Knowledge Management Capacity and Ambidexterity: how to
become ambidextrous by reconfiguring and realigning your
knowledge capacities.
University of Amsterdam Amsterdam Business School
Executive Programme in Management Studies - Strategy Track MSc Thesis
Author: Ben van Vliet Student number: 10475702 Date: June 30th, 2014 Version: Final
First supervisor: Jeroen Kraaijenbrink Second supervisor: Sebastian Kortmann
Acknowledgements
Welcome to the final assignment of the Executive Programme in Management Studies at the Business School of the University of Amsterdam, my master thesis. By taking part in the executive programme and simultaneously working on challenging projects at work, I was able learn a great deal about managing organizations and to develop myself to the utmost. I would like to express my gratitude to the people that made this possible. First of all, I would like to thank my employer providing me with the opportunity to take part in the programme. I would like to thank Jeroen Kraaijenbrink and Sabastian Kortmann for their guidance, and for challenging me. I would like to thank Lies for her unconditional support. Lastly, I would like to thank fellow students whom I have worked with on projects and assignments; working together, thinking out loud, sharing knowledge, experiences, insights, and ideas, not being afraid of making mistakes, arriving at creative and effective solutions together, was truly inspiring.
Abstract
This study examines the relationship between knowledge management capacity and organizational ambidexterity. Both the effect of internal and external knowledge management capacity on ambidexterity is studied. It is proposed that both capacities will enhance the ability of organizations to effectively balance exploitative and exploratory innovations, and subsequently have a positive effect on performance. Evidence from data collected from business units in The Netherlands (n = 72), revealed that only internal knowledge management capacity has an effect on ambidexterity. The data also reveals a possible direct relationship between external knowledge management capacity and performance, indicating that outsourcing knowledge management capacity only contributes to short-term financial performance. Furthermore, environmental dynamism and competitiveness do not alter these relationships. By combining the integrative model on knowledge management and ambidexterity, this study contributes to the existing literature by revealing how knowledge management leads to innovative and financial performance, and by revealing a novel antecedent to ambidexterity.
Index ACKNOWLEDGEMENTS ... 3 ABSTRACT ... 4 1. INTRODUCTION... 8 2. THEORY ... 12 2.1ORGANIZATIONAL AMBIDEXTERITY ... 12
2.2KNOWLEDGE MANAGEMENT CAPACITY ... 15
2.3ENVIRONMENT ... 21 3. METHODS ... 25 3.1RESEARCH DESIGN ... 25 3.2SAMPLE ... 25 3.3KEY INFORMANT CHECK ... 27 3.4NONRESPONSE BIAS ... 27 3.5MEASURES ... 29
3.6VALIDITY AND RELIABILITY ... 31
3.7CONFIRMATORY FACTOR ANALYSIS ... 34
3.8COMMON METHOD BIAS ... 34
4. RESULTS ... 36
4.1TEST OF HYPOTHESES ... 37
5. CONCLUSION & DISCUSSION ... 44
5.1CONCLUSION ... 44
5.2DISCUSSION ... 48
5.3LIMITATIONS AND FUTURE RESEARCH ... 50
REFERENCES ... 53
Index tables and figures
Table 1. Knowledge management capacities ... 16
Table 2. Descriptive statistics ... 28
Table 3. Psychometric properties of first-order measurement scales ... 33
Tabel 4. Test of hypotheses, model 1 and 2 ... 40
Tabel 5. Test of hypotheses, model 3 and 4 ... 41
Tabel 6. Post-hoc, KMC and Exploratory and exploitative innovation ... 43
Table 7. Knowledge management capacities ... 47
Table 8. Measurement itemss and validity assessment ... 63
Table 9. Dutch translation measurement items ... 65
Table 10. Confirmatory factor analysis ... 67
Table 11. Common method variance analysis ... 68
Figure 1 conceptual model... 24
Figure 2 Results of structural equation modeling – Internal and external knowledge management capacity. ... 36
1. Introduction
Today organizations face multiple trends that put a continuous strain on their competitive advantage and their chances for long term survival, these are: a.) the increase of the pace of technological innovation, b.) the fragmentation of customer needs, and c.) a flattening of the global playing field (Boston Consulting Group, 2012). The trade-off organizations face in an environment characterized by these trends is clear; to be successful in the short run, managers must continuously strive for incremental improvements in the alignment of the strategy, structure, processes, and culture of today’s business (i.e. exploitation), to be successful in the long haul, managers must adopt new strategies, structures, and processes (i.e. exploration) that could potentially destroy today’s business. Organizations that are able to successfully balance this trade-off have been labeled ambidextrous (O’Reilly & Tushman, 2008).
Organizational ambidexterity has attracted substantial interest from the research community and has emerged as a new research paradigm in recent years. Duncan (1976) and March (1991) are often cited as the authors being responsible for the excessive interest in ambidexterity. Duncan (1976) is seen as the first to use the term ambidexterity and March (1991) proposed that exploration and exploitation are two fundamentally different types of organizational learning and that firms need to balance their attention and resources between them. Whereas earlier studies often concluded that the ambidexterity trade-off is insurmountable, later studies revealed a wide range of solutions, enabling organization to effectively and simultaneously balance exploration and exploitation (e.g. Gibson & Birkinshaw, 2004; Jansen et al., 2006). Today a comprehensive model that contains multiple antecedents, moderators, and environmental factors, a distinction of the different kinds of organizational ambidexterity, and a robust positive relationship between ambidexterity and
performance, is emerging (Reich & Birkinshaw, 2008). This study will use this robust relationship between ambidexterity and performance as a starting point, but proposes – in line with recommendations of Raisch and Birkinshaw (2008) to span disconnected research fields to stimulate cross-fertilization – to widen the scope of antecedents of ambidexterity. A promising domain for ambidexterity research is to move from its traditionally intra-firm focus to a inter-firm focus, because of the fact that the locus of innovation is increasingly moving towards the community or ecosystem in which the organization is embedded (Chesbrough, 2006; O’Reilly & Tushman, 2013). Firms across industries move from a relatively closed innovation process, i.e., innovation oriented activities being conducted inside organizational boundaries, to a more open innovation process, i.e., innovation oriented activities being conducted in collaboration with external partners, customers, and stakeholders. This fundamentally changes the way firms conduct their business (Chesbrough, 2006; Lichtenthaler, 2007). In an ecosystem characterized by open innovation, it is argued - in line with the knowledge-based perspective (Grand, 1996) - that organizational knowledge management capacity, is key to explaining inter firm heterogeneity in profiting from open innovation and thus potentially a major source of competitive advantage (Lichtenthaler, 2009). By considering knowledge exploration, retention, and exploitation both inside and outside organizational boundaries, Lichtenthaler and Lichtenthaler (2009) arrive at an integrative framework on managing organizational knowledge, containing six different capacities, these are: inventive, absorptive, transformative, connective, innovative, and desorptive capacity. Knowledge management capacity refers to an organizational dynamic capability that is able to reconfigure and realign these knowledge capacities.
