How do different sources of knowledge influence dynamic
capabilities?
Master thesisMSc. Business Administration -‐ Strategy track University of Amsterdam
Author: Maaike Vogel Student ID: 6039049
Supervisor: Sebastian Kortmann
Final Version submitted on 31st of January 2015
Abstract
This paper explores the commonalities of dynamic capabilities cross effective
development teams. Thereby it focuses on the role of knowledge resources. According to the latest strategy literature, firms should be able to reconfigure their resources in favour of the new market trends and developments. This is done with dynamic processes and routines defined as dynamic capabilities. Although the resource configurations resulting from dynamic capabilities are unique, the foundations for efficient dynamic capabilities seem to be quit homogenous across firms. In dynamic markets the processes that manipulate knowledge resources are especially important. Therefore, this thesis investigates commonalities regarding dynamic capabilities that (1) gain, (2) integrate and (3) reconfigure knowledge resources. The focus is on dynamic capabilities in the product development process.
By addressing this agenda, attempt is made to support current dynamic
capabilities literature with empirical evidence and connect this literature with that of specific dynamic processes.
The conceptual model proposes a direct effect of team tenure and use of explicit knowledge sources on the dependent variable innovative output. In addition, the indirect effect of implicit knowledge present in the inventor team is investigated. The hypotheses are answered using a patent database of the US mobile telecom industry.
Similar to prior research commonalities are found across firms in dynamic processes gain, integrate and reconfigure knowledge resources. The conceptual model explains a variance of the dependent variable significantly. Team tenure, gaining of explicit knowledge externally and knowledge brokering are proven to be important for
Statement of originality
This document is written by Maaike Vogel who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of contents
1. Introduction ... 5 2. Literature review ... 8 2.1. Dynamic capabilities ... 8 2.2. Knowledge resources ... 12 2.3. Gaining resources ... 132.4. Integration of resources ... 15
2.5. Reconfiguration of resources ... 17
3. Conceptual model and hypotheses ... 20
3.1. Team tenure ... 21
3.2. Explicit knowledge ... 22
3.3. Implicit knowledge ... 24
4. Data and Method ... 26
4.1. Use of patent data ... 26
4.2. Sample ... 28
4.3. Dependent variable ... 29
4.4. Independent variables ... 31
4.5. Control variables ... 32
5. Results ... 34
5.1. Direct effect of team tenure and explicit knowledge on innovative output ... 34
5.2. Mediating effect of implicit knowledge ... 37
6. Discussion ... 38
6.1. Discussion of results ... 38
6.2. Implications and future research ... 40
6.3. Limitations ... 41
7. Conclusion ... 43
1. Introduction
Throughout the years strategy scholars have been trying to find out what kind of firms outperform others and what are the sources of their competitive advantage. According to the resource-‐based view owning valuable assets and resources is essential to be competitive (Barney, 2001; Collis & Montgomery, 1999; Priem & Butler, 2001). However recent studies (Buderi, 1999; Eisenhardt & Martin, 2000; Teece, Pisano, & Shuen, 1997) claim just owning valuable resource is not enough in markets where the competitive landscape is constantly changing. Firms should be able to create and renew asset and resource configurations to gain competitive advantage. Especially in
industries where competition is fierce and technologies change rapidly, innovation has become crucial for firms to survive (Nonaka & Takeuchi, 1997).
The ability to constantly reconfigure internal and external competencies in order to be innovative is called ‘dynamic capability’. The theory by Eisenhardt and Martin (2000) states that dynamic capabilities are processes that integrate, reconfigure, gain and release resources to match or create market change. Examples of these processes are product-‐development, strategic decision-‐making and alliancing, which are mostly performed on individual or team level. These capabilities often have substantial
research streams related with them. Prior research shows effective dynamic capabilities have commonalities across effective firms. This thesis attempts to identify such
commonalities.
Dynamic capabilities are present in all shapes and sizes, however this study focuses on dynamic processes in which knowledge-‐based resources are reconfigured, gained or integrated. Knowledge-‐based resources are, more than other kinds of resources, subject to change. The value of knowledge-‐based resources decays in time
very quickly compared to other resources, like capital-‐ or location-‐based resources. Therefore, it is crucial that the knowledgebase is renewed, reconfigured and integrated in an efficient manner. Nonaka (1994) even claims that innovation is synonym for creation or recombination of knowledge.
