Organizational learning in Professional Knowledge Intensive Business Services:

Organizational learning in Professional Knowledge

Intensive Business Services:

A two-case study on the portfolio of knowledge sharing mechanisms employed by P-KIBS

By

Jan Joris Kram Supervisors:

René van der Eijk & Hendrik Snijders

University of Groningen

Faculty of Economics and Business

2

Table of Content

Abstract ... 4

1.

Introduction ... 5

1.1

Innovation and Knowledge in KIBS ... 5

1.1.1

Research Objective ... 7

1.1.2

Research Boundaries ... 7

1.1.3

Research Question ... 7

1.2

Research Model ... 8

1.3

Structure of the paper ... 8

2.

Literature Review ... 9

2.1

Knowledge Intensive Business Services ... 9

2.1.1

Definition of KIBS ... 9

2.1.2

Services vs. Manufacturing ... 11

2.1.3

KIBS vs. Services ... 12

2.1.4

P-KIBS vs. T-KIBS ... 14

2.2

Theoretical Background ... 14

2.3

Organizational Learning ... 16

2.3.1

The 4I’s ... 19

2.3.2

Feedback-Feed forward ... 23

2.4

Organizational Knowledge ... 25

2.4.1

Knowledge characteristics ... 25

2.4.2

Dimensions of Knowledge ... 26

2.4.3

Knowledge sharing ... 28

2.5

Conclusion ... 33

3.

Conceptual Framework ... 35

4.

Methodology ... 36

4.1

Case study design ... 36

4.1.1 Multiple case study ... 37

3

4.1.3

Case selection ... 37

4.2

Validation and reliability ... 38

4.3

Data Collection Methods ... 39

4.5

Data analysis ... 41

4.6

Conclusion ... 41

5.

Analysis ... 42

5.1

Summary of results ... 42

5.2

Statements of interviewees ... 44

6.

Discussion ... 50

6.1

The propositions ... 50

6.2

Remarkable findings... 51

7.

Conclusion ... 54

7.1

The research question... 54

7.2

Limitations and future research ... 55

7.3

Implications and recommendations ... 55

Appendix A. Interview protocol ... 57

Appendix B. Case Data ... 58

4

Abstract

This study looks at the way professional knowledge intensive business service (P-KIBS) firms employ knowledge sharing mechanisms in order to make the best use of the knowledge which resides in the organization. This study acknowledge multiple literature bases and combines theory about organizational learning, organizational knowledge and KIBS firms in order to get an integral understanding in which managers and researchers can find answers about how P-KIBS firms actually share knowledge within the organization. The theory led to the formulation of three propositions which are illustrated through practical case study analysis. The cases existed of two successful Dutch P-KIBS organizations. The results of the study indicate that knowledge sharing mechanisms are used for creating awareness of ‘who knows what’ and or actual knowledge transfer. Different knowledge sharing mechanisms are used at an individual and group level, mechanisms that are used at an organizational level are solely used to create awareness of ‘who knows what’. The strategic importance of the choice of knowledge sharing mechanisms is supported by the case study results. The results show that the use of inter personal knowledge sharing mechanisms contributes to a faster development of the knowledge in contrast with codified knowledge sharing mechanisms. Summarized, this research shows the function that the different knowledge sharing mechanisms fulfill and with that the potential abundance of specific types of mechanisms and the lack of others can be identified.

Keywords: Organizational learning, KIBS, Knowledge sharing mechanisms, Organizational Learning

5

1. Introduction

1.1 Innovation and Knowledge in KIBS

In most industrialized nations services have become the dominant sector in the economy, for instance in the US it covers over 80% of the GDP (Goffin & Mitchell, 2005; Song & Song, 2009). Prominent explanations of this shift from manufacturing to services include the growing ability of new industrialized countries to compete on labor costs in manufacturing. Another explanation is the tendency to spend more on services with increased welfare and the fragmentation of supply chains which resulted in contracting out the service functions to other firms (Abreu, Grinevich, Kitson & Savona, 2010).

Although scholars begin to understand the importance of this dominant sector it is still the least studied and most poorly understood part of the economy (Metters & Marucheck, 2007).

It is accepted that innovation is a necessity for long-term success, growth and sustainable performance (Doyle, 1999). While a large share of innovative efforts in business is related to the development of new services (Howells, 2000), in literature services were long perceived as secondary from the viewpoint of innovation (Toivonen & Tuominen, 2009).

Within the service industries, the knowledge intensive business services sector is rapidly growing due to the unfolding of the knowledge-based economy (Den Hartog, 2000). This sector is considered to play a key role in processes of innovation by acting as ‘bridges of innovation’ between firms and science (Koch & Strotmann, 2008). It is interesting to better understand processes of innovation in KIBS firms because services such as these will account for an increasingly larger share of innovation and value creation (Bettencourt, Ostrom, Brown & Roundtree, 2002), and because it is increasingly recognized that such firms occupy a central position in modern knowledge-based economies (Freel, 2006).

6

Not only is knowledge a precondition for KIBS firms to innovate, more general many scholars suggest that a firm’s competitive advantage is increasingly depended on organizational learning and knowledge (Gagné, 2009; Riege, 2005; Von Krogh, Nonaka & Aben, 2001; Nonaka, 1994; Stata, 1989). Henderson and Clark (1992) mention that in order to innovate it is critical that managers have the ability to combine or link together knowledge within the firm. By sharing individual knowledge the value of the knowledge grows and multiplies (Cabrera &Cabrera, 2002). Davenport, De Long and Beers (1998) mention that when knowledge is shared in a systematic manner throughout the organization, the knowledge is not only accessible to a wider group, it creates synergy through sharing, reusing, discussing and re-interpreting the knowledge.

Because of the high mobility of employees it is becoming more crucial for organizations to make individual knowledge accessible for other members of the organization (Von Krogh et al., 2001) and by doing this make the organization less harmful when knowledgeable individuals leave the organization (Crossan, Lane & White, 1999).

It is generally recognized that knowledge and organizational learning are important for firms in order to innovate and to create competitive advantage. This is especially true for KIBS firms, but researchers have examined KIBS mainly from the perspective of the knowledge supplying firm (Bettencourt, Ostrom, Brown & Roundtree, 2002; Muller and Doloreux 2009) to the neglect of how organizational learning takes place within the KIBS firm. This represents a notable gap in KIBS research. This is consistent with prior research which indicated that although there is a reasonable understanding of individual and group-level learning, there is not a clear understanding of how learning flows between the levels (Crossan et al., 1999). How organizations actually manage learning is absent in the literature (Spender & Grant, 1996; Murray & Peyrefitte, 2007).

