(MANAGEMENT INFORMATION SYSTEMS). Submitted ...... KMS, which refers to the degree of computerization and integration of KM processes as well as the ...
AN EMPIRICAL INVESTIGATION OF FACTORS PROMOTING KNOWLEDGE MANAGEMENT SYSTEM SUCCESS by BOBBY DALE THOMAS, JR., B.B.A. A DISSERTATION IN BUSINESS ADMINISTRATION (MANAGEMENT INFORMATION SYSTEMS) Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY Approved
Donald R. Jones Chairperson of the Committee
Karma S. Sherif
Ralph E. Viator
Peter H. Westfall
Accepted
John Borrelli Dean of the Graduate School
August, 2006
Copyright 2006, Bobby D. Thomas
ACKNOWLEDGEMENTS
I would like to express my sincerest gratitude to my family and to my committee members as well as to all of the faculty members and fellow doctoral students that have provided support and guidance, which has helped to shape this dissertation. Your contributions and support have been invaluable.
ii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS................................................................................................ ii ABSTRACT.........................................................................................................................v LIST OF TABLES............................................................................................................. vi LIST OF FIGURES .......................................................................................................... vii LIST OF ABBREVIATIONS.......................................................................................... viii CHAPTER I.
INTRODUCTION ...................................................................................................1 1.1 Purpose of the Research....................................................................................1 1.2 Contribution ......................................................................................................2 1.3 Overview of Research Model ...........................................................................4 1.4 Overview of Methodology................................................................................5 1.5 Synopsis ............................................................................................................6
II.
LITERATURE REVIEW ......................................................................................10 2.1 Knowledge Management Systems—An Overview ........................................10 2.2 Implications of Information System Success Literature.................................13 2.3 The Role of KM Strategy................................................................................18 2.4 The Impact of having a Manageable Core Set of KMS Success Factors .......21 2.5 Rationale for Selection of Factors for the KMS Success Model ....................23
III.
THEORY, MODEL, AND HYPOTHESES..........................................................31 3.1 Model Development........................................................................................31 iii
3.2 Discussion of Model and Hypotheses.............................................................33 3.2.1 Knowledge Management Strategy...................................................33 3.2.2 Top Management Leadership ..........................................................36 3.2.3 Collaboration....................................................................................38 3.2.4 Quality of Knowledge in the KMS ..................................................40 3.2.5 Compensation Schemes ...................................................................42 3.2.6 Summary..........................................................................................44 IV.
METHODOLOGY ................................................................................................46 4.1 Overview of Research.....................................................................................46 4.2 Development of Survey ..................................................................................46 4.3 Sample Selection and Data Collection............................................................48
V.
DATA ANALYSIS................................................................................................51 5.1 An Overview...................................................................................................51 5.2 Validity and Reliability...................................................................................51 5.3 Regression Analysis........................................................................................55 5.4 Path Analysis ..................................................................................................58 5.5 Alternative Analyses.......................................................................................60
VI.
CONCLUSIONS....................................................................................................62
REFERENCES ..................................................................................................................67 APPENDIX........................................................................................................................72
iv
ABSTRACT
The growing popularity of the knowledge-based theory of the firm, the view that organizational knowledge is one of the last remaining sources of long-term sustainable competitive advantage, has led to management’s growing interest in knowledge management (KM) and knowledge management systems (KMS). To date, organizations that have implemented KMS have encountered mixed results. This research contends that existing KM studies fail to give adequate consideration to the importance of KM strategies in determining critical KMS success factors. The rationale behind this research is that by properly considering the moderating effect of KM strategy on the factors that influence KMS success one can explain the success of a KMS (or lack thereof) using a greatly simplified list of success factors. This research draws on existing IS and KM frameworks, models, and literature and selects four organizational factors that are believed to be critical for the success of a KMS; this study hypothesizes which of these factors are more critical for a knowledge exploration strategy (KRS) and which of these factors are more critical for a knowledge exploitation strategy (KIS). A web-based survey utilizing existing scales, some with slight adaptations, and a newly created strategy scale was administered to test the model; 204 complete responses were collected. The results contribute to the literature by empirically confirming the hypothesized positive relationships between the identified success factors and KMS success. This research can serve as a foundation for future studies, which can help identify additional factors critical for KMS success. v
LIST OF TABLES
2.1
Comparison of Variables for Various KM Models................................................28
5.1
Results of Exploratory Factor Analysis .................................................................54
5.2
Results for Odd Hypotheses...................................................................................56
5.3
Results for Even Hypotheses .................................................................................58
5.4
Results of Path Analysis ........................................................................................59
5.5
Regression and Path Analyses Using a Five-Question Strategy Average .............61
A.1
Answer Choices .....................................................................................................79
A.2
General Statistics by Factor ...................................................................................81
A.3
Correlation Matrix Based on Responses from all 204 Respondents......................82
A.4
Strategy Averages by Organizational Structure Type ...........................................83
vi
LIST OF FIGURES
1.1
The KMS Success Model.........................................................................................6
2.1
DeLone and McLean’s (1992) I/S Success Model ................................................14
2.2
Updated DeLone and McLean (2003) IS Success Model......................................14
2.3
Jennex and Olfman’s (2003) KMS Success Model ...............................................15
2.4
Holsapple and Joshi’s (2000) Influences on the Management of Knowledge.......17
3.1
The KMS Success Model (with hypotheses) .........................................................33
4.1
The KMS Success Model (with associated scale information) .............................49
5.1
Model for Testing Odd Hypotheses with Regression Analysis .............................55
5.2
Model for Testing Even Hypotheses with Regression Analysis............................57
A.1
Invitation Letter .....................................................................................................72
A.2
Survey Login Screen..............................................................................................73
A.3
Survey Instrument..................................................................................................74
vii
LIST OF ABBREVIATIONS
Abbreviation: Meaning IS: Information System IT: Information Technology KCS: Knowledge Codification Strategy KIS: Knowledge Exploitation Strategy KM: Knowledge Management KMS: Knowledge Management System(s) KPS: Knowledge Personalization Strategy KRS: Knowledge Exploration Strategy
viii
CHAPTER I INTRODUCTION
Knowledge management systems (KMS) are “a class of information systems applied to managing organizational knowledge [by supporting and enhancing] the organizational processes of knowledge creation, storage/retrieval, transfer, and application” (Alavi and Leidner 2001 p. 114). For a system to be classified as a KMS, a chief purpose of the system must be to promote one or more of the four organizational processes just mentioned: knowledge creation, storage/retrieval, transfer, and/or application. Examples of knowledge management systems include: collaborative systems, group decision systems, data mining tools, expert systems, knowledge repositories, Intranets, electronic bulletin boards, groupware, Lotus Notes, portals, data warehouses, etc.
1.1 Purpose of the Research The purpose of this research is to study the effect of KM strategy emphasized by an organization on the blend of success factors that determines KMS success. More specifically, this research will (1) investigate the wide range of outcomes (from highly successful to highly unsuccessful) that many organizations have experienced as the result of implementing knowledge management systems; (2) examine and explain some of the mixed research findings in terms of the influence of various factors on KMS success; (3) segment firms into different categories based on the type of knowledge management 1
(KM) strategy that the firm emphasizes; and (4) empirically investigate, using the four organizational factors identified from prior literature, whether KM strategy emphasized by an organization moderates the blend of success factors that determines KMS success. This research contends that firms that report having highly successful KMS have achieved success as a result of having sufficient resources to experiment with various blends of KMS success factors until an appropriate blend developed within the organization or having luck (i.e., the factors critical to the success of the KMS were in place at the time the system was implemented).
1.2 Contribution In a 2000 study by Holsapple and Joshi, the authors start out with a literature review that yields eight factors that potentially influence KM success and conclude with a list of eighteen factors, which they divide into three categories: managerial influences, resource influences, and environmental influences. Clearly, there is a very diverse list of factors capable of influencing KMS success, and this list is by no means exhaustive. According to Alavi and Leidner (2001 p. 126), “[w]hile much theory exists on knowledge management, little empirical work has been undertaken. Hence, there are large gaps in the body of knowledge in this area.” Although numerous information system (IS) and KM articles have addressed the issue of success and have identified numerous success factors, companies that have implemented KMS have seen a wide range of outcomes from enormous savings to significant losses (Ambrosio 2000; Bose 2004; Malhotra 2005; Rigby, Reichheld, and 2
Schefter 2002). Because the ultimate success of the KMS will depend on the knowledge management processes in place as well as the organizational culture (Massey, MontoyaWeiss, and O’Driscoll 2002), this research concentrates on four key success factors that are believed to be necessary to promote successful KM processes and a supportive KM culture throughout the organization. The empirical findings of this study can confirm the positive relationships between the identified success factors and KMS success while serving as a foundation upon which future KM studies can build by examining the moderating effect of KM strategy on additional KMS success factors with the ultimate goal of this line of research being to develop two small, well-rounded sets of KMS success factors, one set of success factors for KMS emphasizing a KRS and another set for KMS emphasizing a KIS. The identification of these core sets of KMS success factors would provide practitioners with a means to more easily assess whether or not their organization has in place the conditions necessary for implementing a successful KMS emphasizing their chosen KM strategy. This information will also help practitioners to determine how best to invest in a KMS to avoid costly KMS failures. In short, the survey instrument developed in this study will provide a means by which firms may assess their strategy and their current levels of the success factors tested by the instrument. Based on the findings of this study, firms should gain initial insights into which of the tested factors are most critical for their KM strategy and hence should be focused on to improve the success of their KMS. Academicians may further extend the findings from this stream of research in order to further improve the success rates of various types of KMS. 3
In addition to providing a contribution to the IS literature, the empirical findings of this research should offer substantial savings to practitioners. By understanding the few organizational factors that are essential for a KMS emphasizing a specific KM strategy (exploration or exploitation), top management at knowledge intensive firms can better assess whether its organizational structure is sufficient for promoting a successful KMS. This information will also help companies to recognize what organizational characteristics to focus on in order to create an environment that is compatible with the type of KMS needed by the company. This technique can be used to prepare a company for a new KMS or to help a company improve an existing, unimpressive KMS.