A distinction can be made between two types of knowledge, explicit and implicit knowledge (Smith, 2001). Explicit knowledge is codified information that can be
transferred through tangible media like academic papers and manuals. With efficient dynamic capabilities, explicit knowledge sources can be combined for new purposes, also known as knowledge brokering. Implicit knowledge is captured within individuals in the form of know-‐how. Personal contact is essential for transfer of implicit
knowledge from one person to another. To manage the gaining, integration and reconfiguration of resources effectively, dynamic capabilities are needed.
This study investigates the common features among effective dynamic capabilities. Thereby, the focus is on dynamic capabilities that gain, integrate or reconfigure knowledge resources. The research question is therefore: What types of implicit and explicit knowledge sources have positive effect on dynamic
capability of a development team?
The theory of dynamic capabilities is often criticized as being non-‐specific, vague and tautological (Mosakowski & McKelvey, 1997; Priem & Butler, 2001). In this
research attempt is made to contribute to the dynamic capabilities literature by empirically supporting the statement (Eisenhardt & Martin, 2000) that dynamic capabilities are concrete, identifiable processes, rather than an abstract concept. In addition, a connection is made between the dynamic capabilities literature and the literature streams connected to the specific dynamic processes like knowledge
brokering and team organization. Thereby, common features among effective firms and project groups are identified. The outcomes of the research can be useful for further research into the commonalities among dynamic capabilities. Managers can use the information to understand and optimize their own dynamic capabilities.
In order to answer the research question, a patent database of the U.S. wireless telecom industry from 1971 till 2006 is used. The wireless telecom industry is suitable for this research because it is a high-‐velocity and high-‐tech industry. The need to innovate for firms in this industry is high; therefore the innovative activity is clearly visible in patent data. Dynamic capability is measured by the amount of forward citations a patent received. This gives a good indication of the economic value of the patent and thus the efficiency of a firm’s dynamic capability. Patent data is a widely accepted measure for innovative activity. In contrast to R&D expenditures, patent data measures the inventive output instead of its input.
In the first chapter, the different views of dynamic capabilities and several processes that use knowledge resources are discussed. Secondly, the findings from prior research are connected in the proposed conceptual model and hypotheses. In the fourth chapter the use patent data as research method is discussed and the different variables are presented. Next, the empirical tests and results will be described. The meanings of the results are presented in the discussion, as are the strengths and weaknesses of this research. Lastly, a conclusion of the entire study is given.
2. Literature review
The objective of this literature review is to give a clear idea of what dynamic capabilities are and which specific dynamic processes manipulate knowledge resources. Thereby, the dynamic capabilities literature and the literature streams of knowledge-‐based processes are linked. Through this, common features of dynamic capabilities are identified. First, the different theories of dynamic capabilities are discussed. Next, the difference between explicit and implicit knowledge is addressed. Finally, the processes of gaining, integrating and reconfiguring knowledge resources are discussed.
2.1. Dynamic capabilities
In the existing literature there are several different views on dynamic capabilities. The most important theories are discussed in this chapter. In 1997, Teece and others first introduced the dynamic capabilities as the ability to constantly reconfigure internal and external competencies in order to be innovative (Teece et al., 1997). Later, Teece (2007) continues investigating dynamic capabilities and states enterprise success depends upon the discovery and development of opportunities. In a rapidly changing
environment, gains from one single opportunity decay in time and the firm will have to come-‐up with something new. For an organization to obtain sustainable competitive advantage, innovation should become a strategy and a continuous process.
In his model of dynamic capabilities, Teece subdivides the dynamic processes into three categories; sensing (and shaping), seizing and transformational activities. Sensing is recognizing windows of opportunity or shaping your own opportunities. Sensing new opportunities is about scanning, learning and interpretation of existing
fitting service or product and commercialize it with an innovative business model. Lastly, as a firm grows through innovation it has more assets to manage.
Transformational activities are related to the effective management and reconfiguration of these assets.
Zollo & Winter (2002) claim dynamic capabilities result from several learning mechanisms. They define the concept of dynamic capabilities as “a learned and stable pattern of collective activity through which the organization systematically generates and modifies its operating routines in pursuit of improved effectiveness”. In addition to the explanation of what dynamic capabilities are, Zollo & Winter address how dynamic capabilities evolve. They state dynamic capabilities develop through three learning mechanisms namely; tacit accumulation of past experience, knowledge articulation, and knowledge codification processes. For example, a firm developing an effective joint venture based on prior experiences in cooperating with other firms, has effective dynamic capability.