In addition to a change in conducting business, as mentioned earlier, a change in arriving at being ambidextrous is emerging; therefore, knowledge management capacity could
reveal untapped additional explanatory power as an antecedent of ambidexterity and subsequently performance. The logic behind knowledge management capacity as an antecedent to ambidexterity is that effective exploration, retention, and exploitation of knowledge in an organization, enhances strategic decision making processes concerning the ambidexterity trade-off. Relative few studies have been conducted that explicitly link both internal and external knowledge management capacity to ambidexterity, indicating a void in the literature (O’Reilly & Tushman, 2013). By opening up this avenue, this study could reveal that both internal and external knowledge management capacity are sufficient or even necessary factors in the path to organizational ambidexterity, whether or not the organization operates in an ecosystem characterized by open innovation.
Earlier studies reveal environmental dynamism and competitiveness as being important moderators of effects concerning ambidexterity (e.g., Jansen et al., 2005). The logic behind these findings is that as dynamism and competition intensifies, the pace of change also intensifies, and therefore the tension to simultaneously exploit existing competencies and explore or develop new competencies becomes more severe. This study will examine how environmental dynamism and competitiveness influences the effect of both internal and external knowledge management capacity on ambidexterity.
The following contributions to our understanding of knowledge management capacity and ambidexterity are being made. First, it enhances knowledge on the antecedents of organizational ambidexterity. The knowledge management capacity framework helps to understand how organizations can capture value created during open innovation processes and become, not only more innovative, but potentially more ambidextrous. Second, this study further substantiates the knowledge on how environmental dynamism and competitiveness effects organizational ambidexterity. Third, an operationalization of the knowledge
management capacity framework is provided, plus guidelines on how to create the higher and lower order constructs of both the knowledge management capacities, and thus a basis for further empirical research. Together, these insights contribute to moving the ambidexterity theory into the paradigmatic status.
Managers might also benefit from the finding of this study; they are able to assess under which circumstances the development of certain knowledge management capacities is needed to contribute to resolving the ambidexterity paradox in their organization; under which circumstances it is important to extract knowledge from the external stakeholders, how that knowledge can be applied, for example, to create new products, enter new markets or to foresee that it is best to focus on optimizing current business because developing new products or entering new markets at this stage might contain too much risk.
Section 2 further describes the theoretical background of this study and introduces the associated hypotheses, section 3 provides a full description of the methods used in this study, section 4 contains the results of the research, and section 5 provides an interpretation of the results, i.e., how this impacts both ambidexterity and knowledge management research and practice, plus a description of limitations and subsequently directions for future research.
2. Theory
2.1 Organizational Ambidexterity
In an environment characterized by the disruptive trends described in the previous section, i.e., an increase in the pace of innovation, the fragmentation of customer needs, and globalization, it is often the incumbent that struggles most with adaptation, change, and the creation of dynamic capabilities needed in such a environment (Teece et al., 1997; Christensen, 1997; O’Reilly & Tushman, 2008). A central issue in these organizations is the tension between the exploration of new possibilities and the exploitation of existing certainties (Schumpeter, 1934; Kuran, 1988). The theory of rational search assumes that there are several investment alternatives each characterized by a probability distribution related to returns (Radner & Rothschild, 1975; Hey, 1982). These distributions are initially unknown; where the probability distribution of alternatives related to exploration is usually unstable and dependent on uncontrollable environmental factors; the probability distribution of exploitation is more stable and controllable. Thus, rational calculation often suggests concentrating the investments on exploitation of existing certainties with short-term results (March, 1991). This bias towards organizational short-termism can also explained by theories of limited rationality; it is assumed that the choice between exploration and exploitation is determined by targets or aspiration level (Cyert & March, 1963). Exploratory activity is limited or not preferred if current organizational performance is meeting or is close to today’s targets; this is also known as satisficing (Simon, 1955). An important mechanism that causes stable probability distributions related to exploitation, organizational inertia and subsequently a bias towards exploitation is path-dependency (Arthur, 1989). The current trajectory and behavioral patterns of organizations is to a large extent influenced by their current success, as well as competences
and capabilities that are focused on the current routines and cultures. This pattern causes a self-reinforcing dependency on the current direction of the organization (Arrow, 2003; Arthur, 1989; Williamson, 1981). Furthermore, Joseph Schumpeter stated that newcomers to the industry more often implement radical innovation and that consequently incumbents often face fierce competition by these newcomers (Schumpeter, 1942). But, more recent studies reveal that, contrary to the assumption of Schumpeter, incumbents are able to adapt to changing environments by developing dynamic capabilities. Teece et al. (1997) state that a dynamic capability is the firm’s ability to integrate, build and reconfigure internal and external competences, in reaction or anticipation to changes in the environment. Dynamic capabilities are sometimes referred to as ‘routines to learn new routines’ (Eisenhardt & Martin, 2000) which form the capacity for firms to modify their resource base (Ambrosini & Bowman, 2009). While the concept of dynamic capabilities has been used in many studies to explain how certain incumbents successfully transitioned to another industry; O’Reilly and Tushman (2008) state that there is a need for clear articulation of the routines and competencies that helps firms gain the ability for long-term sustainability; according to them, using the concept of ambidexterity can fulfill this need. O’Reilly and Tushman (2008) frame ambidexterity as a dynamic capacity and contribute the difficulty of incumbents to create this capacity to conflicting routines and cultures associated with exploitation and exploration. Combining these two contradictory cultures inevitably causes a clash, and with path-dependencies pulling organizations towards the status quo, exploitation often prevails (March, 1991; Levinthal & March, 1993). However, in order to secure short-term profits and long-term survival organizations have to balance exploration and exploitation, i.e., become ambidextrous (Ghemawat & Ricart Costa, 1993; van den Bosch, Volberda & de Boer, 1999; Gibson & Birkinshaw, 2004). Organizations that are ambidextrous are able to foster both cultures, meet
current demands, and simultaneously reconfigure their culture, routines, and capabilities, to meet future demands, break path-dependencies and survive disruption (Tushman & O'Reilly, 1996).
Current research acknowledges that the antecedents of organizational ambidexterity can be categorized into three perspectives; a.) structural mechanisms, b.) organizational context, and c.) leadership characteristics. First, structural or architectural solutions to cope with competing organizational demands were introduced (Tushman & O’Reilly, 1996). It was argued that efficiency and flexibility cannot coexist within the same business unit and should therefore be separated; e.g. one business unit with a mechanistic structure – that is characterized by standardization, centralization, and hierarchy – that supports efficiency, and another business unit with an organic structure – characterized by decentralization and autonomy – supporting flexibility (Burns & Stalker, 1961). In this way, organizations could become ambidextrous by having an ambidextrous portfolio. The second perspective focuses on organizational context, with the aim of achieving contextual ambidexterity (Gibson & Birkinshaw, 2004). Contrary to the former perspective, it is argued that ambidexterity can be achieved within the business unit by using behavioral and social means to enable individuals or teams to balance exploration and exploitation oriented activities. The third stream of research concerning organizational leadership, accentuates, in line with upper echelon theory (Hambrick & Mason, 1984), the role of top management in supporting innovations and the focus on balancing the ambidexterity trade-off (Kleinschmidt et al., 2007; Swink, 2000). This perspective frames organizational ambidexterity as a top management team capacity that enables organizations to align with their environment by means of ambidexterity oriented decisions (Kortmann, 2014). Furthermore, several studies have incorporated environmental factors as moderators or as important boundary conditions to organizational ambidexterity
(Leviathan & March, 1993; Volberda, 1996; Volberda & Bruggen, 1997). This contingency perspective argues that environmental factors, e.g. competitiveness and dynamism, shape the extent to which organization are able to balance and synchronize exploration and exploitation (Jansen et al, 2005). But, the fundamental question addressed by ambidexterity research is whether ambidexterity actually contributes to the performance of an organization. In spite of unclarity surrounding the measures and some scholar receiving no empirical support for this relationship (Raisch et al., 2009), the overall pattern appears to be clear: organizational ambidexterity seems to be positively associated with increased firm innovation, better financial performance, and higher survival rates (He & Wong, 2004; O’Reilly & Tushman, 2013).