1.3 Overview of Research Model Knowledge management systems are successful when an organization maintains a blend of organizational factors that balance managerial, resource, and environmental influences on KMS success (Holsapple and Joshi 2000). This research identifies two pairs of KMS success factors, (1) top management leadership and collaboration and (2) quality of knowledge in the KMS and compensation schemes, which balance managerial, resource, and environmental influences on KMS success and are appropriate for organizations emphasizing one of the two primary knowledge management strategies: KRS (emphasizes seeking out or creating new knowledge) and KIS (emphasizes reusing existing knowledge). In short, the KM strategy emphasized by the organization moderates the blend of KMS success factors essential for promoting KMS success. A KMS is deemed to be successful when its use has positive impacts on individual users 4
(e.g., high perceived usefulness) and the organization (e.g., high perceived organizational efficiency). In this paper, KMS success is being defined solely in terms of individual and organizational impacts (DeLone and McLean 1992). In short, a KMS that is incapable of generating positive individual and organizational impacts is consuming scarce resources (is generating costs) without generating benefits. Distinguishing the key organizational characteristics essential for a successful KMS emphasizing either a KRS or a KIS will have extensive impacts on researchers and practitioners. The proposed model, a KMS Success Model, follows (see Figure 1.1). Note that this model is not intended to, and does not, include all relevant success factors; rather, it includes a subset of the relevant success factors, which are used to test the moderating role of KM strategy emphasized.
1.4 Overview of Methodology To test the KMS Success Model, a web-based survey was administered. The survey utilized existing validated scales with minimal changes where possible. The constructs: KM Strategy Emphasized, Top Management Leadership, Collaboration, Quality of Knowledge, and Compensation Schemes are defined in section 3.2. A detailed discussion of the scales chosen along with the reasons for them being chosen can be found in section 4.2. The full-survey appears in the Appendix.
5
Figure 1.1. The KMS Success Model. The unit of analysis is KMS users. A number of statistical techniques are utilized including: factor analysis, path analysis, and moderated regression analysis. Prior to administering the survey, pre-testing is conducted as needed to refine the instrument and to ensure the reliability and validity of the scales. Validity and reliability measures are reported.
1.5 Synopsis This research draws on existing IS and KMS studies, frameworks, and models, that have identified or investigated factors that potentially influence system success (DeLone and McLean 1992, 2003; Holsapple and Joshi 2000; Jennex and Olfman 2003). This research asserts that not all of the factors identified by these studies are equally 6
influential on the success of systems in different organizations. While knowledge management systems can differ substantially from one organization to the next, they all tend to emphasize either a KRS or a KIS (Zack 1999) depending on the organizational environment and the types of knowledge with which the KMS is associated (e.g., existing knowledge or new knowledge). The theory being proposed in this research is that the KM strategy emphasized by a firm moderates the importance or influence of potential success factors on KMS success. This suggests that amid the sea of factors that potentially influence KMS success: there exists a much smaller core set of factors that are critical to the success of a KMS that emphasizes a KRS, and there is a different core set of factors that are critical to the success of a KMS that emphasizes a KIS. The purpose of this research is to empirically test the positive relationships between the proposed success factors and KMS success as well as the existence of the moderating effect of KM strategy emphasized. Assuming that KM strategy emphasized is indeed found to be a moderator, this research will serve as the foundation for future studies that can further this research by conducting additional studies to identify the smaller core sets of factors critical for the success of a KMS emphasizing either a KRS or a KIS. The end result of carrying out this line of research should be the ability for individuals to explain the success of a KMS (or lack thereof) using a greatly simplified list of success factors (i.e., a shorter list of factors that are most critical to the success of a KMS emphasizing a specific KM strategy, KIS or KRS, as opposed to having to consider all of the factors that potentially influence the success of a KMS). 7
To test this theory, four factors expected to influence KMS success are selected from the existing literature. Two of the factors, top management leadership and collaboration, are expected to be more essential for a KMS emphasizing a KRS, and the other two factors, quality of knowledge in the KMS and compensation schemes, are expected to be more essential for a KMS emphasizing a KIS. Eight hypotheses are proposed. Four hypotheses are designed to test whether a positive relationship exists between the level of each success factor and the level of KMS success—as reported by KMS users (H1, H3, H5, and H7). Two hypotheses are designed to test whether the positive relationship between top management leadership (collaboration) and KMS success is stronger for organizations following a knowledge exploration strategy than for organizations following a knowledge exploitation strategy (H2, H4). The final two hypotheses are designed to test whether the positive relationship between the quality of knowledge in the KMS (compensation schemes) and KMS success will be stronger for organizations following a knowledge exploitation strategy than for organizations following a knowledge exploration strategy (H6, H8). The KMS Success Model proposed in this research examines four factors that potentially influence KMS success rather than all of the factors that potentially influence KMS success. There are several reasons for limiting the model to four success factors. First, because the data necessary to assess the model will be collected using a web-based survey, only a limited number of questions can be asked, which in turn limits the number of factors that can be assessed. Second, the four factors included in the KMS Success Model were carefully chosen to capture all three categories of influences on KM success 8
identified by Holsapple and Joshi (2000); the four factors included in the KMS Success Model capture both managerial and resource influences with environmental influences on KMS success being captured through the KM strategy emphasized variable. These four factors were also chosen by carefully considering the factors that seemed essential to facilitating (1) knowledge exploration and (2) knowledge exploitation. Finally, given the careful selection of these four success factors, any significant moderating effect of KM strategy emphasized on the relationships between individual KMS success factors and KMS success should be detected, provided that sufficient power is achieved. Steps to achieve sufficient power will be discussed in Chapter IV. Again, keep in mind that the proposed KMS Success Model focuses only on a subset of the factors that are expected to predict KMS success—for the purpose of testing the moderating role of KM strategy emphasized. Chapter I has outlined the purpose of this line of research and the potential contributions. Chapter I has introduced a KMS success model and has proposed a research method for testing the model. Chapter II will provide a review of existing KM, KMS, and IS literature that will set the stage for Chapter III, which discusses the theory behind the model and presents testable hypotheses. The method used to test these hypotheses will be discussed in detail in Chapter IV. Chapter V will discuss data analysis issues, and Chapter VI will present conclusions drawn from this research.
9
CHAPTER II LITERATURE REVIEW
Until now, the puzzle of KMS success has been missing a few pieces; one such piece is the moderating role of KM strategy in determining critical KMS success factors. The importance of tailoring a KMS to an organization is reflected in the following quote. According to Turch (2002 p. 15), “[m]ost executives intuitively feel that ‘one size does not fit all’ when it comes to solutions and that their enterprises need to develop their own brand of KM, one which is aligned with their business needs and objectives [i.e., aligned with their strategy].”