This thesis will be mainly based on the dynamic capabilities theory of
Eisenhardt & Martin (2000, p.1107). Their theory originates from the resource-‐based view, in which unique assets and resources are thought to be source of competitive advantage. Eisenhardt and Martin state that in a changing market environment, a firm will not be able to sustain competitive advantage from merely owning valuable
resources. To stay competitive, a firm must constantly adapt and create unique and valuable resource configurations. They define dynamic capabilities as follows:
“The firm’s processes that use resources—specifically the processes to integrate, reconfigure, gain and release resources—to match and even create market change. Dynamic capabilities thus are the organizational and strategic routines by which firms
achieve new resource configurations as markets emerge, collide, split, evolve, and die.” Eisenhardt and Martin make three claims about dynamic capabilities. First, dynamic capabilities are identifiable processes and routines that react to, or create change in the marketplace. They categorise these processes as those that, 1) gain and release, 2) integrate and 3) reconfigure resources. Gain and release of resources is done knowledge creation processes, exit routines exemplify a release process. An example of a process that integrates resources is the product development process. Different skills and experiences are integrated to create profitable new products and services. Dynamic capabilities that focus on reconfiguration of resources are mainly knowledge-‐based processes. For example brokering of knowledge from previous inventions to combine it into new solutions.
The categories by Eisenhardt and Martin and those proposed by Teece (2007) can be related to each other. A firm should gain knowledge about their customer’s needs and wants and to sense opportunities in the marketplace. Next, it should integrate skills and experience to create new products and services to seize these opportunities. Lastly, it should constantly reconfigure its resource configurations to be innovative and stay competitive in a changing marketplace. In short, it can be said that the categories by Eisenhardt and Martin (2000) describe the process itself and Teece (2007) describes the purpose of these processes.
The second claim made by Eisenhardt and Martin is that these processes have commonalities across effective firms and project teams. In the product development process for example, teams that have joint experiences, such as working together on a task or brainstorm session, are more efficient. The actual synergies created between different team members are however different from case to case. Because the
functionality of effective capabilities can be copied across firms their value lies in the resource configurations (synergies) they create, and not in the capabilities themselves. According to this theory, dynamic capabilities are necessary for competitive advantage but having efficient dynamic capability is not a competitive advantage by itself.
Lastly, it is noticed that there are several differences in dynamic capability patterns between moderately dynamic markets and high-‐velocity markets. Moderately dynamic markets are markets with stable industry structure, clear market boundaries and players. Change in moderately dynamic markets is roughly predictable. The dynamic capabilities therefore are stable, detailed and efficient processes. They rely mostly on existing knowledge and rules of thumb.
In high-‐velocity markets, industry boundaries are blurred, players are
ambiguous and changes are unpredictable. As a consequence, dynamic capabilities are rather simple routines based on new knowledge. Technologies and trends shift quickly and processes have to adjust accordingly. Simple routines should provide just enough structure to guide people through the dynamic processes and create situation-‐specific knowledge.
In this thesis, attempt is made to support the statements made by Eisenhardt and Martin with empirical evidence. Three specific dynamic capabilities will be investigated, namely processes that integrate, gain and reconfigure knowledge resources. This
research will try to confirm that there are commonalities among effective dynamic capabilities; the goal is to discover antecedents that enable effective resource
reconfiguration, integration, and gain. The claims made regarding capabilities in high-‐ velocity markets are taken in account since the industry investigated is the high-‐velocity mobile telecom industry.
2.2. Knowledge resources
According to the resource based view (RBV) (Barney, 2001; Collis & Montgomery, 1999), valuable resources can take variety of forms; they can be tangible, like capital and a strategic manufacturing location; or intangible, like brand name or access to relevant knowledge. In this thesis the focus will be on dynamic capabilities that are integrating, reconfiguring and gaining knowledge resources. Company archives,
scientific papers, construction manuals, R&D reports and even employee experience are all knowledge-‐based resources. As stated before, dynamic capabilities are processes that update and upgrade resource configuration in order compete in changing environment. The value of knowledge-‐based resources decays in time very quickly (Nonaka & Takeuchi, 1997). Therefore, the manipulation of knowledge resources, in particular, is critical in dynamic markets, especially compared to other kinds of resources like capital. Prior research shows the ability to create gain knowledge from outside the firm as well as within the firm has substantial impact on the innovative capability of the firm (Caloghirou, Kastelli, & Tsakanikas, 2004).