7

1.1.1 Research Objective

Because knowledge is such an important asset of a firm, the management of knowledge and intellectual assets more generally is very important. This is especially true for P-KIBS firms because the content of the service itself is to transfer knowledge to a client firm (Leiponen, 2006). The objective of this research is to better understand the internal process of learning at an organizational level in P-KIBS firms. To better understand this process this research uses theoretical literature to 1) develop a series of propositions about how P-KIBS firms employ knowledge sharing mechanisms in order to share knowledge, 2) and offer practical guidelines for these types of organizations to manage knowledge sharing within the organization.

1.1.2 Research Boundaries

This research will not investigate to what extend disseminating knowledge throughout the organization will lead to new innovations (for this research see: Lin, 2007; Leiponen, 2006; Liao, 2006). This research will also not investigate how transferring knowledge throughout the organization will result in competitive advantage. This research will also not examine the role of organizational culture within the knowledge sharing/organizational learning process. By focusing on a specific part of the process of organizational learning namely the mechanisms which are used for sharing and disseminating (individual) knowledge throughout the organization in P-KIBS firms this research enhances our understanding of organizational learning in P-KIBS firms and organizational learning in general.

1.1.3 Research Question

The research objective and research boundaries lead to the formulation of the following research question:

8

1.2 Research Model

This research is based on literature from both the areas of organizational learning and organizational knowledge and a case study performed at two P-KIBS firm. This research model shows the different stages of the research.

Figure 1. Research Model

1.3 Structure of the paper

The paper starts with a literature review in chapter one which is structured in three sections. The first section starts with a thorough description of the KIBS sector and characteristics of the KIBS sector and characteristics of P-KIBS firms in particular in order to thoroughly understand the context of this research. The second section reviews literature on the main streams of research on organizational learning and organizational knowledge. Section three zooms in on organizational knowledge by focusing on the types and characteristics of knowledge and knowledge sharing mechanisms. The literature review will result in a set of propositions which combines research on P-KIBS, organizational learning and organizational knowledge in order to provide an answer on the main research question.

9

from the empirical research the propositions are illustrated and its theoretical and practical implications are discussed.

2. Literature Review

2.1

Knowledge Intensive Business Services

In order to understand how knowledge flows between organizational levels in P-KIBS it is necessary to get a deep understanding of KIBS. This is acknowledged by Seely-Brown and Duguid (1991) who mention the importance of understanding the context when trying to understand learning.

In the next paragraph there will be taken a closer look at how the KIBS sector is defined in literature. To understand the peculiarities of KIBS firms there will be made a distinction between; service firms and manufacturing firms, service firms and KIBS firms and between P-KIBS and T-KIBS (see figure 2). At the end of paragraph 2.1 there will be given a summary of characteristics which will be used to formulate propositions about which types of mechanisms are used to share knowledge within and between organizational levels in KIBS.

Figure 2. Research Model of the literature review on the characteristics of P-KIBS

2.1.1 Definition of KIBS

The sector of KIBS comprises, among others, software firms, financial services and management consultants. In the most of the worlds’ economies this sector is one of the fastest growing (Toivonen, 2004). This sector comprises of firms which can be characterized by their knowledge intensity and the orientation of their services to other organizations (Koch & Strotmann, 2008).

P-KIBS vs. T-KIBS Services vs. Manufacturing

KIBS vs. Services

10

Bettencourt et al. 2002) define Knowledge-intensive business services (KIBS) as: ”Firms whose primary value-added activity consist of the accumulation, creation, or dissemination of knowledge for the purpose of developing a customized service or product solution to satisfy the client's needs (e.g., information technology consulting, technical engineering, software design).” Another definition is provided by Tovoinen (2006), who describe KIBS as “expert companies that provide services to other companies and organizations”.

Den Hertog (2000) suggested a more precise definition of KIBS: “Private companies or organizations that rely heavily on professional knowledge, i.e., knowledge or expertise related to a specific (technical) discipline or (technical) functional-domain to supply intermediate products and services that are knowledge based”.

These definitions comprise different elements such as: business services in which firms and public organizations are the client and not private consumers. Another element is Knowledge intensive which refers to the kind of labor which is involved in such services or the terms of the conditions in which transactions between the service provider and user take place. Knowledge intensive firms undertake operations of an intellectual nature where human capital is the dominant factor (Muller & Doloreux, 2009). In the literature a subdivision in the sector of KIBS is gradually taking root, namely technological services (t-KIBS: IT related services, engineering, R&D consulting, etc.) and professional services (P-KIBS: business and management services, legal and accounting activities, market research, etc.) (Freel, 2006; Doloreux & Muller, 2007). This research will focus on P-KIBS.

11

second view and focuses on KIBS as innovators in their own right; the role played by KIBS in their clients’ innovations, is not studied. KIBS firms are the unit of analysis and the purpose of this study is to explore how individual knowledge is shared in P-KIBS firms.

2.1.2 Services vs. Manufacturing

In existing literature the differences between services and physical products is a well covered topic. In general; services differ from physical goods because they are intangible, there is simultaneity of production and consumption and they are experienced or heterogeneous (Sirilli & Evangelista, 1998; Dolfsma, 2004), other researchers add that services are perishable (Avlonitis, Papastathoulou & Gounaris, 2001). These differences make service innovation mostly incremental with small changes in processes and procedures and easier to imitate (De Jong & Vermeulen, 2003)

Because there is a large variation in the service’s nature, in contrast with manufactured goods rolling of an assembly line which have a high degree of uniformity services are more likely to differ from one service delivery to the other (Murray, Kotabe & Westjohn, 2009).

12

service as a system of competences, technical characteristics and final characteristics. Innovation consists than as changes in one or more of these elements.

This research builds upon the synthesis approach by using a combination of old and new theories in order to study a specific element of service innovation. This is in line with the approach of Gallouj and Weinstein (1997) because this research acknowledges that innovation in KIBS results from the creation of combinations of old and new knowledge (Amara et al., 2009).

