Description
Organizations have recognized that knowledge constitutes a valuable intangible asset for creating and sustaining competitive advantages
Knowledge Sharing and Management in Large, MultiNational Corporations
Joseph G. Davis Associate Professor of Information Systems University of Wollongong Wollongong, NSW 2522 AUSTRALIA Email: [email protected] Phone: 61 2 4221 3648 Fax: 61 2 4221 4474 Eswaran Subrahmanian Principal Research Scientist Institute for Complex Engineered Systems (ICES) Carnegie Mellon University Pittsburgh, PA 15217 Email: [email protected] Arthur W. Westerberg University Professor and Swearingan Professor of Chemical Engineering Carnegie Mellon University Pittsburgh, PA 15217 Email: [email protected] (Please address all correspondence to Joseph G. Davis) This research was funded by a grant from the Carnegie Bosch Institute, Pittsburgh, PA under the ‘Impact of Global Information Revolution on International Management’ program. Previous versions of this paper have been presented at the bi-annual Carnegie Bosch Institute conference held in San Francisco, October 21-23, 1999, at the Prototyping Research Laboratory, Department of Mechanical Engineering, Stanford University, Palo Alto, CA on October 19, 1999, at the Du Pont Experimental Station, Wilmington, DE on December 16, 1999, and at the Du Pont Photopolymers & Electronic Materials headquarters at Research Triangle Park, NC on December 17, 1999. We have benefited from comments and suggestions from several participants at these presentations and seminars. We are indebted to Joe Miller, Du Pont’s Vice President, Research and Rita Seelig Ayers and Margaret Isselmann of Corporate Information Science for their cooperation and assistance without which the empirical segment of this study could not have been completed. We also thank the 44 respondents and Steven Miller (Leader of the Communication, Collaboration, and Coordination Program Office at Du Pont) for being generous with their knowledge and time. We acknowledge Professors V.S. Arunachalam and David Hounshell of Carnegie Mellon University for their guidance.
1
Knowledge Sharing and Management in Large, MultiNational Corporations
Abstract
Knowledge sharing and management (KSM) has emerged as an important issue for international management. However, there is considerable confusion as to what constitutes organizational knowledge, whether and how it can be systematically managed, and what are some of the effective organizational and technological mechanisms for facilitating knowledge management. This study seeks to unravel the complexities associated with it, initially by developing a typology of such mechanisms. This will provide a starting point for a detailed field study carried out in a large, multinational company (Du Pont) focusing on the critical issues, concrete practices, bottlenecks, and constraints in knowledge sharing and management in two functional areas, two business units, and four countries. Using the field data, we also elucidate a dynamic model of knowledge with emphasis on ongoing interpretation and contextualization of previously generated knowledge. Design guidelines for implementing computer-based systems to support KSM are also presented.
2
Knowledge Sharing and Management in Large, MultiNational Corporations
1.0 Introduction The notion of contemporary organizations as knowledge producing, sharing, and disseminating entities is gaining rapid currency among researchers in Organizational Studies and International Management, among others. While the critical role played by the stock and application of knowledge in economic development at the macro-economic level is relatively well understood (Machlup, 1980; Nelson and Winter 1982; Eliasson et al. 1990), its centrality in the management of individual firms is more of a recent concern. This interest is perhaps a response to the challenges posed by an increasingly complex business environment characterized by intensified competition, greater globalization, and compressed product life cycles and the consequent information overload for senior management. Concurrently, advances in new information and communications technologies (NICT) in the form computersupported cooperative work (CSCW) systems, groupware, knowledge-based systems, internet, intranet, and the world-wide web (WWW) promise capabilities for developing effective solutions to the KSM problem.
The issues referred to above have been faced in a more acute form by large, multinational corporations (MNCs) for which the forces of global integration, local differentiation, and worldwide innovation have become stronger and more compelling. Several scholars have argued that such firms have had to devise means to enhance their global flexibility and learning levels in order to stay competitive (Bartlett and Ghoshal, 1991; Doz and Prahalad, 1991, among others). This is increasingly achieved through the adoption of new organizational capabilities for pooling world-wide knowledge and to transfer and adapt innovative product and process technologies and project management knowhow to international markets. We investigate empirically the organisational and technological mechanisms employed by MNCs to promote knowledge sharing and to develop and manage their intellectual resources. The implicit assumptions that underpin the category of ‘knowledge’ in this context as reflected in
3
the relevant literature and contemporary organizational practices will be explored and critiqued. We also address the barriers to effective knowledge management and discuss approaches to overcoming them through an alternative conceptualization of the dynamics of organizational knowledge and the design of appropriate NICT-based systems to facilitate its sharing and management.
This study was carried out in three phases. In the first, we developed a comprehensive typology of organizational mechanisms for knowledge sharing and management. An exhaustive review of the growing academic and practitioner literature provided the basis for this typology. In the second phase, we completed a detailed field-based, case study of the conceptualization of knowledge and KSM practices in R&D and project management in a large MNC - Du Pont, a diversified, manufacturing company with headquarters in the United States. The company has extensive R&D and project management operations in a number of countries. The data gathering was carried out through in-depth interviews with key R&D and engineering project management (EPM) executives in four different countries in two separate business units. The typology developed in the first phase and a preliminary set of interviews with senior Du Pont managers were used to develop semi-structured questionnaires for the data gathering. In the third and final phase, set of system requirements for the design of an integrated knowledge management system was generated based on our analysis of the field data.