To clarify what different kinds of knowledge are of influence for dynamic capabilities, a distinction is made between two types, explicit and implicit knowledge (Smith, 2001). Implicit knowledge is ‘know-‐how’ acquired by any kind of personal experience in for example university or work practice. The work practice where implicit knowledge is used is often spontaneous and responds to change. Implicit knowledge includes the skill of creative and flexible thinking, which develops new insights and eventually explicit knowledge. An important characteristic of implicit knowledge is the fact that it cannot be documented; it can only be transferred from people-‐to-‐people (Smith, 2001).
Explicit knowledge is ‘know-‐what’ described in text or other media to be
transferred to other people. For example, knowledge generated from scientific/business journals, which proved to have significant influence on innovative activity (Caloghirou et al., 2004). Explicit knowledge is often codified in organised routines and work processes. Explicit knowledge assumes a predictable environment (stable competition and customer needs) and creates knowledge objects like scientific papers or work manuals.
To continue, specific dynamic capabilities that manipulate implicit and explicit knowledge sources will be discussed. Thereby, these capabilities will be categorized as processes that gain, integrate and reconfigure resources.
2.3. Gaining resources
As mentioned before some dynamic capabilities focus on gaining resources. Gaining of knowledge resources is done in knowledge creation routines and acquisitions. These capabilities are particularly important in industries where cutting-‐edge knowledge is essential for performance. In previous research several commonalities are recognized among firms with effective knowledge gaining routines (Caloghirou et al., 2004; Hargadon, 2002; Smith, 2001).
Knowledge resources can be created in-‐house via research or can be gained externally. According to Caloghirou, Kastelli, & Tsakanikas (2004) it is equally important for upgrading innovative performance to be open towards knowledge sharing as it is to have knowledge in-‐house. Eisenhardt and Martin (2000) state that in moderately dynamic markets, dynamic capabilities are mostly dependent on existing
information, while in high-‐velocity markets they are mostly dependent on new knowledge.
Turbulent environments require firms and processes to react quickly to changes in the market place. Thereby, it cannot rely only on its own knowledge base but needs to benefit from the experience and knowledge of other economic actors (Schulz, 2001). Acquiring external knowledge requires fewer resources than internal knowledge creation. In addition, internal knowledge creation often entails managerial challenges. Hereby, the process of gaining knowledge externally is a relatively simple process that is easy to adjust to market changes. Thus, gaining knowledge externally contributes to effective dynamic capability by ensuring strategic flexibility and potential first mover advantages (Caloghirou et al., 2004).
Examples of explicit, external knowledge sources are scientific papers, conferences, customer evaluation forms, newspapers etc. These sources are mostly available to the public and can be easily gained in libraries or Internet. Implicit knowledge however can only be transferred from people-‐to-‐people it can only be
gained externally via recruitment of experts or acquisition of other firms. Unfortunately, the dataset used for this research does not contain any information about newly
recruited personal or acquisitions. Therefore, the search is limited to gaining of explicit knowledge.
Explicit knowledge can be divided in two categories, technical knowledge and market knowledge. This research focuses on sources of technical nature, because technical knowledge is the knowledge that is actually gained and transformed by dynamic capabilities in the development process. Market knowledge is used for application and commercialization of technological knowledge (Lichtenthaler, 2009).
Beside the fact that gaining knowledge externally is an easy to adapt, dynamic process: gathering information that is already out there, not only in your own domain but also outside the industry borders, can help to overcome a narrow search horizon. Overcoming a narrow search horizon is extremely important because, the more different kinds of knowledge gained the more new combinations can be made. In addition, new insights can reframe and upgrade existing knowledge routines. A common pitfall for firms, is to become prisoners of their own deeply ingrained assumptions, information filters, and problem-‐solving (Teece, 2007). Katz & Allen (1982) call this pitfall the ‘not-‐invented-‐here-‐syndrome’; a (project) group tends to believe it possesses the monopoly of knowledge in the field. Other studies describe the false belief that the same practice that led to a past success will necessary lead to a future one, as ‘the lock-‐out effect’ or ‘competency trap’(Levitt & March, 1988; Lichtenthaler, 2009).
In short, use of externally gained explicit knowledge, and especially technical knowledge, can have positive influence on dynamic capability because looking at the knowledge that is out in the field can help overcome a narrow local search horizon. Secondly, gaining knowledge externally requires fewer resources than creating it internally. In addition, it is an easier and more flexible task to manage from a
managerial point of view. Therefore, it can be easily adapted and changed in order to react to changes in the marketplace.