A large amount of research has enhanced our understanding of how organizations actually tackle the seemingly insurmountable paradox, but there are also still multiple issues surrounding ambidexterity that are in need of further clarification (Raisch et al., 2009). One of these issues is how knowledge management contributes to resolving the ambidexterity paradox (Raisch et al., 2009; O’Reilly & Tushman, 2013).
2.2 Knowledge Management Capacity
In knowledge management research the distinction between knowledge exploration, i.e. the creation of knowledge, and knowledge exploitation, i.e. the application of knowledge is widely acknowledged (Nonaka, 1994; Argote et al., 2003; Bogner & Bansal, 2007). Furthermore, to bridge the time between knowledge exploration and exploitation, organizations need to build a capacity called: knowledge retention (Garud & Nayyar, 1994; Zollo & Winter, 2002). Recent studies underscore the possibility of organizing these processes beyond the boundaries of the organization (Grant & Baden-Fuller, 2004; Cassiman & Veugelers, 2006).
Furthermore, multiple authors, not only note that it is possible to organize these processes externally, but go one step further and argue that inter organizational knowledge transactions could enhance internal knowledge exploration, retention, or exploitation (Argote et al., 2003; Gulati & Gargiulo, 1999; Andersen & Drejer, 2008; Hargadon & Sutton, 1997; von Hippel, 1988). An integrative framework that includes both internal and external knowledge management capacity is developed by Lichtenthaler & Lichtenthaler (2009). This framework consists of six knowledge management capacities, see table 1.
Table 1. Knowledge management capacities
Knowledge exploration Knowledge retention Knowledge exploitation
Internal Inventive capacity Transformative capacity Innovative capacity
External Absorptive capacity Connective capacity Desorptive capacity
Framework by Lichtenthaler & Lichtenthaler, 2009
Internal and external knowledge management processes could be seen substitutes when analyzing individual projects. For example, an organization could face a ‘make-or-buy’ decision related to knowledge exploration, ‘integrate-or-relate’ decision related to knowledge retention, or an ‘keep-or-sell’ decision related to knowledge exploitation (Cassiman & Veugelers, 2006; Lichtenthaler, 2007; Lichtenthaler, 2009). But analyzing these decisions from the organizational level, the internal and external processes are rather complementary (Grant & Baden-Fuller, 2004). This underscores the notion that coordination on an organizational level is required to successfully reconfigure and realign these processes in order to make optimal use of these capacities, adapt to changing environments, and subsequently enhance performance (Eisenhardt & Martin, 2000). This dynamic capability is called knowledge management
capacity and refers to the ability of organizations to successfully manage its knowledge base over time (Zollo & Winter, 2002; Zahra et al., 2006; Lichtenthaler & Lichtenthaler, 2009).
Several scholars state that the dynamic capability that enables organizations to reconfigure, realign, and integrate both their internal and external knowledge processes, also enable organizations to become ambidextrous (Raisch et al., 2004; O’Reilly & Tushman, 2013). This is in line with multiple scholars that state that organizations requires knowledge management capabilities that support an organization’s complex set of decisions and routines that enable them to sense and seize new opportunities through the continuous reallocation of their organizational skills and assets, i.e. achieve ambidexterity (O’Reilly & Tushman, 2008; Taylor & Helfat, 2009). Furthermore, some scholars state that it is important to initiate exploratory initiatives to acquire new knowledge externally and that exploration beyond the boundaries of the organization has more impact on innovative performance compared with exploration within the boundaries of the organization (Rosenkopf & Nerkar, 2001; Raisch et al., 2004). Combining the knowledge management capacity model of Lichtenthaler and Lichtenthaler (2009) and the recommendations of O’Reilly and Tushman (2013), this study concludes that both high internal and external knowledge management capacity contributes to the level of ambidexterity of an organization, because knowledge management capacity enables organizations to acquire or develop knowledge needed to develop future profitable product–market combination, retain knowledge that is needed in the future, make optimal use of knowledge by exploiting it either internally or externally, contributing to balancing exploration and exploitation and subsequently cause the organization to performe.
The higher order construct of internal knowledge management capacity consists of three lower order constructs; these are: inventive, transformative, and innovative capacity. Inventive capacity refers to the ability of an organization to internally explore knowledge, i.e.,
to generate new knowledge inside the organization. The capacity to internally retain knowledge, created either internally or externally, over time is referred to as transformative capacity (Garud & Nayyar, 1994). Contrary to information, which consists of facts that can be codified, knowledge consists of know-how, which requires active management in order to keep the knowledge ‘alive’ if it is not implemented right away (Szulanski, 1996; Lane et al., 2006). Innovative capacity refers to the exploitative component of absorptive capacity (Lane et al., 2006; Zahra & George, 2002), meaning the ability of an organization to exploit or apply the knowledge it has created or required, internally. Innovative capacity includes converting knowledge into an organization’s products or services (Brown & Eisenhardt, 1995; Khilji et al., 2006).
Prior knowledge on the related field determines to a great extent, the level of inventive, transformative, and innovative capacity (Garud & Nayyar, 1994; Khilji et al., 2006). Therefore, this dynamic capability, i.e., internal knowledge management capacity, is required when organizations need to develop new creative ideas, develop new capabilities and resources that enable both exploration and exploitation innovation (O’Reilly & Tushman, 2008; Taylor & Helfat, 2009). Organizational ambidexterity thus is partly reliant on the ability of organizations to realign, reconfigure, and balance knowledge exploration, retention, and exploitation in a way that organizations can renew their routines when necessary (Kauppila, 2010). Enhanced decision-making processes related to the ambidexterity trade-off enables the organization to leverage existing assets and capabilities from the mature side of the business to gain competitive advantage in new areas. Leading to the following hypothesis concerning the higher order construct of internal knowledge capacity:
The higher order construct of external knowledge management capacity consists of three lower order constructs, these are: absorptive capacity, connective capacity, and desorptive capacity. External knowledge exploration is referred to as absorptive capacity. Cohen and Levinthal (1990) state that absorptive capacity is critical to innovative capabilities and that it can enhance problem-solving skills internally. Zahra and George (2002) reconceptualize the definition of absorptive capacity of Cohen and Levinthald (1990) as a dynamic capability that consists of knowledge creation and knowledge utilization which subsequently will enhances the ability of an organization to gain and sustain competitive advantage. Absorptive capacity in this framework is defined as the capacity to externally explore new knowledge (Zahra & George, 2002; Lane et al., 2006). External knowledge retention is referred to as connective capacity, meaning the ability to retain knowledge via inter-organizational relationships or alliances, but not necessarily acquiring the knowledge (Grant & Baden-Fuller, 2004; Gulati & Gargiulo, 1999). Similar to transformative capacity, knowledge that is accessible via external relationships requires active management (Grant, 1996). Kale and Singh (2007) state that an alliance learning process involves articulation, codification, sharing, and internalizing of alliance management know-how, and that the capacity to externally retain knowledge had both direct and indirect benefits. Next to the direct effect of optimization of knowledge retention, knowledge retained with alliance partners evolves over time, and thus helps organizations to extract additional knowledge from their partners (Grant, 1996). Desorptive capacity refers to knowledge exploitation outside the boundaries of the organization (Lichtenthaler, 2007). Knowledge created or retained, either internally or externally, could, next to internal exploitation, be used for external exploitation, meaning transferring knowledge outward by means of e.g. technology alliances or licensing (Lichtenthaler, 2007; Lichtenthaler &
Lichtenthaler, 2009) and therefore create additional income. But it can also provide additional benefits, e.g., enhancing internal knowledge exploitation and furthermore, it could be used to collect feedback from customers.