Characteristics of Services

Characteristics of manufacturing Source

Intangible service Tangible product (de Jong et al. 2003; Sirilli & Evangelista, 1998)

Simultaneity of production and consumption

asynchronous production and consumption

(Sirilli & Evangelista, 1998; Dolfsma, 2004

Experienced or heterogeneous

Products are homogenous (Murray, Kotabe &

Westjohn, 2009; Consoli & Elche-Hortelano, 2010). Perishable, not able to

stockpile services

Stockpiling is possible (Avlonitis, Papastathoulou & Gounaris, 2001

“Need pull” innovation “Technology push” innovation (Howells, 2000) Table 1. Comparison between Services and Manufacturing

2.1.3 KIBS vs. Services

KIBS firms provide products as for example advice; these products are intangible because the product that is offered has no physical substance (Sirilli & Evangelista, 1998). The KIBS firms provide knowledge based services to client firms in order to improve the client firm’s performance. To do this KIBS firms rely on experts who try to understand the unique needs of the client, this process involve non routine tacit knowledge (Camacho and Rodriguez 2005; Muller and Doloreux, 2009).

13

To enhance the performance of a client firm KIBS must cooperate with client firms, user participation in production is a widely debated and distinguishing characteristic of services and appear especially relevant for KIBS (Corrocher, Cusmano & Morrison, 2009). This cooperation with clients and uniqueness of each service delivery results that KIBS firms draw less on formal R&D than other services (Wong, & He, 2005). This also results in innovation with an ad hoc nature instead of planned, structured innovation (Sundbo, 1997).

Because the products of KIBS rely heavily on tacit knowledge and on a close interaction with client firms, KIBS are inseparable which refer to the degree to which production and consumption of a service occur in different time and space. Often close interaction between KIBS and clients is needed in order to be able to provide the right product (Murray, Kotabe & Westjohn, 2009)

Other researchers underline that KIBS’ intensive interaction with clients and their knowledge environment are not spatially neutral. As complex cognitive processes not only need codified knowledge but also rely heavily on tacit knowledge which need close interaction in order to be transferred (Corrocher, Cusmano & Morrison, 2009; de Jong et al., 2003). Another KIBS characteristic which is discovered in literature is the high training spending intensity in comparison with other services (Leiponen, 2007). This is explained easily by the fact that highly qualified human capital represents a key strategic asset (Wong, & He, 2005) because the tacit knowledge resides largely in individuals. Because of the nature of the products that KIBS firms offer and the dependence on specific individuals for different tasks KIBS firms are less structured than other types of service firms (Bettencourt et al., 2002). The characteristics that are found in the comparison between KIBS and services literature are summarized in table 2.

Characteristics of KIBS Characteristics of other services

Source

Highly customized to meet a customer’s unique needs

Less customized to meet customer’s needs

(Bettencourt et al. 2002; Camacho & Rodriguez 2005; Muller and Doloreux, 2009)

High training spending intensity

Less training spending intensity (Wong, & He, 2005; Leiponen, 2007)

Less structured More structured (Bettencourt et al.,2002)

More dependence on external knowledge

Less dependence on external knowledge

(Doloreux & Shearmur, 2009).

14

2.1.4 P-KIBS vs. T-KIBS

In the literature a subdivision in KIBS sectors is made by Miles et al. (1995), namely technological services (t-KIBS: IT related services, engineering, etc.) and professional services (P-KIBS: business and management services, legal and accounting activities, market research, etc.). A number of empirical studies support this differentiation in KIBS sectors (Muller & Zenker 2001; Freel, 2006; Zenker & Doloreux, 2008; Corrocher et al., 2009).

The pattern of specialization of P-KIBS differs from that of T-KIBS. Professionally oriented P-KIBS firms focus on client firms specific needs which entails a combination of general knowledge, problem solving and on-the-job-learning and a tailor-made solution (Bettencourt et al., 2002; Miles, 2005). In the provision of the services of T-KIBS the context is almost neutral for the supply of standardized service packages (Consoli & Elche-Hortelano, 2010). The focus on the client requirements limits replicability across assignments which are less pronounced in T-KIBS which strive at the application of repetitive routines (Consoli & Elche-Hortelano, 2010).

Characteristics of P-KIBS Characteristics of T-KIBS Source More variability across

assignments

More standardization across assignments

(Consoli & Elche-Hortelano, 2010

Table 3. Comparison between P-KIBS and T-KIBS

2.2

Theoretical Background

In literature there are four important topics which are related to this study namely: Organizational learning, the learning organization, organizational knowledge and knowledge management, which at first glance seem very similar. In order to be able to contribute and build upon a particular research stream one must get an understanding of these topics (Easterby-Smith & Lyles, 2003).

15

This distinction helps us understand the difference between the terms as described by Tsang (1997) who say that organizational learning refers to the academic study of the learning processes of and within organizations. Such studies aim primarily on the understanding of what is taking place in organizations. This makes the studies in this stream of research descriptive and targeted to academic audiences (Vera & Crossan, 2003).

The learning organization on the other hand is seen as an entity which has the capacity to learn effectively, a kind of ideal type of organization. Literature with as topic the learning organization is more practical by aiming at the understanding of how to create and improve learning capacity (Easterby-Smith & Lyles, 2003) and make this stream mainly prescriptive and targeted to managers (Vera & Crossan, 2003).

The same distinction can be made between the terms knowledge management and organizational knowledge. Literature about knowledge management generally adopts a technical approach aimed at disseminating and leveraging knowledge throughout the organization with a more practical approach. The driving force of this terminology has come from the major consultancy companies, seeking to capitalize on the enormous potential of information technology (Grint & Case, 1998). Literature about organizational knowledge tries to understand and conceptualize the nature of knowledge that is contained within organizations (Spender, 1996).

16

2.3

Organizational Learning

As one could read in the previous paragraph learning is the process of acquiring a type or degree of understanding that exist at a point in time (Chakraverthy, Mc Evily, Doz & Rau, 2003). This is a very broad definition of learning and in order to get a better understanding of the concept this paragraph will discuss several definitions and their differences and similarities.

Spender (1996) defines organizational learning as the process of experiencing and analyzing knowledge which is previously generated by others or as the process of communicating this knowledge. Dodgson (1993) argues that when individuals transfer knowledge to other members in the organization, the organization has learned. By stating this Dodgson suggests that individuals are the primary learning entity in organization.

This is in contrast with researchers who state that organizational learning is not simply the sum of each employee’s learning. (Fiol & Lyles, 1985) Although all initial learning occurs through individuals, an organization has learned when the learning is stored in organizational repositories of knowledge.