2.1 Knowledge Management Systems—An Overview In recent years, the resource-based view of the firm, the view that firms that are able to achieve sustainable competitive advantages and/or earn superior returns are able to do so as the result of the optimal utilization of the firm’s unique mix of resources and capabilities, has been widely accepted; more recently, knowledge has come to be viewed “as the most strategically important of the firm’s resources,” which has led to an “outgrowth of the resource-based view” that has come to be known as the knowledgebased view of the firm (Grant 1996b p. 110). This new view contends that firms are able to achieve sustainable competitive advantages and/or earn superior returns as the result of the firm optimally managing its unique knowledge resources. Unlike tangible resources, which can often be easily acquired on the open market by competitors, knowledge 10
resources are more difficult to acquire. Every organization has a unique knowledge base that consists of the collective knowledge, experiences, and expertise of its employees; since every firm has different employees, no two firms have identical knowledge bases. In a competitive environment, there are at least two firms: a leading firm (i.e., the firm with the competitive advantage) and a competing firm that is trying to catch the leading firm. For the competing firm to acquire the knowledge, experience, and expertise held by a leading firm, the competing firm needs time to identify available knowledge resources, hire new employees, train existing employees, or take other measures to acquire the necessary knowledge resources. While the competing firm is trying to acquire and utilize newly acquired knowledge resources to try to minimize the leader firm’s competitive advantage, the leading firm is likely managing its knowledge to promote further innovation with the intent of increasing its competitive advantage (i.e., widening its lead), which, if successful, forces competing firms to again acquire more knowledge. In short, due to the time that it takes a competing firm to acquire similar knowledge, a leading firm that carefully manages its knowledge resources in order to promote continuous innovation is often able to stay one step ahead of competitors and sustain a competitive advantage (Nonaka 1991). Based on this view of knowledge as being the key to sustainable competitive advantage, practitioners and researchers have shown a growing interest in ways to effectively manage knowledge; in other words, both practitioners and researchers have shown a growing interest in knowledge management (Alavi and Leidner 2001; Grant 1996a, 1996b; Spender 1996a, 1996b). Upon realizing the vast base of knowledge, 11
experiences, and expertise that exists within a firm and that resides with a firm’s employees, practitioners and researchers quickly realized that even if all of the knowledge could be captured and stored on paper, finding and retrieving the desired knowledge from storage would be exceptionally tedious and in all likelihood, ineffective. However, if the same knowledge were captured in digital format, then search engines could be used to facilitate knowledge retrieval and the knowledge base could be made available to all employees regardless of their physical location. This revelation has led to a surge of interest in knowledge management systems by both researchers and practitioners. Knowledge management systems are a special class of information systems designed to facilitate KM, the leveraging of organizational knowledge through knowledge creation, knowledge storage/retrieval, knowledge transfer, and knowledge application (Alavi and Leidner 2001). Some companies, such as Ford, Chevron, and Texas Instruments estimate that their KMS have saved them millions of dollars (Bose 2004). Meanwhile recent articles report that seventy percent of all KMS fail to meet the KM objectives originally established for the system (Ambrosio 2000; Malhotra 2005; Rigby et al., 2002). A seventy percent failure rate clearly leaves room for improvement. To date, the KM literature has suggested many potential KMS success factors (e.g., Holsapple and Joshi 2000; Jennex and Olfman 2003—see discussion in Section 2.2); identifying a smaller, more manageable core set of KMS success factors should promote a more favorable KMS success rate (see discussion in Section 2.4).
12
2.2 Implications of Information System Success Literature Given that KMS are a special class of information systems and that KMS success is still a relatively new topic with little empirical support (Alavi and Leidner 2001), the extensive empirical research on, the more established topic, IS success should serve as a starting point for research on KMS success. In 1992, DeLone and McLean reported that “in searching for an I/S success measure, rather than finding none, there are nearly as many measures as there are studies” (p. 61). As indicated from DeLone and McLean’s extensive (1992) review of studies on IS success factors that cites 180 articles, of which 100 are empirical in nature, information system success is based on a delicate blend of qualities, which DeLone and McLean consolidate into a model with six factors. According to DeLone and McLean, IS success begins with system quality and information quality, which determine use and user satisfaction, which drive individual and organizational impacts; the individual and organizational impacts are viewed as the indicators of IS success (Figure 2.1). In 2003, DeLone and McLean published the Updated DeLone and McLean IS Success Model (Figure 2.2). The main differences between the 2003 and 1992 models by DeLone and McLean are (1) the addition of a service quality dimension (2003), (2) the addition of an intention to use dimension (2003), (3) the replacement of individual and organizational impacts (1992) with a broader dimension called net benefits (2003), and (4) the addition of a feedback loop from net benefits to intention to use and user satisfaction. Also in 2003, Jennex and Olfman published a KMS Success Model (Figure 2.3), which is essentially an adaptation and expansion of the Updated DeLone and 13
Figure 2.1. DeLone and McLean’s (1992) I/S Success Model.
Figure 2.2. Updated DeLone and McLean (2003) IS Success Model.
McLean (2003) IS Success Model. In Jennex and Olfman’s (2003) KMS Success Model, the system quality dimension focuses on: (1) how well the KMS is able to support knowledge creation, storage/retrieval, transfer, and application; (2) how much of the organization’s knowledge is codified; and (3) how well the KMS is supported by the IS staff as well as the organization’s infrastructure. Jennex and Olfman (2003) break down system quality into three constructs: (1) technological resources, which refers to the 14
Figure 2.3. Jennex and Olfman’s (2003) KMS Success Model.
ability of “an organization to develop, operate, and maintain a KMS”; (2) the form of the KMS, which refers to the degree of computerization and integration of KM processes as well as the accessibility of computerized knowledge; and (3) the level of the KMS, which concerns how well the KMS enables search and retrieval functions (p. 2533). Jennex and Olfman define the knowledge/information quality dimension as the extent to which the system captures and makes available the right knowledge with sufficient context to the right users at the right time. Jennex and Olfman break down the knowledge/information quality dimension into three constructs: (1) knowledge strategy/process, which examines the format and formalization of organizational processes that are in place to promote knowledge creation, storage/retrieval, transfer, and application; (2) linkages, which refers to topic maps and listings of expertise available to the organization; and (3) richness, which reflects the degree of context stored with captured knowledge as well as the accuracy and timeliness of the knowledge and its context. 15
Jennex and Olfman view service quality as a component of system quality. The other dimensions are virtually the same as those identified by DeLone and McLean (1992, 2003) and are largely self-explanatory. For example, the intent to use/perceived benefit construct tries to capture the likelihood that users will use the KMS if usage is voluntary. The use/user satisfaction construct measures use and satisfaction with the “content, accuracy, format, ease of use, and timeliness” of the outputs of the KMS (Jennex and Olfman 2003 p. 2536). Finally, the net benefits construct seeks to measure all impacts of the KMS (e.g., impacts on individuals, the organization, society, etc.). Based on the works of DeLone and McLean (1992, 2003) and Jennex and Olfman (2003), it appears very likely that the success of any type of information system (e.g., KMS) will require success on multiple dimensions before the system can be an overall success. According to Holsapple and Joshi (2000), KM success is dependent on the existence of a delicate blend of organizational factors. In a 2000 study by Holsapple and Joshi, the authors start out with a literature review that yields eight factors that potentially influence KMS success (culture, leadership, technology, organizational adjustments, evaluation of knowledge management resources/activities, governing/administering knowledge resources/activities, employee motivation, and external factors); the authors then expand several of these eight factors (e.g., culture and external factors) into more specific influences. This expansion results in a framework with fourteen factors (Figure 2.4; eighteen factors if you break down GEPSE climate) that are categorized as: managerial influences (leadership, coordination, control, and measurement), resource influences (human, knowledge, financial, and material), and environmental influences 16
(fashion, markets, competitors, time, technology, and GEPSE—governmental/economic/ political/social/educational climate). While these influences are separated into three categories, the authors suggest that any predictor or measurement of KM success
Figure 2.4. Holsapple and Joshi’s (2000) Influences on the Management of Knowledge.
must consider influences from each category of influences (managerial, resource, and environmental). In short, the framework shown in Figure 2.4 “identifies three major kinds of forces that conspire [emphasis added] to influence the knowledge management episodes that ultimately unfold in an organization” (Holsapple and Joshi 2000 p. 239).