Similar to internal knowledge management capacities, prior related knowledge in the related field determines to a great extent, the level of absorptive, connective, and desporptive capacity (Lane et al., 2006). Therefore, the same type of dynamic capability, i.e., external knowledge management capacity, that is able to realign and reconfigure the lower order capacities, is required that enables organizations to identify and acquire new technological knowhow, creative ideas, acquire new capabilities and resources outside of the organization, which consequently enables both exploratory and exploitative innovation. In other words, ambidexterity requires internal knowledge integration and balancing mechanisms (Kauppila, 2010; Lichtenthaler, 2007). In sum, external knowledge management capacity contributes to the level of ambidexterity via enhanced decision making processes related to the ambidexterity trade-off. Leading to the following hypothesis:
H2: External knowledge management capacity is positively related to ambidexterity.
In line with Gibson and Birkinshaw (2004), this study also argues that ambidexterity partially mediates the relationship between both internal and external knowledge management capacity and organizational performance. High knowledge management capacity can contribute to organizational performance outcomes more fully, if the organization is able to leverage the knowledge through decisions that enable ambidexterity. In other words, when ambidexterity has not been developed, knowledge management capacity on its own influence performance, but knowledge management capacity will over time help organizations to become more
ambidextrous. This is in line with both Bartlett and Ghoshal (1994) and Adler et al. (1999) who show that the development of these types of dynamic capacity takes time - and subsequently enables the organization to better leverage the knowledge management capabilty. Therefore, it is proposed that ambidexterity partially mediates the relationship between internal and external knowledge management capacity and performance, leading to the following hypotheses:
H3: Ambidexterity partially mediates the relationship between internal a) and external b) knowledge management capacity and performance.
2.3 Environment
To establish a more comprehensive understanding of the relationship between knowledge management capacity and ambidexterity, this study incorporates factors stemming from outside the organizational boundary and assesses how these environmental factors moderate this relationship. Many scholars have underscored the importance of environmental factors, when researching internal processes (e.g., Burgelman and Rosenbloom, 1989; Volberda, 1996). In line with previous research that incorporated environmental factors when researching innovation, ambidexterity, and organizational performance (Levinthal & March, 1993; Gibson & Birkinshaw, 2004; Atuahene-Gima, 2005; Jansen et al., 2006; Kuenzi & Schminke, 2009) – this study proposes that environmental dynamism and competitiveness moderates the hypothesized positive impact of both internal and external knowledge management capacity on the ability of an organization to resolve the ambidexterity paradox.
Environmental dynamism is reflected through the amount and the unpredictability of change (Dess & Beard, 1984). This could be change in technology, variations in customer
preferences, or fluctuation in demand or supply of materials (Jansen, 2006). These changes could have a major impact on organizations, potentially making current products or services obsolete. Thus, organizations continuously need to be aware of what is going on in their environment, and subsequently being able to react, by for example timely invest in technologies, new business models, and the creation of product-market combinations, to prevent becoming obsolete. Consequently, this study proposes that external knowledge management capacity is more decisive in an environment characterized by dynamism, compared to internal knowledge management capacity. Absorptive capacity enables the organization to monitor potential disruptive trends emerging outside the organization; this monitoring could also be enhanced through connective capacity with alliance partners, and also via desorptive capacity getting feedback on developments in demands in technology, customer’s preferences and possible disruptive trends. Contrary, organizations with low external knowledge management capacity are more prone to missing these trends, and consequently not being able to create the right balance between incremental and disruptive innovation, this leads to the following hypotheses:
H4: Environmental dynamism negatively moderates the relationship between a) internal knowledge management capabilities and ambidexterity and positively moderates the relationship between b) external knowledge management capabilities and ambidexterity.
Environmental competitiveness is reflected through the amount of competitors and the number of areas in which there is competition (Jansen et al., 2006) with high environmental competitiveness causing a pressure for efficiency and lower prices, means less slack and lower
margins (Zahra, 1996). Thus, investing in exploratory activities becomes more difficult. Furthermore, when organizations perceive high competitiveness, they are more reluctant to share or transfer knowledge to other organizations, causing organizations to be more inward looking (Faems et al., 2007). This study therefore proposes that in an environment characterized by competitiveness, high internal knowledge management capabilities are better able to resolve the ambidexterity trade-off. Contrary, organizations with high external knowledge management capacity are less able to create the right balance between incremental and disruptive innovation, leading to the following hypotheses:
H5: Environmental competitiveness positively moderates the relationship between a) internal knowledge management capabilities and ambidexterity and negatively moderates the relationship between b) external knowledge management capabilities and ambidexterity.
In sum, knowledge management capacity is decisive in strategic decision making processes related to the ambidexterity trade-off. In other words, being able to explore, retain, and exploit knowledge from within and beyond organizational boundaries enables organizations to better allocate organizational resources to either exploration or exploitation that is beneficial to both short-term performance and long-term survival. This study tests if this proposed relationship holds or changes in different context, i.e. in environments characterized by dynamism and competitiveness, see figure 1.
3. Methods
3.1 Research design
In order to test these hypotheses a self-administered survey with closed questions was distributed via email to collect cross sectional data from multiple organizations in multiple industries. This study is part of a larger project on ambidexterity and is executed in cooperation with a fellow participant of the executive programme in Management Studies of the Amsterdam Business School. An advantage of using a self-administered online survey is the fact that a large amount of respondents can be reached in a short amount of time, which is needed to create a representative sample and minimize the possibility of history effects (Saunders & Lewis, 2012). A disadvantage of this type of research is the fact that collecting data at one point in time, does not allow dynamics that develop over time to be analyzed. The absence of interviewer assistance ensures confidentiality, but does not allow for dissolving any ambiguity that might arise when filling out the survey. Furthermore, self-administered surveys, which include both the endogenous and the exogenous constructs, are prone to common method bias (Blumberg et al., 2011). In order to reduce the possibility of this bias, different measurement scales were used for different constructs and endogenous and exogenous constructs were measured in different parts of the survey (Podsakoff et al., 2003).
3.2 Sample
The study focuses foremost on the Technology, Media, and Telecommunications industry because this industry is especially prone to display the relevant constructs (i.e. ambidexterity, environmental dynamism, and environmental competitiveness) and experience technological
disruption, as a result of earlier described trends (Boston Consulting Group, 2012). To make this study more generalizable, organizations in other industries were also invited to participate. The unit of analysis of this study is the strategic business unit. Managers of small organizations were instructed to consider their organizations as a business unit.