17

The conformity in previous definitions is the belief that knowledge transfer within organizations is crucial for organizational learning. For the first definitions knowledge sharing is the way in which organizations learn and the other definitions see knowledge sharing as a means to not only learn from each other but also to create new knowledge.

This research adopts the view of Levitt and March, 1988; Cohen and Levinthal, 1990; Nonaka, 1994 and Crossan et al. (1999) and sees organizational learning as a way to create new knowledge from existing stocks of knowledge by sharing it with other members of the organization.

Although the objective of this research is not to understand how the process of creating new knowledge takes place, this research takes into account the important role of knowledge sharing mechanisms in the organizational learning process. So in order to understand the choice for particular types of knowledge sharing mechanisms one must understand the purpose for which these mechanisms are the means.

This research builds upon the 4I framework organizational learning framework of Crossan et al. (1999). These scholars identify four processes (intuiting, interpreting, integrating and institutionalizing) within organizational learning which are present at three levels (individual, group and organizational levels). Different authors have identified these levels of knowledge creation in the organization (Nonaka, 1994; Crossan et al., 1999) but they do not explicitly present details about how different types of mechanisms influence the flows and stock of knowledge within and between the identified levels.

To answer this question the model of Crossan et al. (1999) will be used to get an understanding of the underlying processes which take place within organizational learning. This model is used because: 1) the authors see organizational learning as a means to be able to renew the organization. This is consistent with the underlying motivation for this research which concerns innovation in KIBS firms. 2) The authors mention that their framework should serve as a map to help researchers understand the flows of learning in an organization.

18

highlight the need to understand the mechanisms that improves the flows and stocks of learning. In the next paragraph the 4I model of Crossan et al. (1999) will be elaborated.

2.3.1 The 4I Organizational Learning Framework

Easterby- Smith, Crossan, and Nicolini (2000) reflect on organizational learning research and mention that it progressed exponentially in recent years. Beginning with the idea that learning is primarily cognitive and simply equal to the sum of learning of individuals, research has developed to recognize the concept as being socially constructed and context being placed in the learning activities of individuals, groups and organizations. According to the 4I framework, learning occurs over time and across levels (Crossan et al., 1999) it arises as a tension between assimilating new knowledge and actions (feed-forward) and exploiting what already has been learned (feedback). Throughout the feed-forward and feedback processes there is an interactive relationship between cognition and action because understanding guides action, but action also informs understanding. Organizational learning links cognition and action, which differentiates it from the fields of knowledge management and intellectual capital.

As described earlier Crossan et al. (1999) identify four socio-psychological processes (intuiting, interpreting, integrating and institutionalizing) which link learning at different levels (individual, group and organizational).

Learning is seen as a combination of stocks and flows of knowledge, even when individuals, groups, and the whole organization act as repositories of knowledge, knowledge flows across these levels in the form of feedback and feed-forward linkages through the 4I processes which are illustrated in figure 3 (Dutta & Crossan, 2005).

19

At the individual level there occur intuiting and interpreting processes, at the group level there occur interpreting and integrating processes and integrating and institutionalizing occur at the organizational level.

The framework of Crossan et al. (1999) offers three important elements of organizational learning which are taken into account in this research: 1) the framework recognizes three levels which brings together individual, group and organizational levels of analysis. 2) It shows the dynamic nature of learning by indicating the feedback and feed-forward processes. 3) it describes four different processes which occur at the different levels (the 4I’s) in the next paragraphs these elements are discussed which will contribute to the formulation of the three propositions.

2.3.1 The 4I’s

In this paragraph the 4I’s which form the basis of the framework of Crossan et al. (1999) will be elaborated and at some points additional literature is referred to further explain the four processes.

Intuiting

20

makes expertise highly subjective; deeply rooted in individual experiences; and very difficult to surface, examine and explain.

The second view on intuiting which is mentioned by Crossan et al. (1999) is entrepreneurial intuition which, in contrast to the expert view (which is past pattern oriented), is future possibility oriented. Entrepreneurial intuition refers to the ability to make novel connections and to discern possibilities. In the entrepreneurial view no two situations are the same, and patterns, while similar are never identical. This view has more to do with innovation and change, entrepreneurial intuition allow people to make novel connections, perceive new relationships that have not been identified previously.

Individual intuiting results in an inexplicable sense of the possible, of what might be done because no language exists to describe the insights or explain the intended action. And although intuition guides the actions of the individual, this intuition is difficult to share (Nonaka & Takeuchi, 1995).

Interpreting

The individual unconscious hunches, sensations or feelings which were the result of the intuiting process are made conscious through a process what Crossan et al. (1999) call interpreting. This process entails that individuals develop cognitive maps about the various domains in which they operate. Language plays a central role in the development of such maps, because it enables individuals to name and explain the unconscious outcome of the intuiting process. And once things are named, more explicit connections among the elements can be made. Furthermore Crossan et al. (1999) state that interpreting takes place in relation to a certain domain or an environment. Individuals interpret the same stimulus differently, because they have different cognitive maps. This can lead to a different or equivocal meaning for different people. Equivocality is often resolved through a group interpretive process. Just as language plays a central role in developing individual cognitive maps, it is also pivotal in developing a sense of shared understanding. The process of interpreting is a social activity that creates and refines common language, clarifies images, and creates shared understanding.

21

naturally blends into the integrating process when individual interpretive processes come together around a shared understanding and it becomes embedded within the workgroup.

Integrating

The focus of Interpreting is the change in the individual’s understanding and actions whereas integrating focuses on coherent action. Shared understanding is a condition for such coherence (Crossan et al., 1999). Through continuous conversation among members of the organization and shared practice that shared understanding develops and mutual adjustment and negotiated action takes place (Simons, 1991).

According to Crossan et al. (1999) in order for an organization to learn, its language must evolve. Because language not only helps to learn, but it also preserves what has been learned. Conversation can be used to convey the established meaning but also to evolve new meaning by conveying both the message and a deep interconnected meaning. The process of integrating is the evolution of language which extends the process of interpreting to interactions among individuals in workgroups. Actual practice is not what is specified in manuals, but it is captured and revealed in stories told by organizational members. Crossan et al. (1999) state that storytelling is a significant part of the organizational learning process, because stories reflect the complexity of actual practice rather than the abstractions taught in classrooms. By evolving stories the understanding of the phenomenon is developed which result in new integrated approaches to solving problems. Stories become the repository of knowledge instead of exclusively the individual organizational member.