2.4. Integration of resources
The next category of dynamic capabilities is concerned with integration of resources. Working in a team to develop new products or ideas is such a dynamic process.
Different skills, expertise and backgrounds are integrated to make an effective development team. Assuming group members don’t have identical experience, cross fertilization of knowledge and ideas within a group makes a team more effective than an individual (Ancona & Caldwell, 1992; Horwitz & Horwitz, 2007; Katz, 1982).
The experience and skills an individual possesses can be seen as implicit
knowledge resources (Nonaka & Takeuchi, 1997). When integrating implicit knowledge in a development team, communication is very important. Implicit knowledge cannot be documented but only transferred face-‐to-‐face. Prior research claims that interpersonal communications is the primary means, rather than scientific research and other written documents, to collect and transfer important information and new ideas in development teams (Allen, 1977; Katz, 1982).
The better the individuals are connected the more efficient the development process will be (Brown & Eisenhardt, 1995). Many studies have investigated which team characteristics lead to high performance (Ancona & Caldwell, 1992; Horwitz & Horwitz, 2007; Huckman & Staats, 2011). There are two team characteristics that are said to have influence on the knowledge sharing and communication within a team, namely, team-‐size and team tenure.
Team organization stimulates the transfer of implicit knowledge from one team member to other team members (Zucker, Darby, & Armstrong, 2002). The transfer of implicit knowledge requires one-‐on-‐one communication, which becomes more difficult when the team size increases. Team size should balance between sufficiently large for individuals to be able to create mutual synergies, while not greater than sufficient to keep internal communication effective.
A characteristic that can increase internal communication and thus transfer of implicit knowledge is team tenure. Katz & Allen (1982) found that performance increased up to 16 months average tenure but that group performance gradually decayed with increasing higher levels of group stability. Familiarity can help group members to set common goals and priorities. Concrete experiences with others on the development team create a common experience base and language that facilitates communication among functionally distinct people (Eisenhardt & Martin, 2000).
Thus, communication between group members and external professional communication first increases with familiarity. However, communication declines after a certain breakpoint (Ancona & Caldwell, 1992; Huckman & Staats, 2011). After this breakpoint, the team becomes more and more isolated from external (knowledge) sources and team members become less receptive for ideas that contradict with the familiar work practice. The team becomes conservative, what is not effective for dynamic capabilities. Thus, tenure might have a two sided, curvilinear effect on group performance.
2.5. Reconfiguration of resources
Lastly, processes that reconfigure knowledge resources will be discussed. A well-‐known name for the process of reconfiguring knowledge resources is knowledge brokering (Hargadon, 2002; Schulz, 2001; Zollo & Winter, 2002). Knowledge brokering is the routine by which existing knowledge is collected from different sources and combined in new ways, resulting in synergy. The knowledge combined can be implicit or explicit. However, Reconfiguration of explicit knowledge begins with internalizing it as implicit knowledge. Thereafter, it is used to broaden, reframe the existing implicit knowledge.
Explicit knowledge sources can also directly be reconfigured to explicit knowledge, like an accountant collects financial information and puts it together in a financial report. However, this will not lead to breakthrough innovations (Nonaka & Takeuchi, 1997).
Knowledge, in particular explicit knowledge, is mostly context specific. However, with analogic thinking (finding similarities in non-‐similar things) new knowledge can be created by moving it from one context to another (Hargadon, 2002).
Explicit knowledge can be gained, copied and reconfigured from, Internet, scientific resources, and existing products and technologies. By brokering explicit knowledge it will become implicit knowledge when it becomes embedded in human behaviour. At this point it can be applied to a wider variety of local situations and thus becomes more effective for dynamic capability (Zollo & Winter, 2002).
Implicit knowledge, as mentioned before, is ‘know-‐how’ acquired by personal experience in practice. Experience can be seen as brokering of knowledge trough time. From a philosopher’s point of view: ‘experiences act upon a blank slate to imprint knowledge or wisdom like a sculptor moulds soft clay’. In this case there is no distinction between experience and knowledge (QUIŃONES, Ford, & Teachout, 1995).
It is already widely accepted that experience increases (job) performance (Ancona & Caldwell, 1992; Schmidt, Hunter, & Outerbridge, 1986). Experience leads to the acquisition of skills, techniques and methods. The primary effect of experience is accumulation of implicit knowledge that in turn leads to better performance.
