level (Peltokorpi, 2014). Especially the effect of TMS on innovation performance has remained inconclusive (Akgun et al., 2006; Lewis & Herndon, 2011) as
research has remained focused on team level outcomes. These findings are especially valuable in broadening our understanding of TMS on different levels (e.g. organizational), and is different contexts signified by different functional capacities (e.g. innovation) as well as varying levels of knowledge intensity (e.g. industry).
The most significant contribution of this thesis to prior literature is on coordination and integration mechanisms and as means to handle dispersed organizational knowledge. As such, the findings correspond to overlooked aspects of organizational coordination design theories in the TMS literature (Kogut and Zander, 1996; Peltokorpi, 2014). First, although the importance of centralized through formal linkages has been established (Jansen et al., 2009) recent studies suggest that formal coordinating mechanisms alone are insufficient (Faraj and Xiao, 2006). The inadequate working is related to the lacking capabilities of centralized functions to gain overview of available knowledge (Tsoukas, 1996), and the sensitivity organizational members might feel about sharing information when its ordered from a centralized position (Moreland, 1999). This