Institutionalizing

The process of institutionalizing is the difference between organizational learning and individual or ad hoc group learning. Because some learning is embedded in systems, structures, strategy, routines, prescribed practices of the organization and investments in information systems and infrastructure, that what organizational members have learned individually or in groups doesn’t leave with them when they leave the organization (Crossan et al., 1999). This is consistent with the views of other scholars which also assume that organizations are more than simply the sum of the learning of its members (Nonaka & von Krogh, 2009; Nonaka, 1994; Cohen & Levinthal, 1990).

22

Because their small size, open communication and their information based on common interest, individual and group learning dominate in young organizations. As organizations mature patterns of interaction and communication between individuals begin to fall into patterns, and the organization attempt to formalize them. By doing this organizations can leverage the learning of the individual members because structures, systems and procedures provide a context for interactions.

As the prior learning becomes embedded in the organization and begins to guide the actions of organizational members the spontaneous individual and group learning become less prevalent. This causes that organizations naturally outgrow their ability to spontaneous interpreting, integrating and coherent action taking.

When routines are formalized there is a need to ensure that the routines continue to be carried out. Simons (1999) describe “diagnostic systems” with which organizations regulate the day to- day routines of the business in order to exploit the current understanding of the business. Another type of formal system is “interactive” (Simons, 1999). Organizations use these kinds of systems to consider how the future differs from the past. The process of learning becomes less fluid and incremental and becomes more abrupt the more it moves from individual to the organizational level (Crossan et al., 1999).

In general, institutionalized learning has received a certain degree of consensus or shared understanding among influential members of the organization, so when a formal organizational structure or system is changed it generally undergo a process of consideration. This will ensure that once something is institutionalized, it usually endures for a period of time (Crossan et al.,1999).

Because it takes time to institutionalize learning it cannot capture all the ongoing learning which take place at the individual and group levels.

23

2.3.2 Feedback-Feed forward

An important aspect of the 4I framework is the emphasis that learning is a dynamic process. It does not only occur over time and across different levels but it also creates a tension between what already has been learned and assimilating new learning (Crossan et al., 1999). This aspect of the framework directs the attention to the dynamic nature of organizational learning and is important to understand when one is looking at the mechanisms which are used to share knowledge. There are different flows of learning at any time at different levels. In this paragraph this tension will be explained and the implications for this research are discussed.

Crossan et al. (1999) describe how through feed-forward processes new ideas and actions flow from the individual to organizational levels. At the same moment what has been learned feeds back through group and individual levels. The simultaneous nature of these two processes creates a tension.

The feed forward process of moving from interpreting to integrating requires a shift from individual learning to learning among individuals or groups. This process entails making personal cognitive maps and integrates them in such a way that a shared understanding in a group is developed (Crossan et al., 1999). There are two main challenges in creating a new shared reality (feed-forward). The first challenge is the communication of their cognitive maps. Because many aspects of cognitive maps are tacit, this process needs to make tacit knowledge explicit which requires a process of surfacing and articulating ideas and concepts.

The second challenge arise when the cognitive maps are made explicit and can be shared there need to be the same collective interpretation of the maps. Everybody has its own cognitive map which complicates the process because they act as unique filters on the communication (Crossan et al., 1999).

The feedback process of moving from institutionalizing to intuiting can also be problematic. When there is a high degree of institutionalized learning the intuiting process requires “creative destruction” as Schumpeter (1959) calls it. This can be difficult because members of the organization ignore proven, objective methods and follow unproven experimentation (Crossan et al., 1999).

24

When learning is institutionalized in the organization it runs the risk of becoming irrelevant and it may obstruct feed-forward learning flows (Crossan et al., 1999).

An example of this is when a consultancy company have created a model of how a solution to certain type of problem can be created, when this model is out of date and an individual have other ideas of how to find a solution for such problems this individual must act against the proven shared understanding, which somebody must be willing to do and even than it can be difficult.

But although institutionalizing can create problems by inhibiting feed-forward learning it is necessary to reap the benefits of what has already been learned (Crossan et al., 1999).

The 4I framework identifies feed-forward (exploration) and feed-back (exploitation) processes which are fundamental challenges of strategic renewal (Crossan et al., 1999). Institutionalized learning may be out of date as a result of a changing environment. This creates a gap between what an organization knows what to do in a certain event and what it needs to do (Crossan et al., 1999). The process of strategic renewal is inhibited or facilitated by many factors such as reward systems, information systems, structure etc. (Crossan et al., 1999).

Looking at learning in such way raises the possibility to facilitate or constrain the learning flows by for example by changing the organizational structure which have a strong impact on who talks to who which may facilitate conversation (Crossan et al.,1999). This research attempts to understand the choice for different types of knowledge sharing mechanisms in P-KIBS organization. In trying to understand this choice this tension need to be taken into account, because the choice for particular types of knowledge sharing mechanisms can inhibit or promote new or existing knowledge to flow between different levels (Murray & Peyrefitte, 2007).

25

knowledge sharing mechanism for knowledge sharing within and one mechanism for knowledge sharing between different organizational levels which is allows for efficient knowledge sharing it is presumed that in P-KIBS firms with its specific characteristics that the processes of interpreting, integrating and institutionalizing are facilitated by different knowledge sharing mechanisms. This leads to the following proposition:

Proposition 1: P-KIBS employ different types of knowledge-sharing mechanisms at and between different organizational levels: individual, group and organization.

2.4 Organizational Knowledge

Knowledge is a commonly used word and concept in literature and everybody seems to know what is meant by it. But knowledge is more complex than one should think at first glance. As could be read in paragraph 2.3 even in literature the concepts of organizational knowledge, knowledge management and organizational learning are used interchangeably. There is no general accepted definition of knowledge and the concept is often used interchangeably with information. The consequence of this lack of understanding of knowledge is that there is also very little understanding of how organizations actually manage knowledge and knowledge sharing (Nonaka, Toyama & Komo, 2000). To avoid this pitfall and to get a deeper understanding of (organizational) knowledge not only a definition of knowledge will be given, but this paragraph also elaborates on different characteristics and dimensions of knowledge which will result in a classification of different types of knowledge sharing mechanisms which differ in terms of ‘reach’ and ‘richness’.