Ernst et. al. (2000) call this experience effect, knowledge duplication: the amount of ideas an individual can develop depends on the amount of ideas he or she had before. When the amount of ideas grows, the possible combinations of prior ideas are
from one context in another. This means that once an inventor gains more experience or explicit knowledge, his productivity will grow exponentially. In longer term this causes a skewed distribution of performance among development teams. Only few teams will be responsible for most of the innovative output (Lotka, 1926). Lotka’s law states that the number of authors making n publications is about 1/n2 of those making one
contribution. Thus, 1/3rd of the people who make one publication make three
publications and 1/5th make five, etc. Lodka’s law also applies for productivity of
development teams. Brokering knowledge or experience will positively contribute to efficient dynamic capability. Zollo and Winter (2002) even claim that dynamic
capabilities are shaped by accumulation of experience. In their opinion routines are an outcome of trial and error learning and are a way to codify implicit knowledge.
Brokering knowledge or experience will positively contribute to efficient dynamic capability. Zollo and Winter (2002) even claim that dynamic capabilities are shaped by accumulation of experience. In their opinion routines are an outcome of trial and error learning and are a way to codify implicit knowledge.
Concluding, the reconfiguration of knowledge resources is called knowledge brokering. In this process existing knowledge, explicit or implicit, is combined in new ways through analogic thinking. Knowledge from one context is transformed and used for solutions in another context. Personal experience of inventors can be seen as implicit knowledge accumulation and therefore leads to better dynamic capabilities. Experience and gaining explicit knowledge has an exponential effect on performance whereby more knowledge or experience, leads to even more possible reconfigurations.
3. Conceptual model and hypotheses
The aim of this study is to examine the effects of implicit and explicit knowledge on dynamic capabilities. The process displayed in the conceptual model is an example of a dynamic process resembling the product development process. Innovative output is proposed as the outcome of effective dynamic capabilities. Innovation and effective dynamic capabilities are closely related. Dynamic capabilities are processes that continuously create new products, new organizations forms in reaction to changes in the market place. According to Schumpeter (Ruttan, 1959) innovation is setting up a new production function which covers new products, new organization form, mergers and more. Effective dynamic capability is thus a prerequisite for innovation. Therefore, by testing innovative output, the effectiveness of dynamic capabilities is tested.
As mentioned before the definition of dynamic capabilities used in this thesis is as follows:
“The firm’s processes that use resources—specifically the processes to integrate, reconfigure, gain and release resources—to match and even create market change. Dynamic capabilities thus are the organizational and strategic routines by which firms achieve new resource configurations as markets emerge, collide, split, evolve, and die.”
The development process is chosen because it includes three out of four types of dynamic capabilities, namely gaining, integration and reconfiguring of resources. The three processes and their mutual dependence are displayed in figure 1. Next, the antecedents and the proposed hypotheses are discussed.
Figure 1: Conceptual model
3.1. Team tenure
The first antecedent in the model is team tenure. Routines that evolve when working in a team can be seen as processes that integrate knowledge resources. These routines like brainstorming, performance evaluation and also casual lunch conversations, are expected become more effective when the team formation stays the same over a longer period of time (Katz & Allen, 1982). Individual scientist or inventors with different skills, expertise and backgrounds are put together in a team to work on a project. The longer these team members work together, the better their different specialties will create synergies. Through learning by trail and error, team members will develop standard work patterns that are familiar and comfortable and thereby result in effective knowledge sharing. For example, by sharing individual experiences and opinions, team members will create a better mutual understanding and as a result more efficient
knowledge sharing (Katz, 1982). This corresponds to the theory of Zollo and Winter (2002) that effective dynamic capabilities evolve from learning.
As written in the literature review, cross fertilization of knowledge and ideas within a group makes a team more effective than an individual. Team organization stimulates the transfer of implicit knowledge from one team member to other team members (Zucker et al., 2002). This can be seen as a process by which integration and combination of implicit knowledge leads to effective output.
H1: Team tenure will lead to efficient integration of implicit knowledge and thus innovative output.
3.2. Explicit knowledge
Members of the development team collect external explicit knowledge sources to increase and broaden its search for creative solutions. This is a process of gaining external resources. It is stated before that gaining of external knowledge sources is beneficial for effectiveness of dynamic capabilities for the following reasons. First, use of externally gained explicit knowledge, and especially scientific research, can have positive influence on dynamic capability because, looking at the knowledge that is out in the field can help overcome a narrow local search horizon. Secondly, the process of gaining external knowledge is easier to adapt to a changing environment than creating it in-‐house. This research will focus on technical knowledge sources like scientific papers, conferences and product manuals. In addition, this model only investigates the externally gained knowledge of explicit nature. Technical implicit knowledge is much
harder to gain externally, usually only via mergers or acquisitions.