A business-oriented daily newspaper based in Amsterdam provided contact information of 6,764 business unit managers in The Netherlands; this included both subscribers and non-subscribers to the newspaper in question, avoiding as much as possible a bias related to the newspaper in question. The business unit managers were randomly selected.
Prior to the actual data collection, a pre-test with a small (n = 36) but representative sample, was executed to check the response rate, drop-off rate, and assess the clarity and internal consistency of the measures. For each knowledge management capacity construct, the five items with the highest standardized factor loading were included in the actual survey, reducing the number of questions to ensure participants were able to complete the survey in an acceptable amount of time. Measures related to performance, ambidexterity, environmental dynamism and competitiveness, produced satisfactory results, meaning no ambiguity and high internal consistency.
The e-mail, used to invite the 6,764 business unit mangers to participate, contained information on the purpose of the study, the time it would take to complete the survey, and the gains from participation, i.e., receiving the results and recommendations after completion of the study. Also, confidentially related to individual scores was guaranteed. The survey did not contain any routing; all participants were asked to answer the same amount of questions. Out of the 6,764 business unit managers that were invited to participate in the survey, 495 opened the survey, 211 started the survey, and 115 finished the survey. Analysis of the data revealed that of the 115 managers, 33 were reluctant to report data on the performance of their business
unit and thus were excluded. Furthermore, 10 managers were excluded because they were part of a business unit that consists of less than three full time employees, resulting in a sample size of n = 72. Based on 6,764 invites and 211 respondents starting survey, the response rate is 3.2%; based on 211 respondents starting and 72 finishing the survey the drop off rate is 65.9%. Table 2 contains descriptive statistics on the variables: firm’s size and type of industry.
3.3 Key informant check
The reported job title, presented in table 2, was used to do a key informant check, i.e., a check to see if the manager was qualified to report on the constructs under investigation in this study. All of the participants were part of the management team or able to report on the variables of interest to this study. Therefore, none of the participants were excluded as a result of this check.
3.4 Nonresponse bias
Potential nonresponse bias was assessed by analyzing the difference between the early responses, i.e., the first 75%, and the late responses, i.e., the last 25%, and the difference between those who finished to survey and those who did not (Kortmann, 2014). The t-test revealed significant differences for two items of inventive capacity. A significant difference between the scores of the early and late respondents, indicating a small nonresponse bias (Armstrong & Overton, 1977). Contrary to the first two tests, comparisons of basic organizational traits indicated no pattern of differences between respondents between organizations that participated and the organization that did not, which minimizes concern of
nonresponse bias (Ward & Duray, 2000). In conclusion, the results of this study should be approached with caution due to the fact that the results could be distorted by this nonresponse bias.
Table 2. Descriptive statistics
n = 72 % Job Title CEO 5 6.9% Chairman 3 4.2% CMO 2 2.8% Controller 1 1.4% COO 1 1.4% Director 21 29.2% Director Strategy 1 1.4%
Manager Business Development 2 2.8%
Managing director 4 5.6% Partner 1 1.4% President 1 1.4% Senior Manager 30 41.7% Firm Size* 3 ― 10 2 2.8% 11 ― 20 7 9.7% 21 ― 50 14 19.4% 51 ― 150 21 29.2% 151 ― 500 9 12.5% 501 ― 1500 11 15.3% > 1500 8 11.1% Industry** Financial Services 5 6.9% Healthcare 3 4.2% IT 38 52.8% Manufacturing 4 5.6% Media 6 8.3% Other 6 8.3% Professional Services 6 8.3% Real estate 1 1.4% Utilities 3 4.2%
CEO = chief executive officer; CMO = chief marketing officer; COO = chief operating officer. * = Number of full-time employees. ** = IT and Media are combined into a dummy variable called TMT (Technology, Media & Telecommunications), see further analysis.
3.5 Measures
The core constructs of this study were measured using both established multi-item scales that were validated in previous studies (Jansen et al., 2006; Auh & Menguc, 2005) and non-established measured that were developed especially for this study. Based on the earlier described pre-test, some measurements were adjusted to tackle ambiguity. The measurement items and associated standardized loading factors are included in table 8 of the appendices; table 9 contains the Dutch translations.
Dependent variable: performance
Performance (α = .94) was measured using two dimensions: a.) effectiveness, including profit growth, sales growth, and market growth, and b.) efficiency, including: profitability, return-on-investment, return-on-sales, and return-on-assets. Effectiveness captures long-term performance, whereas efficiency captures short-term performance (Auh & Menguc, 2005). All items were measured, asking the business unit manager to rate performance over the last three years compared to that of their main competitor, on a 5-points Likert scale (1 = much worse; 5 = much better). The business unit managers were not obliged to disclose information on performance.
Mediating variable: ambidexterity
The Ambidexterity construct was measured by using two dimensions: explorative and exploitative innovation (Jansen et al., 2009). Exploratory innovation (α = .88) and exploitative innovation (α = .87) were measured using respectively six and five items. Business unit
managers were asked to rate statements capturing explorative and exploitative innovation, on a 7-point Likert scale (1 = fully disagree; 7 = fully agree).
Independent variables: knowledge management capabilities
The organizational dynamic capability called knowledge management capacity consists of two higher order constructs, i.e., internal knowledge management capacity and external knowledge management capacity. The lower order constructs of internal knowledge management capacity: inventive (α = .89), transformative (α = .86), and innovative capacity (α = .90), were all measured with five 7-point Likert scales, asking the business unit manager to rate the extent to which their business unit contains these capabilities (1 = fully disagree; 7 = fully agree). The same was done for the lower order constructs of external knowledge management capacity: absorptive (α = .94), connective (α = .95), and desorptive capacity (α = .90). These constructs were also measured using five 7-points Likert scales. The measurement items for inventive, absorptive, transformative, connective, and innovative capacity were adopted from Lichtenthaler and Lichtenthaler (2009). The measurement items for desorptive capacity were developed for this study, based on the literature of Zahra and George (2002) and Lichtenthaler and Lichtenthaler (2009).
Moderating variables: environment
This study measured environmental dynamism and competitiveness of the organization by asking the business unit manager to indicate the extent to which he or she agrees with statements regarding these constructs on a 7-point Likert scale. Both the items on dynamism (α = .74) and competitiveness (α = .84) are measured with 5 items (Jansen et al., 2005).
Control variables
This study controlled for possible alternative explanations for ambidexterity and organizational performance by including relevant control variables, thereby isolating the effects of the constructs under investigation in this study. Two types of control variables were included: a.) characteristics of the organization and b.) characteristics of the business unit manager. The latter is only included to assess if the respondent is suitable to answer on the questions, this was done by checking the job title of the respondent. Size of the organization, measured in number of fte within the organization, is included because large organizations are better able to simultaneously implement exploration and exploitation projects due to the presence of slack resources within the organization. Furthermore, Age of the organization, measured in number of years, R&D spending, measured as a percentage of total revenue and type of industry are measured. All these control variables are in line with earlier empirical studies on ambidexterity (e.g., He & Wong, 2004; Jansen et al., 2009; Kortmann, 2014).