2.4.1 Knowledge characteristics

To better understand knowledge it is useful to look at characteristics of knowledge which many scholars have studied (Szulanski, 1996). Widely recognized characteristics of knowledge are its tacitness, the human nature and its context specificness (Nonaka, 1994; Zander & Kogut, 1995; Nonaka, Toyama, & Komo, 2000) and knowledge is causal ambiguous (Simonin, 1999).

26

knowledge. Knowledge is related to human action what provides the humanistic character (Nonaka, Toyama, & Komo, 2000). Knowledge is causal ambiguous because it is difficult to understand the linkages between input and output and actions and outcomes (Simonin, 1999).

2.4.2 Dimensions of Knowledge

In knowledge management and organizational learning literature there are two dimensions of knowledge which are widely discussed (Lam, 1997). These dimensions are: the epistemological and the ontological (Lam, 2000; Nonaka, 1994). The former concerns modes of how knowledge is expressed, namely the tacit vs. explicit distinction (Nonaka, 1994). The latter relates to the locus of knowledge which can be at different levels, namely individual vs. collective (Spender, 1996). Because this research focuses on which mechanisms are used to share knowledge throughout the organization in P-KIBS firms these dimensions of knowledge are most relevant and are elaborated in this paragraph.

Tacit and Explicit (Epistemological)

According to Lam (2000) there are three areas in which there are differences between knowledge which can be articulated explicitly or which manifest implicitly (tacit). The first and most relevant area are the mechanisms used for transferring knowledge and codifiability. Explicit knowledge can be, or is captured and shared in a medium such as a scientific formulae and specifications. It can be abstracted and stored and understood and shared without a ‘knowing subject’. Nonaka (1994) describe that explicit knowledge is expressed in a particular way. It can be, or is captured and shared in a medium such as a scientific formulae and specifications. This kind of knowledge can easily be transferred and is not context specific. Nonaka & von Krogh (2009) refer to explicit knowledge as having a “universal” character. Tacit knowledge, is contrasting in a way that it cannot be easily transferred because it is difficult to communicate and articulate. This type of knowledge is difficult to capture in a medium because tacit knowledge is an accumulation of practical skills or experiences which may not be separable from individual’s actions (Nonaka, 1994). Unlike explicit knowledge which can be formulated, abstracted and transferred independently of the knowing suspects across time and space tacit knowledge requires close interaction before it can be transferred among them (Lam, 2000).

27

and logical deduction. In contrast, the only way tacit knowledge is acquired is through practical experience in the relevant context, i.e. ‘learning-by-doing’. There are critical factors which determine the generation and accumulation which are the variety of experience and the individual’s involvement in the context (Nonaka, 1994).

The third area is the difference in their potential for aggregation and appropriation modes. Explicit knowledge does not need the knowing subject to be appropriated and can be stored everywhere in objective forms. Tacit knowledge on the other hand is personal and contextual so it needs close involvement and cooperation of the knowing subject to be aggregated and appropriated (Lam, 2000).

Nonaka (1994) argue that knowledge is created through the interaction and combination of these two types of knowledge. This means that the quality and interaction between explicit and evolving tacit knowledge types may lead to superior performance. Firms are different in the extent that they are capable of fostering such interaction and the relevant importance of the two types may also vary. But the creation of knowledge necessarily involves the use and generation of tacit knowledge (Lam, 2000).

This is why Lam (2000) states that: “The learning capability of an organization is critically dependent on its capacity to mobilize tacit knowledge and foster its interaction with explicit knowledge”.

This statement shows the importance for the organization to install the right knowledge sharing mechanisms to make sure that knowledge is mobilized and interaction takes place.

Individual and Collective (ontological)

Another widely discussed dimension of knowledge is the individual vs. collective dimension. Nonaka (1994) state that knowledge can be viewed as existing and created by individuals or in by the collective actions of a group and in relationships within groups. Lam (1997) mentions that individual knowledge moves with the person which can lead to problems in retention and accumulation. A key characteristic of individual knowledge is that an individual has autonomy in its application. Because there are cognitive limits in storing and processing information individual knowledge is always specialized and domain-specific (Lam, 2000).

28

interaction between organizational members (Lam, 1997). This kind of knowledge is largely tacit, composed of cultural norms that are a result of working together (De Long & Fahey, 2000) Collective knowledge exists between individuals rather that within them. It can be a ‘stock’ of knowledge which is stored as hard data or it represent knowledge in a state of ‘flow’ which is the result from close interaction (Lam, 2000).

Because this research examines through which mechanisms knowledge flows in KIBS firms, understanding the tacit vs. explicit and individual vs. collective dimensions of knowledge is very useful. The characteristics of the knowledge which need to be shared determine which mechanisms are most appropriate for sharing that knowledge. In the next paragraph literature about knowledge sharing will be elaborated and in combination with the dimensions and characteristics of knowledge will lead to a typology of knowledge sharing mechanisms.

2.4.3 Knowledge sharing

To maximize the competitive advantage that knowledge which resides in individuals may contain it must be shared and transferred to other members of the organization (Nonaka, 1994; Murray & Peyrefitte, 2007). Some scholars (van den Hooff & de Ridder, 2004; Boland & Tenkasi, 1995) define knowledge sharing as the process of mutually exchanging knowledge and jointly creating new knowledge in the process of reaching a common goal. This relates strongly with organizational learning which also addresses the synergistic collaboration of individuals which results in new knowledge. This is also mentioned by Spender (1996) who state that wether collective knowledge is more or less than the sum of the individual’s knowledge depends on the mechanisms that translate individual into collective knowledge. Other scholars have done studies which have demonstrated that knowledge sharing is essential for the firm because it reduces redundant learning efforts (Scarbrough, 2003; Calantone, Cavusgil & Zhao, 2002). In this paragraph there will be elaborated which factors affect knowledge sharing followed by the mechanisms which can be used to share knowledge.

29

Because this research focuses exclusively on internal knowledge sharing and on an organizational level the source, recipient and context are sufficiently described in previous chapters. The knowledge sharing mechanisms that can be used to share knowledge are discussed in the next paragraph.