There are several reasons why, especially technical knowledge of explicit nature, is expected to have positive effect on the efficiency of dynamic capabilities. Firstly, to recognize the value of new external information, process it, and commercialize the end product, a firm should have so-‐called absorptive capacity. The firm needs some
knowledge overlap with an external knowledge source to successfully gain new knowledge (Schulz, 2001). Narin, Hamilton, & Olivastro (1997) claim that, in order to have absorptive capacity, a firm should have a broad base of background knowledge. This knowledge base can be derived from scientific research, which is mostly done in universities. Caloghirou et al. (2004) compared the influence of the use of journal papers, Internet, patent databases and attendance at conferences, on innovative output. They found journal papers where most effective.
Several studies suggest that scientific research is an effective benchmark for invention and innovation (Hargadon, 2002; Zucker et al., 2002). Scientific research is mostly widely applicable since it contains generalizable knowledge. Therefore, scientific knowledge is a flexible kind of knowledge, meaning it is not context specific. The more flexible a knowledge base is the more easily it can be used to interpret new problems or invent solutions (Hargadon, 2002). The linear model of innovation suggests that
technology, through applied research originates from a scientific base (Narin et al., 1997).
In short, externally gained knowledge can have positive influence on dynamic capabilities because it can help overcome a narrow search horizon. Secondly, the process of gaining external knowledge is less complicated and thus easier to transform to a new environment. In addition, externally gained knowledge of technical and explicit
nature knowledge can have positive influence on dynamic capability because it is valuable for benchmarking and adds to a firm’s knowledge base. Scientific research in particular is a widely applicable and flexible kind of knowledge and therefore positively influences the innovation process.
H2: Gaining of external explicit knowledge resources of technical nature will lead to innovative output.
3.3. Implicit knowledge
As discussed before Recombination of knowledge, knowledge brokering, is combining existing explicit or implicit knowledge in new ways, through analogic thinking.
Knowledge from one context is transformed and used for solutions in another context. Reconfiguration of explicit knowledge begins with internalizing it as implicit
knowledge. Thereafter, use it to broaden and reframe present implicit knowledge. When implicit knowledge is reconfigured into a new solution it can be articulated as explicit knowledge (Nonaka & Takeuchi, 1997). This articulation of reconfigured knowledge can be seen as innovative output.
Schulz (2001) investigated the positive effect of amount of knowledge inflows on the amount of knowledge outflows. He stated that the more knowledge a team gained the more opportunities for brokering and thus the more outflow of knowledge. The knowledge duplication theory, which states gained knowledge or experience will grow exponentially, supports this statement. However, Schulz (2001) did not find empirical support for this hypothesis. According to his discussion, this outcome could be due to
effectiveness of knowledge brokering will be retested.
Concluding, team tenure will lead to more implicit knowledge on the team through integration of skills and reconfiguration of collective knowledge. When the team gains explicit knowledge, and effectively internalizes and reconfigures it, this will also lead to more implicit knowledge. Increase in implicit knowledge, in turn, will lead to effective dynamic capability.
Therefore, implicit knowledge is expected to be a mediator between team tenure, gaining explicit knowledge and innovative output.
H3a: The positive effect of team tenure on innovative output is mediated by an increase of implicit knowledge on the team.
H3b: The positive effect of explicit knowledge on innovative output is mediated by an increase of implicit knowledge on the team.
4. Data and Method
There are many ways in which innovative output has been measured in earlier research. Namely, R&D measured by the amount of laboratories, amount of R&D employees or R&D expenditures. However, these variables are more likely to be inputs of innovative activity than outcomes. Another way to measure innovative output is patent data. A limitation to patent count is that patents are inventions, not yet innovations, and thus do not represent correct economical value. Other works use patent citations as
indicators, which gives a better idea of the value of a patent (Clark, Gertler, & Feldman, 2003). In this study this last method will be used to indicate the level of innovative activity. In this section, the sample, strengths and limitation of patent data research, and the variable constructs are discussed.
4.1. Use of patent data
Although innovation is concept hard to measure in hard data, patents are still widely accepted indication method for innovation activity (Buderi, 1999). Especially a patent’s forward citations and amount of claims can give a good indication of the innovation’s value. Innovation is usually an incremental process with sporadic breakthroughs. This means that a product innovation will not result from one single patent but rather series of patents build on each other. Patents counts are more of an indication of the size of R&D in an industry or firm (Trajtenberg, 1990).