3.6 Validity and reliability
When using the PLS approach the reliability and validity of the measurement scales used to collect data needs to be assessed, this was done by looking at a.) individual item reliability, b.) convergent validity, and c.) discriminant validity (Hulland, 1999; Nunnally, 1978). All individual item loadings reported (appendix a) exceed a loading that is higher than .70; this indicates that all the items load on their corresponding construct. Items that reported a loading that was below .70 were excluded. All path coefficients between the higher and lower order constructs (figure 2 and 3) also show a loading that is stronger than .70, indicating the lower
order constructs load their corresponding higher order constructs as proposed. No lower order constructs were excluded from the model. To test the convergent validity of the constructs, the composite reliability (CR, table 3) and the Cronbach’s alfa (CA, table 3) were calculated. All constructs have a CR and CA above .7, which indicates good internal consistency. Discriminant validity was assessed by calculating the average variance extracted (AVE, table 3). AVE of all items exceeds .50 and its square root should be larger than the correlation coefficients with other constructs in the model. All of the constructs report outcomes consistent with these criteria (table 3). The PLS software used for this is analysis is not able to calculate the AVE of higher order constructs, hence this was done separately; internal knowledge capabilities, external knowledge capabilities, and ambidexterity all report AVE values above .50, respectively .78, .74, and .72, again indicating no problems related to the measurement scales. In sum, all scales used to collect data were both valid and reliable.
Table 3. Psychometric properties of first-order measurement scales ME SD items CA CR AVE 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. KMC1 4.03 1.60 4 .89 .92 .75 .87 2. KMC2 4.34 1.58 5 .94 .95 .80 .66*** .90 3. KMC3 5.11 1.24 2 .81 .91 .84 .60*** .58*** .92 4. KMC4 3.94 1.54 5 .95 .96 .84 .67*** .72*** .59*** .92 5. KMC5 4.95 1.40 3 .91 .94 .85 .62*** .57*** .63*** .49*** .97 6. KMC6 3.82 1.52 5 .90 .93 .72 .45*** .42*** .47*** .60*** .35*** .95 7. Exploitative innovation 5.55 1.00 5 .87 .91 .66 .48*** .37*** .58*** .41*** .58*** .24** .81 8. Exploratory innovation 4.76 1.34 6 .88 .91 .63 .54*** .32*** .51*** .35*** .57*** .40*** .56*** .79 9. Competitiveness 5.32 1.40 2 .84 .93 .86 .14 .24** .18 .29** .19 .20* .26** .36*** .93 10. Dynamism 5.04 1.29 3 .74 .85 .65 .46*** .43*** .35*** .36*** .39*** .14 .40*** .43*** .44*** .80 11. Firm performance 3.68 0.85 7 .92 .94 .68 .33*** .36*** .33*** .32*** .25** .28** .26** .31*** .07 .35*** .82 12. Firm Age 31.43 34.19 1 - - - -.06 .03 -.12 .02 -.05 -.03 .03 -.10 -.04 -.13 -.06 na 13. Firm Size 4.29 1.59 1 - - - .12 .26** .02 .20* .18 .02 .13 -.01 .18 .08 -.03 .44*** na 14. TMT (dummy) 0.61 0.49 1 - - - .13 .03 .26** .10 .22* .28 .27** .41*** .02 .21* .13 -.23* -.25** na 15. R&D 8.49 11.26 1 - - - .35*** .10 .26** .20* .29** .10 .17 .31*** -.01 .03 -.09 -.19 -.18 .32*** na
a Value on the diagonal is the square root of AVE (bold). na = not applicable. * p ≤ .10; ** p ≤ .05; *** p ≤ .01. ME = mean; SD = standard deviation; CR = composite reliability; CA = Cronbach’s alpha; AVE = average variance extracted.
KMC1 = Inventive capacity; KMC2 = Absorptive capacity; KMC3 = Transformative capacity; KMC4 = Connective capacity; KMC5 = Innovative capacity; Desorptive capacity; TMT (dummy) = Technology, Media & Telecommunications industry; R&D = Research & Development.
3.7 Confirmatory factor analysis
A confirmatory factor analysis was executed using SmartPLS, to check for cross loading of items (appendix b). Items related related the lower order construct external knowledge capabilities, ambidexterity, environmental factors, and performance showed no cross-loadings higher than .7. Some items related to internal knowledge capabilities did report cross-loadings that were higher than .7 and were removed from the model. One item from inventive capacity was removed because of cross loading with both transformative and innovative capacity that was higher than .7. One item from transformative capacity reported cross loading on innovative capacity that was higher that .7 and was removed; a second item from transformative capacity reported cross loading on inventive capacity and was also removed. Two items from innovative capacity were also removed due to cross loading on both transformative and innovative capacity that was higher than .7. In total removing six from the thirty items related to knowledge management capacity.
3.8 Common method bias
To rule out a possible common method bias as a result of using a single survey to collect both the exogenous and endogenous constructs at a single point in time, possibly creating a structural measurement errors, two post-hoc tests were conducted (Podsakoff et al., 2003). First, a Harman’s single factor test (Harman, 1967) was conducted to determine, via a factor analysis of all the items used in the survey, if there is one factor that accounts for most of the variance in data. The test shows that the first factor with an eigenvalue >1, accounts for 38.3% of the total variance, with in total 11 factors with an eigenvalue >1 accounting for 79.6% of the
total variance. This indicates no common method bias. Second, using SmartPLS, a common method factor was created and linked to all the items used to build the model. Next, the variance explained by the common method factor was compared to the variance explained by the constructs the item is intended to load (table 10). The constructs explained on average 74.3% and the method factor on average 1.5%, only 5 out the 47 item loadings of the method factor were significant compared to all the loadings of the constructs. In conclusion, both post-hoc tests do not indicate a common method bias in the data.
4. Results
To test the proposed relationships between multiple higher order constructs (figure 2 and 3) the partial least squared (PLS) approach to structural equation modeling was used. This approach is especially suited for data that is not normally distributed. Furthermore, PLS is suitable when the model contains mediation and moderation effects, as is the case in this study. The sample size of n = 72 is also in line with prior research that used this approach to test hypotheses. The SmartPLS software (Version 2.0 M3; Ringle et al., 2005) was used to build the model and test the relationships.
* p ≤ .10; ** p ≤ .05; *** p ≤ .01.
The arrows pointing towards the ambidexterity construct see figure 3
4.1 Test of hypotheses
The performance of the model is expressed in the Goodness-of-fit (GoF) statistic, which represents the overall fit of structural models. This test reveals a GoF statistic of .58, this indicates the model has a good fit and strong predictive power, because the statistic exceeds the cut-off value of .36 (Chin, 1998; Tenenhaus et al., 2005; Wetzels et al., 2009). Furthermore, this study assessed both the multiplicative score between exploration and exploitation (i.e., Explore × Exploit), and the deviation score (i.e., Explore - Exploit) as mediating variables. The multiplicative score implies that by simultaneously being able to explore and exploit, these two dimensions will enhance each other. On the other hand, the deviation score implies that by balancing or having an equal focus on exploration and exploitation, an organization is ambidextrous. The assessment of both scales indicated that only the model containing the multiplicative score for ambidexterity revealed significant betas, therefore, the model containing the deviation score was rejected.