Knowledge sharing mechanisms

Mechanisms through which knowledge can be shared may differ in ‘richness’ and ‘reach’. Chai, Gregory and Shi (2003) state that the richness of a medium is determined by the amount of information that could be exchanged between individuals in a amount of time (bandwidth), the degree to which the information can be selected and restructured (customization) and the degree to which repeated interaction can take place (interactivity). For example, knowledge retrieved from a first year at a new job is much richer than that acquired by reading an article (Davenport & Prusak, 2000) because it has more bandwidth; it is more customized and interactive. The richer the medium for knowledge sharing, the better the information can be customized to suit the context; it enables interactions to seek clarification and aid reinterpretation of the knowledge (Chai, Gregory & Chi, 2003; Boh, 2007). However the richer the medium for knowledge sharing the less people are able to access the knowledge, which is referred to as less ‘reach’ (Boh, 2007; Chai, Gregory & Chi, 2003; Evans & Wurster, 1997) (see figure 4).

Figure 4. Richness and Reach trade off adopted from Evans and Wurster (1997)

30

earlier. For a complete understanding of knowledge sharing mechanisms one should include the ontological dimension of knowledge.

Boh (2007) differentiate between individualized versus institutionalized mechanisms to describe a second dimension of knowledge sharing mechanisms. According to Boh (2007) Individualized mechanisms support knowledge sharing at an individual level because they have limited reach. These mechanisms tend to be random, ad hoc and informal.

Institutionalized sharing mechanisms enable the transfer of learning and knowledge from an individual to large numbers of individuals by embedding knowledge transfer capabilities in the structure and routines of an organization (Boh, 2007). These mechanisms are formal, structured and have a wider reach or are accessible by a large group of individuals.

So according to Boh (2007) there are codified and inter-personal sharing mechanisms at both the individualized and collective level which result in four types of sharing mechanisms (see figure 4).

Figure 4. Four types of knowledge transfer mechanisms, adopted from Boh (2007).

Individualized- codified sharing mechanisms

These kinds of mechanisms are used to share codified knowledge at an individual level in an ad hoc and informal manner, by for instance email, intranet, written reports, data system etc. (Kane et al., 2005; Murray & Peyrefitte, 2007). These mechanisms are used to facilitate the reuse of intellectual capital which is build by individuals by converting their experience and learning into for instance, project proposals, project plans and presentations. By effectively reusing its intellectual capital it prevents that organizations waste effort on reinventing the wheel (Boh, 2007). However codification mechanisms typically do not provide a rich medium for

Knowledge transfer mechanisms Individualized Collective Interpersonal Inter-personal transfer mechanisms Institutionalized-inter personal transfer mechanisms Codified codified transfer

mechanisms

31

knowledge transfer, because while codification may be an efficient strategy for transmitting large amounts of information, interactions and customizations of solutions to the recipient’s problems are not possible (Boh, 2007).

Individualized- inter-personal sharing mechanisms

Examples of such mechanisms are telephone, e-mail, electronic discussion, knowledge maps, face-to-face conversation, mentoring, apprenticeship, role-playing and storytelling etc. (Kane et al., 2005; Murray & Peyrefitte, 2007). Inter-personal sharing mechanisms provide a richer medium for sharing information because it involves people as mechanisms for sharing knowledge (Argote & Ingram, 2000). Olivera (2000) state that in the process of doing their work individuals generate knowledge that for a large share remain in their heads. Inter-personal sharing mechanisms have the ability to transmit tacit knowledge and allow for discussions and shared interpretations which may lead to new knowledge (Prencipe & Tell, 2001). However there are some risks and costs involved in relying on inter-personal mechanisms for sharing knowledge. Some individuals may be perceived to involve risks of admitting ignorance on a certain field (Menon & Pfeffer, 2003). And knowledge sharing through these kinds of mechanisms can only take place when the knowledge seeker knows what others know and he has access to a knowledge provider who is willing to share the knowledge (Boh, 2007).

Institutionalized- codified sharing mechanisms

32

Institutionalized-inter-personal sharing mechanisms

Examples of such mechanisms are designating specific individuals as subject-matter experts and provide access to these experts, organize work in such a way that individual experts naturally share their experience and knowledge in the process of completing the work or set up a structure such that experts are available to provide guidance to less knowledgeable individuals (Boh,2007). Hansen (1999) describe collective inter-personal sharing mechanisms such as transferring people between offices and placing knowledgeable people in close proximity of people who need to learn from that people.

That knowledge sharing based on direct contact between individuals offer many advantages because of the richness of communication is explained earlier. Instead of trying to institutionalize knowledge sharing through codification, organizations can facilitate inter personal knowledge sharing by employing mechanisms such as the above mentioned (Boh, 2007).

Because P-KIBS firms are less structured and provide highly customized products in combination with the 4I framework was the motivation for proposition 1a. The above referred to literature helps understanding which type of knowledge sharing mechanisms are employed in P-KIBS firms. The characteristics of P-KIBS firms create the need for employing ‘rich’ knowledge sharing mechanisms within and between organizational levels this in contrast with manufacturing firms and for instance T-KIBS which build upon a more predefined of knowledge because of their standardized offerings and could rely on codified sharing mechanisms. This leads to the following proposition:

Proposition 2: P-KIBS firms largely rely on individualized and institutionalized inter personal knowledge sharing mechanisms for the process of sharing knowledge.

Knowledge sharing strategy

33

and find out who is an expert on that topic. Other scholars such as Chiesa and Manzini (1996) state that different types of mechanisms are need to put in place at different stages of product development. Almeida and Grant (1998) agree with this idea and indicate that successful knowledge sharing requires multiple, complementary mechanisms. Both streams of research emphasize that organizations need the right portfolio of knowledge sharing mechanisms and that the set of knowledge sharing mechanisms is determined by the nature of their business. Emphasizing the wrong portfolio of knowledge sharing mechanisms can quickly undermine a business (Hansen et al., 1999). Scheepers, Venkitachalam and Gibbs (2004) add that the set of mechanisms on which an organization has to focus is determined by the task routines of the organization or the nature of their business.

Chai et al. (2003) indicate that there are different stages in the knowledge sharing process. The authors describe the need for creating awareness among different industrial plants about what and where knowledge exists in the organization. Who knows what in the organization is essential to know because the knowledge seeker needs to know where knowledge is residing. When awareness is created the actual knowledge transfer can take place. De Meyer (1991) suggests for instance that face-to-face meetings should be the first mechanism to create awareness in virtual R&D teams. However which knowledge sharing mechanisms at which stage are used in P-KIBS firms remains unclear. This research takes into account that the need for particular types of knowledge sharing mechanisms in the awareness and transfer stages differ. Because the type of mechanism that is needed to create awareness need to have a great ‘reach’ and does not need to be very ‘rich’. This leads to the formulation of the next proposition:

Proposition 3: There is a difference between knowledge-sharing mechanisms that P-KIBS firms use in the awareness stage and the transfer stage.