17
2.3 The Role of KM Strategy “An analysis of KM failures reveals that many organizations who failed did not determine their goals and strategy before implementing KM systems” (Rus and Lindvall 2002 p. 34). This statement suggests that for a KMS to be successful a KM strategy should be chosen before any KMS is designed or implemented. Why should this be necessary? This paper proposes that the primary KM strategy must be selected at an early time because different blends of success factors are associated with successful KMS that emphasize different strategies. According to Zack (1999 pp. 125-6), “[m]y research with more than 25 firms has found that the most important context for guiding knowledge management is the firm’s strategy. . . [T]he link between knowledge management and business strategy, while often talked about, has been widely ignored in practice.” Zack (1999 p. 131) goes on to state that “[t]o explicate the link between [business] strategy and knowledge, an organization must articulate its strategic intent, identify the knowledge required to execute its intended strategy, and compare that to its actual knowledge, thus revealing its strategic knowledge gaps.” Once these strategic knowledge gaps are identified, the knowledge management strategy that the organization needs to emphasize becomes apparent. According to Zack (1999), if the organization is competing in a relatively stable environment, then the organization will likely be able to address the knowledge gaps by exploiting (reusing/reapplying) existing/familiar knowledge—hence, the organization emphasizes a knowledge exploitation strategy (KIS); if the organization is competing in a relatively dynamic environment, then the organization will likely be unable to address the knowledge gaps without seeking 18
newer/unfamiliar knowledge (i.e., will need to create new knowledge)—hence, the organization emphasizes a knowledge exploration strategy (KRS). In short, the environment in which the organization competes will determine the organization’s opportunities for competitive advantages, which will determine the organization’s strategic intent (i.e., business strategy), strategic knowledge gaps, and knowledge strategy. Drawing on Holsapple and Joshi (2000), it appears that the KM strategy emphasized by the organization (KIS or KRS) is determined by the environmental influences on KMS success. Holsapple and Joshi (2000) contend that “[o]rganizations have little control over environmental influences. As such, they pose constraints on an organization’s KM. However, the confluence of environmental influences can also present opportunities for improving KM.” This suggests that the KM strategy emphasized by an organization, which was determined by environmental influences, moderates the blend of KMS success factors (i.e., managerial influences and resource influences) essential for promoting KMS success. Each of the remaining factors in The KMS Success Model (top management leadership, collaboration, quality of knowledge, and compensation schemes) are a mixture of managerial and resource influences. In identifying factors for the model, two factors that are expected to promote KMS success for a system emphasizing a KRS (top management leadership and collaboration) and two factors that are expected to promote KMS success for a system emphasizing a KIS (quality of knowledge and compensation schemes) are included in the model. These four factors were chosen based on existing literature concerning KM strategies, IS success, and KMS success (see discussion in 19
Section 2.5). Definitions and support for the selection of each of these factors will be discussed in Chapter III along with the theory behind why each factor is expected to be more pertinent to the success of a KMS emphasizing a specific KM strategy. Before proceeding to Chapter III, there are two other KM strategies that should be mentioned, the knowledge codification strategy (KCS) and the knowledge personalization strategy (KPS). The KCS centers on the computer; the organization seeks to codify and store all of its knowledge in corporate databases “where it can be accessed and used easily by anyone in the company” (Hansen, Nohria, and Tierney 1999 p. 107). A KCS is typically utilized when an organization’s knowledge is largely explicit (i.e., easily codified) as opposed to being largely tacit (i.e., difficult to codify). When an organization’s knowledge is mostly tacit, a knowledge personalization strategy (KPS) is typically utilized. The KPS relies on the transfer of organizational knowledge between employees through direct person-to-person contacts with the chief purpose of computers being to communicate rather than store knowledge (Hansen et al., 1999). As with knowledge exploration and knowledge exploitation strategies, organizations typically emphasize one strategy (KCS or KPS) with the other strategy being utilized to a lesser extent. For the purpose of this research, the decision was made to focus on exploration and exploitation strategies as opposed to codification and personalization strategies for the following reasons. First, one would expect a firm that emphasizes a KIS to also emphasize a KCS because exploitation relies on knowledge storage and retrieval, which requires the knowledge to be codified (for storage to occur). However, one would likely find it more 20
difficult to guess which strategy (KCS or KPS) is emphasized by a firm that emphasizes a KRS because exploration relies on knowledge creation, which can occur with or without the direct person-to-person contact associated with a personalization strategy. Second, just because an organization’s knowledge is easily codified does not necessarily mean that the firm will choose to codify that knowledge. For example, if the organization is in a rapidly changing (explorative-type) environment the firm may not feel that it is cost effective to codify and store all knowledge that can be easily codified. Therefore, knowing the dominant type of knowledge in the firm (tacit/explicit) does not definitively indicate whether the firm will emphasize a KCS or a KPS, and knowing whether a firm emphasizes a KCS or a KPS reveals very little about the environment in which the firm operates because either could be emphasized in an environment that is predominantly explorative. Because knowing whether a firm emphasizes a KIS or a KRS reveals the general type of environment in which a firm operates, this research will focus on exploration and exploitation rather than codification and personalization.
2.4 The Impact of having a Manageable Core Set of KMS Success Factors Two core concepts in the field of economics are (1) scarcity of resources and (2) opportunity costs. Once implemented, KMS must be monitored and managed. This monitoring and managing consumes a variety of resources including time, labor, and money. In addition, there are opportunity costs associated with dedicating these resources to monitoring and managing the KMS. Any means by which success rates can be improved or time required for monitoring and managing the KMS can be reduced 21
would result in substantial savings for organizations. Holsapple and Joshi (2000) published a framework of 14-18 factors divided into three categories (managerial, resource, and environmental influences) that influence KM success (see Section 2.2). This research contends that virtually any KMS success factor can be classified as: a managerial influence, an environmental influence, a resource influence, or some combination of two or more categories of influences. If a firm wanted to assess the degree of success of its KMS and to determine what changes to make to improve the level of success of its KMS, a firm could devise a survey instrument for KMS users that would assess each of the eighteen factors identified by Holsapple and Joshi; however, these eighteen factors are not believed to be an exhaustive list of factors that are capable of influencing success. After assessing eighteen or more factors, a firm is likely to find deficiencies or room for improvement in several areas (given the low success rate of KMS—30%). How is a firm to determine which factors to dedicate its scarce resources to improving? This research contends that some of the factors are less important than others depending on the environment in which the firm operates (i.e., depending on whether the firm emphasizes a KRS or KIS). By using KM strategy emphasized as a proxy for the environmental factors, a firm now has to assess only one factor rather than the previously identified 6-10 environmental influences. This is more than a fifty percent reduction in the number of influences to be considered. Now that the environmental influences have been handled, a mix of resource and managerial influences appropriate for each strategy is needed. As discussed in Section 2.3 and in Chapter III, a review of the literature on 22
KRS and knowledge creation led to the identification of two factors that represent resource and managerial influences critical to the success of a KMS emphasizing a KRS—top management leadership and collaboration. Similarly, a review of the literature on KIS and knowledge storage/retrieval led to the identification of two factors that represent resource and managerial influences critical to the success of a KMS emphasizing a KIS—quality of knowledge in the KMS and compensation schemes. In short, this research suggests two independent variables to be assessed in addition to the dependent variable, KMS success, depending upon whether a KRS or KIS strategy is emphasized. This research also suggests that by considering the moderating role of KM strategy emphasized on KMS success factors individuals will be able to consider fewer variables to measure, predict, and/or explain KMS success in the future. With fewer factors to focus on and to adjust in order to improve the success of a KMS, it should be easier for firms to better utilize scarce resources and to improve their degree of KMS success (i.e., it is easier to monitor, manage, and allocate resources among fewer factors than for eighteen factors). If this research model could be extended (by including additional, relevant factors), empirically supported, and adopted by practitioners, researchers and practitioners alike should see an increase in the success rate of KMS.
2.5 Rationale for Selection of Factors for the KMS Success Model Holsapple and Joshi’s (2000) Influences on the Management of Knowledge (Figure 2.4) serves as a starting point for the KMS Success Model (Figure 1.1). As indicated by Holsapple and Joshi, 23
[t]he purpose of this research is to develop a comprehensive and unified framework that identifies and characterizes KM influences. Therefore, the scope and the focus is on addressing the question – ‘what are the major factors that govern KM within an organization.’ The framework does not prescribe methodologies to conduct KM effectively within an organization, nor does it attempt to measure the causal relationships between influences and outcomes . . . However, it does offer a foundation on which such investigations can be carried out. (2000 p. 255) Nortel Networks, a global leader in the telecommunications industry, recently redesigned its new product development process to capitalize on its knowledge assets; the redesign process was successful, and the time-to-market for new Nortel products was greatly reduced (Massey et al., 2002). Massey et al. (2002 p. 285) report that [t]he conceptual KM framework offered by Holsapple and Joshi served to focus our interpretation of the Nortel case along broad factors (managerial, resource, and environmental), proposed as enablers of the success of KM initiatives. Our research indicates that [our] efforts . . . represented a complex process of organizational transformation, which was, in fact, facilitated by the confluence of these factors and respective subcomponents. Thus, a key finding of our study is that successful KM initiatives like Nortel’s cannot be disentangled from broader organizational factors and changes.