* p ≤ .10; ** p ≤ .05; *** p ≤ .01.
Figure 3 Results of structural equation modeling, ambidexterity and performance
The hypothesized direct positive effect of internal knowledge management capacity (H1) on ambidexterity is supported (β = .78, t = 6.09, p < .01). The hypothesized direct positive effect of external knowledge management capacity (H2) on ambidexterity is rejected (β = -.15, t = 1.15, n.s.) (model 2, table 4). Both hypothesized mediation effects from internal knowledge management capacity (H3a) and external knowledge management capabilities (H3b) via ambidexterity on performance are rejected, because there is no significant effect from ambidexterity on performance (β = .17, t = .86, n.s.) (model 3, table 5). Model 2 also indicates
no direct significant relationship between internal knowledge management capacity and performance, further supporting the rejection of H3a and H3b.
The hypothesized moderation effect from environmental dynamism on the relationship between internal knowledge management capacity and ambidexterity (H4a) is rejected (β = -.24, t = 1.18, n.s.). The hypothesized moderation effect from environmental dynamism on the relationship between knowledge management external capacity and ambidexterity (H4b) is also rejected (β = .03, t =.16, n.s.). Furthermore, the hypothesized moderation effect from environmental competitiveness on the relationship between internal knowledge management capacity and ambidexterity (H5a) is rejected (β = .02, t = .11, n.s.). The hypothesized moderation effect from environmental competitiveness on the relationship between external knowledge management capacity and ambidexterity (H5b) is also rejected (β = .03, t =.15, n.s.).
Tabel 4. Test of hypotheses, model 1 and 2
Model 1 - Control variables Model 2 - Direct effects
β p-value β p-value Direct effects Internal KMC→ Performance .28 .11 External KMC → Performance .23 .11 Mediation effect Internal KMC → Ambidexterity External KMC → Ambidexterity Ambidexterity → Performance Moderation effect
Internal KMC x Dynamism → Ambidexterity External KMC x Dynamism → Ambidexterity Internal KMC x Competitiveness → Ambidexterity External KMC x Competitiveness → Ambidexterity Control variables Age – Performance -.03 .83 -.02 .83 Size → Performance .19 .27 -.13 .30 TMT → Performance .01 .98 .09 .47 R&D → Performance -.18 .35 -.29 ** .08 Age → Ambidexterity -.01 .93 Size → Ambidexterity .20 .11 TMT → Ambidexterity .38 *** .00 R&D → Ambidexterity .19 *** .01 R2 Performance .05 .23 Ambidexterity .21 * p ≤ .10; ** p ≤ .05; *** p ≤ .01. Two-tailed / One-tailed.
Tabel 5. Test of hypotheses, model 3 and 4
Model 3 - Mediation Model 4 – Full model
β p-value β p-value Direct effects Internal KMC→ Performance -.28 .30 .14 .29 External KMC → Performance .25 * .10 .25 * .09 Mediation effect Internal KMC → Ambidexterity .78 *** .00 .73 *** .00 External KMC → Ambidexterity -.15 .13 -.23 .06 Ambidexterity → Performance .17 .20 .17 .20 Moderation effect
Internal KMC x Dynamism → Ambidexterity -.24 .12
External KMC x Dynamism → Ambidexterity .03 .44
Internal KMC x Competitiveness → Ambidexterity .02 .46
External KMC x Competitiveness → Ambidexterity .03 .44
Control variables Age – Performance -.03 .37 -.03 .36 Size → Performance -.13 .15 -.13 .17 TMT → Performance .04 .38 .04 .37 R&D → Performance -.28 ** .05 -.28 ** .04 Age → Ambidexterity .06 .27 .07 .23 Size → Ambidexterity .02 .41 .01 .47 TMT → Ambidexterity .27 *** .00 .22 *** .01 R&D → Ambidexterity -.05 .27 -.02 .40 R2 Performance .24 .24 Ambidexterity .58 .68 * p ≤ .10; ** p ≤ .05; *** p ≤ .01. Two-tailed / One-tailed.
4.2 Post-hoc analysis
A post-hoc analysis was done to further examine the relationship between both internal and external knowledge management capacity and ambidexterity. The analysis determines the direct effect of knowledge management capacity on the lower order constructs of ambidexterity, i.e., exploitative and exploratory innovation. Table 6 contains three models; a.) a model where the direct effect internal knowledge management capacity on exploitative and exploratory innovation is assessed, b.) a model where the direct effect of external knowledge capacity on exploitative and exploratory innovation is assessed and c.) a model that incorporates both internal and external knowledge management capacity and where the direct effects on exploitative and exploratory innovation are assessed. The results show that, when analyzed separately, internal knowledge capacity has a significant positive direct effect on both exploitative (β = .62, t = 6.25, p < .01) and exploratory innovation (β = .57, t = 5.72, p < .01). Furthermore, the results show that external knowledge management capacity, when analyzed separately, has a significant direct effect on both exploitative (β = .36, t = 2.86, p < .01) and exploratory innovation (β = .33, t = 2.68, p < .01). In contrast to earlier findings, when analyzed together (table 6, model 3), only the direct effects of internal knowledge management on exploitative innovation (β = .74, t = 5.47, p < .01) and exploratory innovation (β = .66, t = 4.87, p < .01) remain significant. The direct effect of external knowledge management capacity on exploitative innovation (β = -.16, t = 1.17, n.s.) and exploratory innovation (β = -.12, t = .82, n.s.) disappears. In conclusion, internal and external knowledge management capacity interacts; in other words, internal knowledge management capacity alters the relationship of external knowledge management capacity of
Tabel 6. Post-hoc, KMC and Exploratory and exploitative innovation
Internal KMC External KMC Internal and external KMC
β p-value β p-value β p-value
Direct effects
Internal KMC → Exploitative innovation .62 *** .00 .74 *** .00
Internal KMC → Exploratory innovation .57 *** .00 .66 *** .00
Exteranl KMC → Exploitative innovation .36 *** .01 -.16 .25
External KMC → Exploratory innovation .33 *** .01 -.12 .41
Control variables
Age → Exploitative innovation .08 .49 .03 .83 0.08 0.41
Age → Exploratory innovation .02 .86 -.02 .88 0.02 0.83
Size → Exploitative innovation .04 .71 .11 .37 0.05 0.63
Size → Exploratory innovation -.02 .88 .04 .73 -0.01 0.93
TMT → Exploitative innovation .19 * .10 .24 ** .04 0.19 ** 0.05
TMT → Exploratory innovation .27 ** .02 .32 *** .01 0.28 *** 0.01
R&D → Exploitative innovation -.09 .36 .05 .60 -0.10 0.26
R&D → Exploratory innovation .02 .79 .16 .11 0.01 0.88
R2
Exploitative innovation 0.43 0.24 0.44
Exploratory innovation 0.48 0.32 0.49
* p ≤ .10; ** p ≤ .05; *** p ≤ .01. Two-tailed / One-tailed.
5. Conclusion & Discussion
The conclusion and discussion section contains a broader examination of the results presented in the previous section. The results of the hypothesis and overall conceptual model will be discussed and the contribution to both knowledge management capacity research and ambidexterity research will be highlighted. Furthermore, this section contains a description of the limitations and possible directions for future research.