2.5 Conclusion

34

across products. These characteristics create the need for the employment of effective portfolio of knowledge sharing mechanisms but they also influence the way knowledge is shared.

35

3. Conceptual Framework

In order to answer the previously stated sub-questions and main research question, this paper connects theory on organizational learning with that of organizational knowledge in the context of P-KIBS firms. The previously elaborated theory has led to the formulation of propositions. By linking these propositions with the results of the case studies at P-KIBS firms an answer on the research question can be formulated.

The following propositions are derived from the literature review:

Proposition 1a:

P-KIBS employ different types of knowledge-sharing mechanisms at and between different organizational levels: individual, group and organization.

Proposition 1b:

Knowledge sharing mechanisms between different organizational levels are used simultaneously.

Proposition 2:

Proposition 2: P-KIBS firms largely rely on individualized and institutionalized inter personal knowledge sharing mechanisms for the process of sharing knowledge.

Proposition 3:

36

4. Methodology

In this chapter the chosen methodology for this research is described and motivated. The first paragraph of this chapter discusses the research strategy which is used to achieve the research objective. The second paragraph describes the chosen research strategy, a case study, in detail. This paragraph includes the motivation for the chosen cases, the unit of analysis and validity and reliability issues of the research.

4.1 Case study design

This research tries to understand how knowledge sharing mechanisms are employed in P-KIBS firms in order to make the best use of the knowledge which is present in the organization. So the research is aimed at phenomenon in situ, where the context of the phenomenon bears directly on the phenomenon (Davison & Hyland, 2006). Such a phenomenon is not suited to survey because the contextual nature provides a richness which is better served by data gathering from within the context. It is also not suited to experiment because testing or measuring is not required (Robson 1996). A multiple case study is chosen because of the ability to provide ‘compelling support’ for the original research propositions, or evidence for the revision of original propositions (Yin, 1994). Robson (1996) include the caution that the purpose of multiple cases is not statistical generalization as with multiple surveys but analytic generalization, where research propositions can be confirmed or disconfirmed with the use of the search for patterns of data.

Yin (2003) describes three types of case studies which each serve a different purpose: exploratory, explanatory and descriptive case studies. Exploratory case studies are used to set up a framework which serves as the basis for further a further study. “In this type of case study, fieldwork and data collection are undertaken prior to the final definition of study questions and hypotheses” (Yin, 2003, p. 6). Explanatory case studies explores why and how an event took place. The purpose of these kinds of studies is to suggest “clues to possible cause-and-effect relationships” (Yin, 2003, p.7). Descriptive case studies are used to develop document that fully illuminates the complexity of an experience and is often used to present answers to questions which are based on theoretical constructs (Yin, 2003).

37

this study the main purpose was to develop an understanding of the knowledge sharing mechanisms employed in P-KIBS firms. The second reason is that descriptive case studies answers questions based on theory. The theory presented in chapter two has led to the formulation of propositions which help to formulate an answer on the formulated research question.

4.1.1 Multiple case study

Yin (2003) states that it is preferred to conduct multiple case studies instead of a single case study, because these may provide substantial analytical benefits. The number of cases selected for this study is based on the availability of resources and the purpose of this study. Besides that it appeared to be very challenging to find companies which were interesting to research and in which people at different levels in the organization wanted to cooperate. Yin (2003) state that a two-case study with contrasting situations is sufficient for a strong start towards theoretical replication, which is sufficient for the purpose of this study.

4.1.2 Unit of analysis

An important component in the case study design is the unit of analysis (Yin, 2003). Typical units of analysis in social science include: individuals, groups, organizations and social artifacts. In this research the units of analysis are the knowledge sharing mechanisms in P-KIBS organizations. In order to gain a deeper insight of the way P-KIBS firms employ knowledge sharing mechanisms, embedded units of analysis have been included in the form of different organizational levels (Crossan et al., 1999) and different stages in the knowledge sharing process (Chai et al.,2003). Because the units of analysis are based on previous research, it enables future research to build upon this research.

4.1.3 Case selection

38

comply fully with the definition of P-KIBS, but they are not alike. The MCF is a commercial company which makes a profit by selling knowledge to other companies in the form of advice. The KI is a not-for-profit independent institute which is founded by the government and has a focus on providing knowledge to the community. The main differences between the companies are: the purpose of existence and the degree of geographically dispersion and size. Both organizations are founded in the mid thirties of the last century which make them suitable to research because they are both mature organizations. In the next paragraph the consequences of the choice of cases and other measures which are taken to increase the validity of the research are discussed.

Characteristic MCF KI

Founding year 1938 1932

Geographically dispersion 3 locations 30 locations

Focus Commercial Not-for-profit

Number of employees +/- 500 +/- 5000

Structure Matrix Matrix

Table 4. Characteristics of Case organizations

4.2

Validation and reliability

Although the main purpose of these research is not to come up with generalizable results, it is important to ensure that the validation and reliability of the research is at the highest level possible. Yin (2003) discusses several measures which can be taken in case study research to enhance its construct- and external validity and reliability. This paragraph discusses the measures which are taken in order to ensure the validity and reliability of this study.

39

from the different organizations it is essential that the interviewees agree with what the author thought of was the essence of their reactions. Every interviewee has received a draft report and had the possibility to comment on every comment, statement or other aspect which was extracted from the interviews. This also enhances the construct validity of the research because now it is clear that what has been used in the report is actually said by interviewees (Yin, 2003) External validity is the degree to which the research may be generalized to a certain domain. Although this research make a strong start towards theoretical replication by testing the propositions in two different cases which are two different organizations in the same sector, its external validity should be strengthened in the future through literal replication. Reliability is the chance that another researcher produces the same results when this case study is conducted again. Every step of the case studies are documented and by explaining the transmission of data to conclusions very clearly readers can ‘follow the derivation of any evidence’ (Yin, 2003:105), this enhances the reliability of the undertaken case studies.

4.3

Data Collection Methods

In the previous section, it became clear that the use of multiple data sources is advantageous over the use of only one source. In this paragraph the methods which are used to collect the data are elaborated. First the way by which secondary data is collected is described, followed by a description of how the interviews were prepared and conducted. Finally, a description is given of the methods used to analyze the data.

Secondary data