Nortel Network’s experience provides further support for Holsapple and Joshi’s framework. However, as previously discussed, many firms will not have the same degree of resources available for KM as did Nortel Networks; therefore, any successful reduction of the number of potential influences (factors) that have to be monitored and managed to promote KMS success (hence, reducing the costs of monitoring and managing KM efforts) should bring KMS success within the reach of more firms (i.e., smaller firms). With this goal in mind and given the limited number of conceptual and empirical studies on KMS success, additional guidance was sought from the more established field, 24
IS success. This led to the consideration of the DeLone and McLean IS Success Models (1992, 2003) as well as the consideration of an extension of their model, which specifically addressed KMS success (Jennex and Olfman 2003)—see Figures 2.1-2.3. In examining these three models, there are several similarities: (1) success has its origin with quality (system quality, knowledge/information quality, and service quality); (2) quality determines use and user satisfaction; and (3) quality determines impacts on individuals, organizations, and other parties (collectively referred to as net benefits), which serve as a proxy for system success. While net benefits can influence future use and user satisfaction, initial use and user satisfaction are established as a result of quality. The potential for the realization of net benefits is established by the various dimensions of quality (i.e., the lower the quality, the lower the potential for net benefits); the degree of net benefits realized should directly and proportionately impact user satisfaction and future intentions for use. In short, lower quality is expected to yield a lower potential for net benefits, which, in turn, adversely affects the potential for user satisfaction and future intentions for use. Given this relationship (i.e., use and satisfaction dimensions are determined directly by quality dimensions and indirectly by net benefits), the KMS Success Model proposed in this research consists of variables for assessing quality and impacts (benefits) rather than variables for directly assessing use, user satisfaction, and intentions to use. Furthermore, the KMS Success Model in this research measures success by assessing only individual and organizational impacts as opposed to net benefits, which is a broader set of impacts; the reason for this decision is that over time individual impacts will lead to 25
organizational impacts, so a KMS that generates positive individual and/or organizational impacts should have a good chance of being or becoming successful. Another reason for not using net benefits is that beyond individual and organizational impacts, the other impacts that comprise net benefits are impacts on parties outside of the organization. Getting back to Holsapple and Joshi’s (2000) framework and drawing on the discussion in Section 2.3, KM strategy is deemed to encapsulate the environmental factors in the framework. Now consider the remaining variables (quality dimensions) to be captured in the KMS Success Model. What determines quality? This research contends that quality has its roots in the resources available to the firm and that ultimately quality is determined by how well these resources are managed; hence, the quality dimensions will be captured using the managerial and resource influences shown in the framework (Figure 2.4). While the meaning of the various resource influences in the framework are rather straightforward, some additional discussion is needed to explain each of the four managerial influences. According to Holsapple and Joshi (2000): Leadership is concerned with building a trusting environment conducive to sharing knowledge. Coordination is concerned with developing and integrating reward and incentive systems that encourage knowledge sharing, as well as scheduling knowledge flows. Control is concerned with governing the content and channels of sharing . . ., ensuring that knowledge that is shared is of adequate quality and that sharing is not counterproductive. . . Measurement can aim at assessing and evaluating the knowledge sharing process. If sharing is linked to reward systems, how can sufficient credit be given to individuals/teams for sharing? What type of knowledge sharing is entitled for reward? How can we measure what and how much is shared, and its impacts on organizational performance? (p. 256)
26
In short, the managerial influences focus on promoting knowledge sharing, rewarding knowledge sharing, regulating the quality of shared knowledge, and evaluating the sharing of knowledge within the organization. Now consider Jennex and Olfman’s (2003) extension of DeLone and McLean’s IS Success Model (2003) (discussed in Section 2.2); the system quality dimension is concerned with assessing (evaluating) how well the KMS supports (i.e., promotes and rewards) knowledge sharing, and how well the system is supported by the IS staff and the organizational infrastructure. The knowledge/information quality dimension is concerned with regulating the quality and quantity of shared knowledge. In summary, both the managerial influences of Holsapple and Joshi’s (2000) framework and the system and knowledge/information quality dimensions of Jennex and Olfman’s (2003) model are concerned with promoting knowledge sharing, regulating the quality of shared knowledge, and evaluating the sharing of knowledge within the organization; each of these tasks involve the use of resources (i.e., resource influences—human, material, financial, and knowledge). In the KMS Success Model, all four factors: top management leadership, collaboration, quality of knowledge, and compensation schemes are tied to knowledge sharing. The top management leadership variable assesses the degree to which top management leadership is promoting knowledge sharing and allocating adequate resources. The collaboration variable is used to evaluate the degree of knowledge sharing in the organization. The quality of knowledge variable is used to assess the quality of
27
shared knowledge. The compensation schemes variable is used to assess the extent to which knowledge sharing and reuse are rewarded. See Table 2.1 for a comparison of variables from the different frameworks and models just discussed. Because environmental influences are largely external and are typically outside the control of the organization, environmental influences will be excluded from Table 2.1—even though they are captured in the KMS Success Model via the variable, KM strategy emphasized. Table 2.1 will focus on the managerial influences because these are the issues that the organization has the most control over (Holsapple and Joshi 2000). All variables in the KMS Success Model will be defined and discussed in detail in Chapter III.
Table 2.1. Comparison of Variables for Various KM Models. Comparison of Variables for Various KM Models Knowledge Management Activity Promoting knowledge sharing Rewarding knowledge sharing Regulating knowledge sharing (quality) Evaluating the degree of knowledge sharing
Holsapple & Joshi (2000)
Jennex & Olfman (2003) an extension of DeLone & McLean (1992, 2003)
Leadership
System quality
Coordination
-------
Control
Information quality
Measurement
System quality
The KMS Success Model Top management leadership Compensation Schemes Quality of knowledge Collaboration
In constructing Table 2.1, the starting point was the managerial influences identified by Holsapple and Joshi (2000). Based on the definitions that Holsapple and Joshi (2000) provided for each of the managerial influences, the next step was to identify the KM activity that each influence was designed to perform. At this point, the system 28
and information quality variables identified by Jennex and Olfman (2003) were added to the table. The final step was to categorize each of the success factors identified in the KMS Success Model proposed in this study using the definitions provided in this study for each factor. For the Jennex and Olfman (2003) model, the promoting knowledge sharing activity is deemed to be captured through the system quality variable. Jennex and Olfman break down system quality into three components; one of which is the form of the KMS, which refers to the degree of computerization and integration of KM processes as well as the accessibility of computerized knowledge; another component is the level of the KMS, which concerns how well the KMS enables search and retrieval functions. In short, a system that makes computerized knowledge readily accessible, easily searchable, and easily retrievable (i.e., a system with high system quality) provides a technological platform that is conducive to (i.e., promotes) knowledge sharing. As for the rewarding knowledge sharing activity, this activity is not captured by the variables in the Jennex and Olfman Model. The system quality variable is also deemed to capture the evaluating the degree of knowledge sharing activity. In Jennex and Olfman’s (2003) KMS Success Model, the system quality dimension focuses on: (1) how well the KMS is able to support knowledge creation, storage/retrieval, transfer, and application; (2) how much of the organization’s knowledge is codified; and (3) how well the KMS is supported by the IS staff as well as the organization’s infrastructure. Before knowledge can be retrieved from the KMS, it must first be shared and codified (i.e., submitted to the KMS in a form that 29
can be stored). Therefore, any assessment of how much of the organization’s knowledge has been codified and stored in the KMS is an indication of the extent (i.e., the degree) of knowledge sharing in the organization. The final KM activity, regulating knowledge sharing (quality), is captured through Jennex and Olfman’s knowledge/information quality variable. Jennex and Olfman define the knowledge/information quality dimension as the extent to which the system captures and makes available the right knowledge with sufficient context to the right users at the right time.
30
CHAPTER III THEORY, MODEL, AND HYPOTHESES
3.1 Model Development This research proposes a list of four KMS success factors: top management leadership, collaboration, quality of knowledge, and compensation schemes. Though all four factors are supportive of KMS success (H1, H3, H5, H7), some factors are undoubtedly more essential than others, given differences in knowledge management strategies (H2, H4, H6, H8). While some combination of these four factors (as determined by the organization’s chosen knowledge strategy) should promote KMS success, organizations should keep in mind that other factors in addition to these four factors may play a role in (1) the achievement of KMS success and in (2) transforming a successful KMS into an overwhelmingly successful KMS. These additional factors will be determined by the unique managerial, resource, and environmental characteristics of the organization (Holsapple and Joshi 2000). After all, the four factors identified in this article are necessary for success but are by no means a comprehensive listing of factors capable of contributing to the overall success of the KMS. Future extensions of this line of research will identify additional KMS success factors that are moderated by KM strategy emphasized and that are more important to either a KMS emphasizing a KRS or a KMS emphasizing a KIS. An organization’s critical KMS success factors are determined by whether the organization adopts a KM strategy emphasizing the exploration or the exploitation of 31
knowledge. The KMS Success Model contends that in order for a KMS to be successful a certain blend of KMS success factors (organizational characteristics) must exist in the organization; this blend is determined by the KM strategy emphasized by the organization. Jennex and Olfman (2005) provide a listing of eight KMS success factors with supporting citations; these factors include: (1) senior management support including allocation of resources, leadership, and providing training (i.e., top management leadership); (2) an organizational culture that supports learning and the sharing and use of knowledge (i.e., collaboration); (3) established measures to verify that the right knowledge is being captured (quality of knowledge); and (4) motivation and commitment of users including incentives and training (i.e., compensation schemes). Massey et al. (2002) contend that a successful KM initiative must first consider the process being focused on, then consider the people involved before finally determining what KM tool (system) is most appropriate; “by focusing on the process, and then combining the richness of human expertise with the efficiencies of technology, an organization will likely be more successful in managing its knowledge assets” (p. 287). Massey et al. (p. 271) describe how “Nortel’s systematic focus on process and the underlying needs of its people led to the development and implementation of a [successful] KM tool, an electronic performance support system (EPSS), that supported the diverse needs of knowledge workers.” Massey et al. also demonstrate how Holsapple and Joshi’s (2000, 2002) influences promoted Nortel’s process-oriented KM strategy; “this case research illustrates that the success of a process-oriented KM strategy depends on key enabling factors: managerial, resource, and environmental influences” (p. 287). 32
3.2 Discussion of Model and Hypotheses See Figure 3.1 below.
Figure 3.1. The KMS Success Model (with hypotheses).