5.1 Conclusion
Due to the robust relationship between ambidexterity and performance, the ambidexterity research has taken an important position in today's management and strategy research (O’Reilly & Tushman, 2013). In today's research, ambidexterity is positioned as a dynamic capability that mediates the relationship of several its antecedents and organizational performance (Raisch et al., 2009). The purpose of this study is to examine a novel antecedent to organizational ambidexterity, i.e., internal and external knowledge management capacity. It is widely recognized that the coexistence of both exploration and exploitation within one organization is a challenge, but not an insurmountable one. The logic behind knowledge management capacity as an antecedent to resolving the ambidexterity trade-off is that this dynamic capability is able to realign and reconfigure the existing knowledge management capabilities, both internal and external, in order to enable organizations to integrate and balance exploratory and exploitative innovations. Based on the research question, five hypotheses were introduced. The first hypothesis states that there is a positive effect from internal knowledge management capacity on ambidexterity; the analysis revealed a significant
positive effect, the hypothesis is therefore supported. Contrary to the first hypothesis, the analysis revealed no significant relationship between external knowledge management capacity, this hypothesis is therefore rejected. Together with the results of the post-hoc analysis, which further analyzed the relationships between knowledge management capacity and ambidexterity, it can be concluded that an organizations does not require both internal and external knowledge management capacity when it wants to achieve ambidexterity. Thus, the earlier notion that concerning knowledge exploration, retention, and exploitation, it is not a ‘make-or-buy’, ‘integrate-or-relate’, or a ‘keep-or-sell’ decision but that these are rather complementary (Powell et al., 1996; Rothaermel & Alexandre, 2009) is not correct. When an organization has high internal knowledge management capacity, building external knowledge management capacity will not make the organization more ambidextrous. The third hypothesis states that ambidexterity mediates the relationship between both knowledge management capacities and performance. The finding with regards to the relationship between ambidexterity and performance was unanticipated. Because of the absence of a statistically significant positive relationship no mediation by ambidexterity could take place, and therefore the third hypothesis was rejected. This is contrary to multiple other empirical studies that did find a significant relationship between ambidexterity and organizational outcomes (e.g., He & Wong, 2004). The fourth and fifth hypothesis related to moderation effects of environmental dynamism and competitiveness, were both rejected. Both environmental factors do not alter the relationship between the knowledge management capacities and ambidexterity. Therefore, these environmental factors present no boundary to the relationship between knowledge management capacity and ambidexterity.
In sum, based on the results of the hypotheses and post-hoc analysis, this study makes the cautious presumption that internal knowledge management capacity is an important
dynamic capability that enables organizations to become ambidextrous. Therefore, internal knowledge management capacity should be considered as one of the antecedents of organizational ambidexterity. Contrary to internal knowledge management capacity, this study makes the cautious presumption that external knowledge management capacity does not confer organizational ambidexterity. The results indicate that external knowledge management does not contribute to resolving the ambidexterity trade-off but that external knowledge management capacity contributes to cost savings related to knowledge management capacity or additional incomes out of existing knowledge. This can be concluded from the non-existing significant positive relationship between external knowledge management capacity and ambidexterity and the near significant direct effect of external knowledge management capacity on organizational performance.
Table 7. Result of hypotheses
Hypotheses Results Notes
H1
Internal knowledge management capacity is positively related to ambidexterity.
Accept
Interal part of the integrative model of Lichtenthaler and Lichtenthaler (2009) framed as a dynamic capability is positively associated with ambidexterity.
H2
External knowledge management capacity is positively related to ambidexterity.
Reject
Interaction with internal knowledge
managenemt causes the positive relationship to disappear, thus no additional contribution to ambidexterity from external knowledge management (Lichtenthaler & Lichtenthaler, 2009).
H3
Ambidexterity partially mediates the relationship between internal a) and external b) knowledge management capacity and performance.
Rject
Because of the unexpected absence of a positive relationship between ambidexterity and performance, mediation of ambidexterity was rejected.
H4
Environmental dynamism negatively moderates the relationship between a) internal knowledge management capabilities and ambidexterity and positively moderates the relationship between b) external knowledge management capabilities and ambidexterity.
Reject
No statistical significant relationship between environmental dynamism and the relationship between internal knowledge management and ambidexterity was found. Although there is some indication of a negative effect on the effect of internal knowledge management capacity on ambidexterity (see table 5).
H5
Environmental competitiveness positively moderates the relationship between a) internal knowledge management capabilities and ambidexterity and negatively moderates the relationship between b) external knowledge management capabilities and ambidexterity.
Reject
Contrary to other studies (e.g. Jansen, 2006) environmental competitiveness does not alter the relationship between external knowledge management capacity and ambidexterity.
5.2 Discussion
5.2.1 Theoretical implications
This study presents a first step towards uncovering the integrative model of knowledge management capacity (Lichtenthaler & Lichtenthaler, 2009) as a possible antecedent of achieving ambidexterity and organizational performance. The main contribution of this empirical study is the fact that, in particular internal knowledge management capacity contributes to organizational ambidexterity and thus should be considered as an important additional antecedent to ambidexterity in the framework proposed by Raisch et al. (2009). Contrary to what was hypothesized, external knowledge management did not reveal a positive effect on ambidexterity. Furthermore, the distinction between internal and external knowledge management made it possible to assess both capacities separately and combined into one model, this was done in the post-hoc analysis. The results of the initial model and the post-hoc analysis revealed that only internal knowledge management should be considered as an antecedent to ambidexterity, and not external knowledge management capacity. The post-hoc also revealed that internal knowledge management has a direct positive effect on both exploratory and exploitative innovation, and that that external knowledge management did not have a positive effect on either exploratory or exploitative innovation. This is contrary to what multiple authors have underlined, i.e., the complementary nature of internal and external knowledge processes in innovation processes (Hargadon & Sutton, 1997; von Hippel, 1988; Andersen & Drejer, 2008). This study, furthermore, contributes to the operationalization of the conceptual framework of Lichtenthaler and Lichtenthaler (2009) on knowledge management capacity by highlighting that the distinction between internal and external knowledge
management capacity is an important one, as these two capacities have totally different direct effects on exploratory and exploitative innovation processes.
The non-significant relationship between ambidexterity and performance in this study further underscores the notion of both Raisch et al. (2009) and Tushman and O'Reilly (2013) that this relationship under some circumstances produces ambiguous results. Hence, it can be concluded that ambidexterity does not per se functions as a mediator between the antecedents and organizational performance.
This study also contributes to understanding the boundary conditions of ambidexterity and knowledge management capacity. The absence of significant moderation effects of both environmental dynamism and competitiveness showed that these factors do not act as boundary conditions to the interplay between knowledge processes and ambidexterity. This is observation contrary to multiple empirical studies on ambidexterity and innovation processes that state that the environment does play an important role (Gibson & Birkinshaw, 2004; Jansen et al., 2005).
5.2.2 Managerial implications
As already noted in the introduction, an important dilemma managers are confronted with today is the seemingly insurmountable trade-off between exploratory and exploitative innovation activities, investment opportunities, or cultural settings. Seemingly, because by becoming ambidextrous, organizations are able to effectively surmount the trade-off. This section discusses how managers can capitalize on the theoretical insights provided in the previous section by better understanding how organizations archive performance; specifically achieve ambidexterity through building effective knowledge management capacity.