Using patent statistics as an indicator of innovative activity has several
limitations. First, invention is not the same as innovation. Innovation is not necessarily only technological innovation but could also be managerial or process innovation (Zollo
but for an invention to become innovation it should have practical value in a product or process.
Second, not all inventions are patentable, because they have to meet the patentability criteria set by USPTO: Novelty, non-‐obviousness, ornamentally,
enablement and definiteness. In addition, some inventions are not patented because the inventors decided to rely on secrecy or other methods to appropriate value from an invention. There is hardly any data about the amount and characteristics of inventions that are not patented(Griliches, 1998).
Third, patents differ greatly in their quality. Simple patent counts are merely an indication of the R&D expenditures of a firm. However, in a study by Griliches (1998) the non-‐used patents appear to be negligible. But even if all patented inventions would be commercialized, variance in economic significance is enormous (Trajtenberg, 1990). Therefore in this study, the forward citation will be used as a weight for importance of a patent. Lastly, patents contain limited amount of information. The problem is inherent with the use of secondary data. The research question must be drawn upon the
available data.
On the other hand there are many advantages in using patent data to indicate innovative activity. First, the data is easily accessible and comparable in a quantitative and thus objective manner. Besides, patents cover practically every field of innovation. They are rich, detailed and consistent in information. Patent data has been stored since the 18th century and thus is a continuous source, where large number of data is
available throughout time (Hall, Jaffe, & Trajtenberg, 2001). Although mere patent count is not a good indication of innovative activity, several studies (Ernst, Leptien, & Vitt, 2000; Trajtenberg, 1990) proved that patent counts weighted by citations are highly
correlated with innovations. Innovation is an incremental process, if the citations keep coming the innovation originating from this invention must be valuable. Lastly, patent data are supplied on a voluntary basis, they grant (temporary) monopoly in exchange for disclosure and therefor more reliable than other types of economic information.
4.2. Sample
The patent dataset used to answer the research question is drawn from the Delphion patent database. The dataset contains U.S. Patent and Trademark Office (USPTO) patents. The sample consist patents from the U.S. wireless telecom industry granted from 1971 till 2006. Only patents applied for from 1971-‐ 2001(20945 patents) will be used, because only those patents were granted at the time of documentation. In addition, in 2001, the industry changed with the introduction of UMTS system. These patents thus cover the four technological standards prevalent during that time period: (1) Advanced Mobile Phone System (AMPS), a 1G analog mobile phone system standard. (2) Global System for Mobile Communications (GSM) a 2G standard, (3) Time Division Multiple Access (TDMA), also a 2G digital standard used in cellular telephone
communication that divides each cellular channel into three time slots in order to increase the amount of data that can be carried, (4) Code Division Multiple Access (CDMA), a competing 2G digital standard.
The wireless telecom industry is suitable for this research, because it is a high-‐ velocity and high-‐tech industry. The need to innovate for firms in this industry is high and because the innovations are mostly related to technology, the innovative activity is clearly visible in patent data.
The dataset includes information about the amount of forward and backward references, application and publication date, applicant location (city, state and country), information about the legal representative, the inventors on the team, inventor location and the field of search. Only patents applied for in the US will be used, to control for country variation. The dataset is still fairly large, 20945 patents, which can lead to biased significant power in several statistical tests. Therefore a random sample of 5% of the total amount of patents is taken for analysis. 980 patents were randomly selected from the complete database.
4.3. Dependent variable
The dependent variable in this study is innovative output. Effective dynamic capability will lead to innovative output, thus by measuring innovative output an conclusion can be drawn about the dynamic capabilities of the firm. Patents, weighted by their forward citations, indicate innovative output. As mentioned before, patents vary greatly in their economic significance. Forward citations are proved to be a good value index for
patented inventions (Griliches, 1998; Hall et al., 2001; Trajtenberg, 1990). According to Usher, major inventions are emerging from the accumulation of several simple
inventions. Forward citations are a visible effect of these accumulations (Ruttan, 1959). If a new invention builds on technology covered by an existing patent, it can cite the patent in question to prevent claims. Thus patent citations indicate how often a certain innovation is used for further research. Because of the legal issues concerned with citing previous patents, one can assume that all but only the relevant patents are cited. As a consequence, patent citations are a good indication of the economic value of a patent (Trajtenberg, 1990). As shown in figure 2 the frequency distribution of patent