3.2.1 Knowledge Management Strategy Every KMS has characteristics that differentiate it from other KMS. Despite this diversity, most KMS tend to emphasize one of two knowledge management (KM) strategies, knowledge exploration or knowledge exploitation (Zack 1999). A knowledge exploration strategy (KRS) typically prevails when the knowledge in the industry is 33
rapidly changing, and an organization finds itself needing to create new knowledge just to maintain the status quo. An organization is pursuing a knowledge exploration strategy if the organization’s primary focus is on creating or acquiring the necessary knowledge to become and remain competitive in its industry (Zack 1999). The strategy is often implemented in organizations facing continuous changes that render most of its experience inapplicable to new situations. Perhaps Nonaka (1991) best describes the KRS in its most extreme form. IN AN ECONOMY where the only certainty is uncertainty, the one sure source of lasting competitive advantage is knowledge. When markets shift, technologies proliferate, competitors multiply, and products become obsolete almost overnight, successful companies are those that consistently create new knowledge, disseminate it widely throughout the organization, and quickly embody it in new technologies and products. These activities define the ‘knowledge-creating’ company, whose sole business is continuous innovation. (p. 22) Examples of firms that might emphasize a KRS include firms that focus on cutting edge research and development projects, firms that focus on fashion and must monitor changing clothing trends, and firms that focus on providing financial services and must monitor changing prices/rates of return/market trends/etc. Alternatively, a knowledge exploitation strategy (KIS) typically prevails when an organization possesses knowledge resources and capabilities above and beyond those needed to maintain a competitive position in its industry. An organization is pursuing a knowledge exploitation strategy if the organization’s primary focus is on further utilizing and capitalizing on all that the existing knowledge platform has to offer the organization. The knowledge platform may be exploited within or across the organization’s competitive niches (Zack 1999) because of the relative stability of the firm’s environment 34
and the recurring problems that the firm continues to solve. An example of a firm that might emphasize a knowledge exploitation strategy is a company that purchases computer components and assembles them into functioning computers; employees could consult the KMS to get knowledge about which components will work together, which components will work best together, which steps need to be taken to resolve common problems encountered in assembly or testing, and which employees are most experienced in handling a specific type of problem. This research will argue that the KM strategy emphasized by a firm dictates which organizational characteristics must be in place for a KMS to be successful. Because a knowledge exploration strategy involves focusing on novel issues and knowledge creation, there is greater risk associated with a KRS than with a knowledge exploitation strategy, which emphasizes knowledge reuse. With the higher risk environment associated with an exploration strategy, the necessary blend of success factors is more rigid and must include top management leadership and collaboration. By definition, knowledge creation is at the core of a KRS. For successful knowledge creation to transpire, employees must willingly and routinely share knowledge and insights with others in a supportive manner; in essence, employees must adopt attitudes and behaviors supportive of collaboration (Alavi and Leidner 2001). One such attitude is trust, and trust can be facilitated through support of top management (Deloitte Research 2001). According to Deloitte Research (2001), top management must support or reward the trusting attitudes that facilitate collaborative efforts within and between organizations
35
while taking an active role to ensure that these efforts are directed in the best interests of the firm. In contrast to a knowledge exploration strategy, the knowledge exploitation strategy focuses on reuse of existing knowledge already held by a company. By definition, a KIS focuses on utilizing existing knowledge rather than searching for or creating new knowledge. This lower requirement for knowledge creation makes a KIS inherently less risky than a KRS, which emphasizes knowledge creation. After all, the outcomes associated with exploiting tried and true knowledge resources are much more predictable than the outcomes associated with applying recently created and untried knowledge resources. The view of the KIS as being an easier (less risky) strategy is supported by Szulanski (1996 p. 31), which states that “[k]nowledge with a proven record of past usefulness is less difficult to transfer. Such a record hints of robustness . . .” In short, there is a reduced risk of failure associated with exploitive type projects due to the high level of knowledge reuse and the familiarity of the domain.
3.2.2 Top Management Leadership Top management leadership “refers to the extent to which [KM] efforts are promoted by the top management of the firm" (Rai and Bajwa 1997 p. 946) where top management refers to the individual or individuals responsible for allocating resources for knowledge management and for specifying the knowledge management program for the organization. According to Massey et al. (2002 p. 282), “[p]ast research suggests that lack of commitment of top leadership to a KM initiative would likely lead to failure in 36
transforming the way(s) in which the organization leverages knowledge assets (Hiebeler 1996; Holsapple and Joshi 2000, 2002; Kotter 1995; Schein 1992).” In addition, numerous articles have emphasized the need for an executive sponsor/KM champion, someone to promote an organizational culture supportive of KM initiatives and activities throughout the organization, as being essential for a successful KMS (Ambrosio 2000; Bock 1999; Earl 2001; Huber 2001; “A Roadmap for Long-Term Knowledge Management Success” 2004; Rus and Lindvall 2002; Seeley and Dietrick 1999; Soliman and Spooner 2000). Massey et al. (2002) report that Nortel’s success is direct evidence of the impact of its KM initiative, which depended on leadership; “as evidenced at Nortel, a key to the success of the . . . KM initiative was the ongoing funding and investment for necessary human and technical resources. By providing financial resources . . . top-level management helped send a clear message that managing knowledge was mission critical” (p. 284). Hypothesis 1: The degree of top management leadership of knowledge management initiatives, as reported by KMS users, will be positively related to the level of KMS success reported by KMS users. Given the risky explorative (knowledge-creating) environment, top management must take a leadership role in promoting knowledge management. “The main job of managers in the knowledge-creating company is to . . . give voice to a company’s future by articulating metaphors, symbols, and concepts that orient the knowledge-creating activities of employees” (Nonaka 1991 p. 40). Setting the direction for the company and emphasizing the importance of managing knowledge is mandatory for a KMS emphasizing a knowledge exploration strategy to be successful. Without top 37
management leadership, employees will be hesitant to engage in knowledge exploration for fear of failure and potential negative repercussions (Leonard and Sensiper 1998). To overcome this resistance, top management engages in coordinating activities to support KM such as: taking an active role in the KM process, staying in contact with KM personnel, providing adequate resources to carryout KM activities, and emphasizing the importance of KM (Rai and Bajwa 1997). Hypothesis 2: The positive relationship between top management leadership and KMS success will be stronger for organizations following a knowledge exploration strategy than for organizations following a knowledge exploitation strategy. Because some companies may not have a traditional hierarchical structure where power is concentrated at the top of the organization, this study defines top management more broadly to accommodate a more diverse set of firms—traditional firms as well as firm’s where power is dispersed throughout the organization or concentrated at levels beneath top management (see the first sentence of this section for the definition of top management). This approach is taken because in any organization, regardless of its structure, some individual or group of individuals is responsible for allocating resources and setting the agenda for the organization’s knowledge management efforts.
3.2.3 Collaboration Collaboration is defined as the extent to which individuals actively communicate, cooperate, and help one another in their work by sharing knowledge and expertise with one another (Hurley and Hult 1998; Lee and Choi 2003; Rus and Lindvall 2002). A culture that promotes the “willingness of employees to share and contribute what they 38
know and to leverage explicit content from inside and outside the organization is a critical success factor for knowledge management” (Earl 2001; Holsapple and Joshi 2000; Seeley and Dietrick 1999 p. 20). Hypothesis 3: The degree of collaboration in an organization, as reported by KMS users, will be positively related to the level of KMS success reported by KMS users. For successful knowledge creation (i.e., knowledge exploration) to occur, employees must “willingly and consistently share their knowledge and insights” (Alavi and Leidner 2001 p. 126). According to Nonaka (1991 p. 26), “[n]ew knowledge always begins with the individual . . . In each case, an individual’s personal knowledge is transformed into organizational knowledge valuable to the company as a whole. Making personal knowledge available to others is the central activity of the knowledge-creating company.” The high-risk exploration environment will cause individuals to resist sharing their ideas with others for fear of sacrificing knowledge that makes them unique and provides them with increased job security. To overcome this resistance, measures to ensure collaboration must be taken. According to knowledge workers “interviewed by Standing and Benson (2000 p. 343) in their study of an emplaced KMS: [1] ‘If I share my knowledge others may take advantage of that. Will they do the same for me?’ [2] ‘People are afraid to share their knowledge and experiences as they feel their positions might be taken away from them’” (Huber 2001 p. 73; Kankanhalli, Tan, and Wei 2005). In essence, management must encourage organization members: to be supportive of others’ KM activities, to interact with organization members both in and outside of a member’s organizational unit, and to just be helpful in general (Lee and Choi 2003). 39
Hypothesis 4: The positive relationship between collaboration and KMS success will be stronger for organizations following a knowledge exploration strategy than for organizations following a knowledge exploitation strategy.
3.2.4 Quality of Knowledge in the KMS The “quality of knowledge in the KMS” construct draws on the “information quality” construct in the DeLone and McLean (1992) IS Success Model. As information quality is deemed essential for IS success, and since a KMS is in simple terms an IS that works with knowledge rather than information, the quality of knowledge in the KMS would appear to be a critical success factor for KMS. The “quality of knowledge in the KMS” construct is defined as the degree to which the knowledge provided by the KMS has the attributes of content, accuracy, and format required by the user (Rai, Lang, and Welker 2002). Jennex and Olfman (2003) identified the quality of knowledge in the KMS as being a critical KMS success factor and included knowledge quality as a variable in their KMS success model. In addition, Holsapple and Joshi (2000) included in their framework, the Influences on the Management of Knowledge, a managerial resource, which they called control. “Control is concerned with ensuring that needed knowledge resources and processors are available in sufficient quality and quantity . . .” (Holsapple and Joshi 2000 p. 240). According to Seeley and Dietrick (2000 p. 21), the quality of the knowledge in the KMS, i.e., the content of the KMS, “is often overlooked as a component of knowledge management strategy, yet it lies at the very heart of knowledge management, as it is largely about the retrieval, distribution and application of previously acquired 40
knowledge to help workers do their job better.” Markus (2001) emphasizes the importance of electronic knowledge repositories for promoting knowledge reuse while indicating that poor knowledge quality (too much/too little knowledge, failure to document rationale, incomplete documentation, etc.) severely restricts the contribution of electronic knowledge repositories. Alavi and Leidner (2001, p. 130) contends that “the success of KMS partially depends upon the extent of use, which itself may be tied to . . . information [knowledge] quality (DeLone and McLean 1992).” In essence, the quality of the knowledge in the KMS is critical to the success of the KMS. Hypothesis 5: The quality of knowledge in an organization’s KMS, as reported by KMS users, will be positively related to the level of KMS success reported by KMS users. Since a knowledge exploitation strategy focuses on the reuse of knowledge as a primary task, it would seem paramount for the knowledge stored in the KMS to be of high quality in order for the system to have positive impacts on the individual users and on the organization (i.e., for the KMS to be successful). More specifically, for knowledge to be available for reuse, knowledge must first be captured, stored, and indexed to facilitate its search and retrieval. If the knowledge that gets captured and stored is incomplete or is incorrect, the knowledge that gets retrieved may lack sufficient detail or may contain so many errors that its reuse is at best inconvenient or at worst impossible. Hypothesis 6: The positive relationship between quality of knowledge in the KMS and KMS success will be stronger for organizations following a knowledge exploitation strategy than for organizations following a knowledge exploration strategy.
41
Since a knowledge exploration strategy requires minimal knowledge reuse, the positive relationship between quality of knowledge in the KMS and KMS success should be weaker than for a knowledge exploitation strategy.
3.2.5 Compensation Schemes Compensation schemes refer to mechanisms developed in the organization to encourage reuse of knowledge resources (Nidumolu and Knotts 1998). Compensation schemes can be monetary (e.g., bonuses) or nonmonetary (e.g., recognizing reuse or including reusability as a component of performance reviews) measures implemented to promote the creation and use of reusable knowledge resources (Nidumolu and Knotts 1998). To date, mixed findings on the impacts of compensation schemes have persisted. Despite the mixed findings (for example: Kankanhalli, Tan, and Wei 2005; Ko, Kirsch, and King 2005), the vast majority seems to agree that KMS success requires some form of compensation scheme to motivate participation in KM efforts (Alavi and Leidner 2001; Ambrosio 2000; Cook 1999; “Do We Know How to Do That? Understanding Knowledge Management” 1999; Earl 2001; Huber 2001; Rus and Lindvall 2002; Szulanski 1996). Furthermore, the majority tends to agree that unlike extrinsic (monetary) compensation schemes, intrinsic (non-monetary) compensation schemes tend to have fewer negative impacts on KMS success (Bock, Zmud, and Kim 2005). As examples, two recent studies show that carefully designed compensation schemes promote KMS success and that non-monetary rewards may be more effective (Garud and Kumaraswamy 2005; Kankanhalli et al., 2005). 42
Markus (2001) discusses the integral role that electronic knowledge repositories play in successful knowledge reuse and points out that successful knowledge reuse requires appropriate incentives. According to Markus (2001 p. 79), “As Dixon [22] points out, successful knowledge transfer or reuse requires a complete solution. It is not just a matter of providing access to information technology and repositories. It also means careful attention to the design of incentives for contributing to and using repositories.” Hypothesis 7: The degree of compensation schemes in an organization, as reported by KMS users, will be positively related to the level of KMS success reported by KMS users. Even though a knowledge exploitation strategy emphasizes knowledge reuse, there is still a minimal level of knowledge creation required to tailor prior knowledge for the situations currently faced. If individuals “satisfice” (i.e., settle for using the best prior solution without customizing it for the situation at hand or storing the new solution in the KMS), then over time the users will consistently reuse knowledge that is more and more outdated and less and less appropriate for the situation at hand. Over time, solutions to current situations will have a tendency to become increasingly sub-optimal. Ultimately, the KMS can become ineffective. According to Bock (1999 p. 25), “the implementation of the knowledge management process is not a one-time activity. Knowledge gets old and declines in value. New knowledge must be sought, captured, and applied.” In short, compensation schemes are necessary to ensure both reuse of existing knowledge assets and the creation/storage of new reusable assets (created by making slight modifications to tailor old assets to individual cases). 43
Hypothesis 8: The positive relationship between compensation schemes and KMS success will be stronger for organizations following a knowledge exploitation strategy than for organizations following a knowledge exploration strategy.
3.2.6 Summary The KMS Success Model includes two independent factors that are believed to be more essential to the success of a KIS and two independent factors that are believed to be more essential to the success of a KRS (based on IS/KMS success/strategy literature, frameworks, and models); these hypothesized relationships will be tested as outlined in Chapter IV to provide empirical evidence of (1) the positive relationships between the selected factors and KMS success and (2) the potential moderating role of KM strategy emphasized. Holsapple and Joshi (2000) identify four managerial influences: leadership, coordination (involves compensation schemes), control (involves quality), and measurement (involves collaboration—an indicator of the degree of knowledge sharing). Upon close examination of the factors in the model, all of them can be viewed as a mixture of managerial and resource influences. Top management leadership involves management promoting KM and specifying how resources will be used to further promote KM. Collaboration is the extent to which individuals actively communicate, cooperate, and help one another in their work by sharing knowledge and expertise with one another (knowledge and expertise are resources; collaboration serves a measurement function—a managerial influence). Quality of knowledge in the KMS encompasses the managerial influence, control, and the resource influence, knowledge. Compensation
44
schemes involve the managerial influence, coordination, and potentially involve the resource influence, financial (i.e., if monetary compensation is given). In short, top management leadership and collaboration are essential in order for a KMS emphasizing a knowledge exploration strategy to be successful. Meanwhile, the essential factors contributing to the success of a KMS emphasizing a knowledge exploitation strategy are the quality of knowledge in the KMS and compensation schemes. Details on how these hypotheses will be tested will be discussed in Chapter IV.
45
CHAPTER IV METHODOLOGY
4.1 Overview of Research To test the hypotheses stated above, a web-based survey was administered. The survey targeted users as the level of analysis; questions were asked to verify that respondents were KMS users. The unit of analysis was the KMS user because the individual users should best be able to judge their perceptions of top management leadership, collaboration, quality of knowledge in the KMS, and compensation schemes as well as the individual impacts that the system has had, which ultimately determines the organizational impacts of the KMS.
4.2 Development of Survey This instrument utilized existing scales, with slight adaptations, addressing KMS success, top management leadership, collaboration, quality of knowledge in the KMS, and compensation schemes. A new scale was created to assess the KM Strategy Emphasized factor. Where multiple scales were found in existing literature, apparent face validity, reported reliability (using Cronbach’s alpha or composite reliability), and closeness of context to my context were considered. In choosing scales, the scale with the highest reported reliability was chosen; if reliability was not reported then the decision was based on face validity and context.
46
Some scale items were reworded (i.e., changing wording from information to knowledge, etc.) to more accurately address the hypotheses being tested. Scales were also modified so that the majority of the questions could utilize a similar (5-point, Likerttype) answer format (where the left-most answer choice for a question was coded as “1” and the right-most answer choice was coded as a “5”); this created a survey instrument with a more uniform appearance. The questions pertaining to KMS Success were scaled such that larger values indicate more successful KMS. Due to the novelty of this study, there is some uncertainty as to whether KM strategy emphasized is best portrayed as: (1) a single factor (a continuum with the extremes being exploitation and exploration) or (2) two factors (an exploitation factor and a separate exploration factor with values representing high or low levels of emphasis). For this study, the KM strategy questions were scaled so that higher values were indicative of a knowledge exploration strategy while lower values were indicative of a knowledge exploitation strategy. The web-based survey consisted of a login screen, a survey screen, and a request for results screen. The survey screen provided a definition of KMS and 4 questions to identify the purpose(s) of the KMS in place (e.g., knowledge: creation, storage/retrieval, transfer, or application). Questions 5-15 assessed the success of the organization’s KMS; the KM strategy emphasized by the organization was assessed by questions 16-20. Top management leadership was measured by questions 21-24, and collaboration was measured by questions 25-29. To assess the quality of knowledge in the KMS, questions 30-36 were included. The extent to which compensation schemes were in place and the types of compensation schemes (e.g., monetary or non-monetary) in place were evaluated 47
by questions 37-41 and 42-43, respectively. Questions 19, 20, and 33 must be reverse scored. Questions 44-54 were used to acquire general background information about the respondents and the organization; several of these questions were adapted from Straub (1989). Scale details are shown in Figure 4.1.
4.3 Sample Selection and Data Collection Over eighty Yahoo groups and similar mailing lists with subject areas relating to knowledge management, information technology, and/or specific business professions (e.g., accountants, etc.) were identified. Multiple e-mails containing a link to the survey were sent to a subset of these groups in order to acquire the data necessary to assess the initial reliability and validity of the survey instrument. Since no survey modifications were made after evaluating the reliability and the validity of the scales, all groups, including those used for pre-test purposes, were sent e-mails requesting their participation. Once all responses were collected (204 in total) validity and reliability were again assessed, questions and/or factors were dropped as deemed necessary (details are disclosed in Chapter V), and average values for each respondent’s answers to the retained questions used to measure each variable (four KMS Success Factors, KM Strategy Emphasized, and KMS Success) were calculated–i.e., six averages were computed for each respondent (SuccessAvg, LeadershipAvg, etc.). The rows were then sorted in ascending order by the KM Strategy Emphasized averages just calculated. The group
48
Figure 4.1. The KMS Success Model (with associated scale information). 49
with KM Strategy Emphasized averages at or below “3.0” was considered to emphasize a KIS, and the group with values above “3.0” was considered to emphasize a KRS. Using this procedure, the KIS and KRS subgroups were approximately equal in size (87 and 117, respectively). As outlined by Aguinis (1995), using pre-testing results to revise scales as needed to help ensure acceptable levels of reliability and having approximately equal subgroups (KIS and KRS in this case) helped to avoid an unnecessary loss of power and improved the likelihood of detecting a moderating effect with adequate power. In addition to these steps, the large sample size also helped to improve power. Aguinis (1995 p. 1148) reported that Stone-Romero and Anderson (1994) “found that what they defined as a small effect size was typically undetected when sample size was as large as 120, and unless a sample size of at least 120 was used, even medium and large moderating effects were, in general, also undetected.” Therefore, this study’s sample size of 204 should be sufficient for detecting a medium to large moderating effect.
50
CHAPTER V DATA ANALYSIS
5.1 An Overview In terms of reliability, Nunnally’s (1967) threshold of 0.70 for acceptable Cronbach’s alpha scores was utilized. When the actual data was received, reliability was assessed for each construct using Cronbach’s alpha; factor loadings were measured for each of the success factors and the strategy factor to see if appropriate loading occurred (a measure of convergent and discriminant validity); regression and path analysis were utilized to test the hypotheses. Harman’s single-factor test was used to check for common method variance (Podsakoff and Organ 1986).
5.2 Validity and Reliability Using the first thirty-five survey responses as a pre-test, a correlation matrix was constructed for questions five through forty-one. The matrix was examined as a preliminary test of convergent and discriminant validity; correlations should be high for theoretically similar measures and low for theoretically dissimilar measures (Trochim 2001). This analysis revealed that question thirty-three, which assesses the quality of knowledge in the KMS (quality for short), had poor convergent validity. Analysis of the correlation matrix also revealed a potential problem with some quality questions correlating more strongly with success questions than with other quality questions. Given that the scales for quality and success are two of the larger scales on the survey, 51
the decision was made to continue to collect data without modifying the survey with the rationale being that one or more questions and worst case one factor (quality) could be dropped if these issues persisted following additional data collection. In regards to validity, all scales (KMS success, top management leadership, collaboration, quality, compensation schemes, and KM strategy emphasized) exhibited values greater than the 0.70 (Nunnally 1967) threshold for Cronbach’s alpha once question thirty-three was dropped from the analysis. Ultimately, 204 responses (including the initial 35) were collected (General information on respondents, and general statistics for each factor appear in the Appendix). Again, a correlation matrix was constructed for questions five through fortyone. Convergent and discriminant validity as well as reliability were assessed as demonstrated for the pre-test data. Analysis of the correlation matrix reveled that question thirty-three (measuring quality) and questions nineteen and twenty (each measuring KM strategy emphasized) suffered from poor convergent validity and should therefore be dropped from future analysis. The correlation matrix also showed strong correlations between multiple success and quality questions, which suggested that the quality factor (questions thirty through thirty-six) be dropped from future analysis (The resulting correlation matrix appears in the Appendix). To further justify dropping questions nineteen, twenty, and thirty through thirtysix from future analysis, an exploratory factor analysis was conducted for questions five through forty-one. The exploratory factor analysis indicated that questions nineteen and twenty did not load on the same factor as the other strategy questions; in fact, they each 52
loaded on a different factor with a loading in excess of 0.40. Additionally, five of the seven quality questions loaded with the success questions with loadings in excess of 0.40. The two remaining quality questions loaded together on a distinct factor (loadings in excess of 0.68) and cross loaded with the success questions (loadings of at least 0.40). Based on the analysis of the correlation matrix and the exploratory factor analysis, questions nineteen, twenty, and thirty through thirty-six were dropped from all future analysis—note that this means that the quality factor in the model is being dropped and that hypotheses five and six cannot be tested in the current study. After dropping these questions, the exploratory factor analysis was repeated and the following results were obtained (Table 5.1); the largest loading for each row is highlighted. Note that the names are a combination of the factor being measured and the survey question number; for example, “Success5” indicates that survey question five is intended to measure KMS success. These results indicate good convergent and discriminant validity because each group of questions loaded together on a separate factor with loadings in excess of the commonly cited 0.40 minimum loading level (Gefen et al., 2000) on the intended factor and with lower loadings on the un-intended factors (lower than 0.40 and at least 0.30 lower than the highlighted loading). Furthermore, all scales had acceptable Cronbach’s alpha values (i.e., values exceeded 0.70). Since the research method did not: (1) utilize multiple methods, (2) collect data using different sources to measure the dependent and independent variables, or (3) utilize a longitudinal design, common response bias, the threat that the “data are related merely because they come from the same source [i.e., the 53
data were collected using a single survey approach to access the entire model],” (Jap and Anderson 2004 p. 8), is a threat. Harman’s single-factor test was used to test for the Table 5.1. Results of Exploratory Factor Analysis.
Success5 Success6 Success7 Success8 Success9 Success10 Success11 Success12 Success13 Success14 Success15 Strategy16 Strategy17 Strategy18 Leadership21 Leadership22 Leadership23 Leadership24 Collaboration25 Collaboration26 Collaboration27 Collaboration28 Collaboration29 Compensation37 Compensation38 Compensation39 Compensation40 Compensation41
Factor1 0.7884 0.8223 0.8496 0.8635 0.8887 0.7990 0.7844 0.8316 0.6569 0.6746 0.7878 -0.0556 -0.1224 -0.0768 0.1571 0.2002 0.2088 0.2795 0.2012 0.2640 0.2167 0.1296 0.0403 0.1412 0.1297 0.1219 0.2231 0.0667
Factor2 0.0519 0.0694 0.0899 0.1002 0.0042 -0.0161 0.0312 0.1812 0.3109 0.3320 0.2582 -0.0953 -0.0436 0.1104 0.0868 0.1210 0.2280 0.2615 0.1605 0.0738 0.1019 0.1640 0.2309 0.8530 0.8363 0.8279 0.5580 0.6240
Factor3 0.0514 0.1359 0.1353 0.0566 0.1317 0.1761 0.1381 0.1113 0.1048 0.1733 0.1166 -0.1440 -0.0673 -0.0441 0.8178 0.8383 0.8073 0.7147 0.2983 0.2121 0.1505 0.1190 0.1534 0.1266 0.1120 0.1542 0.0000 0.1853
Factor4 0.2468 0.2332 0.1555 0.1831 0.1446 0.0726 0.0710 0.0751 -0.0446 0.0779 0.1110 -0.1613 -0.0048 -0.0684 0.2568 0.2472 0.1905 0.1967 0.6114 0.7217 0.7290 0.6992 0.6024 0.1212 0.1427 0.1618 0.1523 0.1219
Cronbach's Alpha Factor5 0.0293 -0.0266 -0.0628 -0.1473 0.0066 -0.1238 0.9536 -0.0722 -0.0204 -0.1178 -0.0154 -0.1227 0.6518 0.8399 0.7034 0.7918 -0.0958 -0.1459 0.9003 -0.1098 -0.0244 -0.0477 -0.2300 -0.2532 0.7997 -0.0319 0.1078 0.1015 0.0428 0.0237 0.8450 -0.2152 -0.0445
presence of a significant common response bias; in accordance with Harman’s singlefactor test an unrotated factor analysis for all of the variables in the study was conducted; 54
since this analysis indicated seven factors with eigenvalues greater than one and since the largest factor accounted for less than the majority, only 36.84%, of the variance, common response bias does not appear to be a serious threat to the validity of the study (Harman 1967, Podsakoff and Organ 1986, Podsakoff et. al 2003). In conclusion, the modified scales exhibited good convergent validity, good discriminant validity, and good reliability
5.3 Regression Analysis Hypothesis testing with regression analysis was done in two parts. First, the model shown in Figure 5.1 was used to assess the odd hypotheses: H1, H3, and H7 (H5 could not be tested because the quality factor had to be dropped from further analysis).
X
Figure 5.1. Model for Testing Odd Hypotheses with Regression Analysis. 55
The regression equation used to test this model is as follows:
SuccessAvg = a + b1LeadershipAvg + b2CollaborationAvg + b3CompensationAvg + e.
A hypothesis is supported if the bi associated with the predictor variable is positive and the associated p-value is less than 0.05. As indicated in Table 5.2, all three bi are positive and have associated p-values of less than 0.05; therefore, hypotheses H1, H3, and H7 are supported. Hence, top management leadership, collaboration, and compensation schemes are each positively related to the level of KMS success reported by KMS users as hypothesized. The model has an r2 of 0.2637.
Table 5.2. Results for Odd Hypotheses.
Parameter a b1 b2 b3
Parameter Estimates 1.92228 0.18918 0.24985 0.13492
p-value