This paper is organised as follows: in the second section we present an overview of the extensive literature on knowledge management and the related topic of organizational learning. This review constitutes the basis for the typology of KSM mechanisms proposed in section 3. Section 4 is devoted to the research methods and techniques employed in gathering the data for the case study. The fifth section includes the details of the case study followed by an outline of some of the generic system requirements for developing systemic solutions to the KSM problem.
4
2.0 Overview of the Relevant Literature 2.1 Organizational Knowledge There is emerging consensus that perhaps the most important source of sustainable competitive advantage in an increasingly turbulent global business environment is knowledge. The organizational capability to create, recognize, disseminate widely, and embody knowledge in new products and technologies is critical when faced with shifting markets, rapid product obsolescence, hyper-competition, and financial upheavals (Nonaka, 1991).
However, despite extensive discussions of concepts such as ‘knowledge management’, ‘intellectual capital’ and related ones like ‘organizational learning’ and ‘organizational memory’, there appears to be a lack of conceptual clarity in the minds of researchers and practising managers on their specific meanings. Davenport et al. (1998) has suggested that it is not useful to worry about the deeper ontological and epistemological dimensions of knowledge; they instead focus on specific knowledge management project as the unit of analysis. This, in our view, is a risky strategy since it is hard to study knowledge management without a coherent conceptualisation of organizational knowledge. We present below, a review of the divergent perspectives on organizational knowledge and the latent assumptions that underpin them.
2.11 Philosophical Issues From an epistemological standpoint, most of what is characterized as organizational knowledge falls within the purview of ‘weak knowledge’ (Goldman, 1999). This is sharply contrasted with ‘strong knowledge’ which is the traditional focus of philosophical inquiry. Strong knowledge imposes very stringent criteria for the use of the term ‘knowledge’ which includes true belief backed up by valid justification or warrant for the belief, and the systematic exclusion of alternative possibilities. Such knowledge is rarely attainable in organizational domains. Besides, this form of skeptical rationalism bears little resemblance to the action-centered processes of generating, sharing, and utilizing knowledge in organizations. It is unlikely that the ‘strong knowledge’ epistemology is particularly helpful1 in view of the inter-subjective and social nature of much of organizational knowledge. The latter tends to
1
The role and influence of strong knowledge in the activities of the R&D department cannot be ruled out.
5
privilege consensus, immediacy, and perceived use value over strong-verifiability, superobjectivity, and transcendence.
Weak knowledge implies true belief without the iron-clad backing of strong warrant or justifiability for the belief. It does, however, imply something more than just information and involves sustainable belief produced by the information that provides some kind of evidence pointing to the validity of the propositions and away from its rivals (Dretske, 1981; Goldman, 1999).
The justification for organizational knowledge is typically reliant on the following bases: i. Coherence: knowledge fragments obtain their justification by virtue or cohering or meshing with the rest of the knowledge base so as to produce a holistic effect (Quine, 1952; Goldman, 1999). ii. Pragmatism: pragmatist epistemology emphasizes the relationship between knowledge and action. This position was initially developed by William James and subsequently expanded on by Dewey who endeavored to bridge the chasm between theory and practice, knowledge and action. The validity and use of knowledge can go only so far as it can “…..pass into actions which rearrange or reconstruct in some way, be it little or large, the world in which we live” (Dewey, 1929). Barring an extreme instrumentalist view, pragmatism attempts to develop an interactive model of knowledge based on engagement with the world through action, experimentation, and experience (Nonaka and Takeuchi, 1995).
2.12 Economics and Organizational Knowledge
Traditional microeconomic theory depicts (technical) knowledge as a quasi-public good. It is characterized by high levels of indivisibility and non-excludability. Its generation is the result of scientific research and general methodological procedures. Its transfer is largely unproblematic and is viewed as a spontaneous aspect of the economic system. The ability to appropriate the knowledge by the innovator is low even though patenting and intellectual property rights can reduce the scope for societal benefits from the knowledge (Arrow, 1969; 1994; Antonelli, 1999).
6
This perspective has been challenged by a number of researchers. The distinction between technological information and technological knowledge is sharply drawn with the latter conceptualised as incorporating a set of capabilities and competences needed to utilize the knowledge which in turn can be leveraged to generate new knowledge. Such knowledge is generated by a process characterized by cumulativeness and path dependence (Jorde and Teece, 1990; David, 1993, among others). Knowledge, according to this view, is highly localized and embedded in the previous background and experience of individual firms. It is the result of a learning process and involves highly specific and relatively ‘tacit’ knowledge processing (Antonelli, 1999).
2.13 Other Representations of Organizational Knowledge
The central role of knowledge in the firm and the organizational processes and mechanisms for its integration and sharing across national borders is the primary basis of Kogut and Zander’s theory of the multinational corporation. They have also highlighted the need for the mechanisms to be sensitive to the degree of tacitness or codifiability of the knowledge (Kogut and Zander, 1992a; 1992b)
A range of definitions and perspectives on knowledge has been presented in the organizational literature. Kerssens Van-Drongelen et al. (1996) defines knowledge primarily in the context of R&D as “… information internalized by means of research, study, or experience, that has value for the organization” (Kressens Van-Drongelen et al., 1996 p.214). Similar conceptualizations of knowledge as the result of processing and refining of information have been implicitly or explicitly employed by a number of authors. This view is an extension of the information-processing paradigm popularized by March and Simon (1959), Simon (1977), and Galbraith (1974). More recently Simon (1996) has argued that the challenge for today’s managers is to filter and extract relevant knowledge from the vast amounts of potential and actual information available from internal and external sources. Effective systems need to be designed to intelligently filter information. In a similar vein, Davenport et al. describe knowledge as information combined with experience, context, interpretation, and reflection and knowledge production as comprising value addition to information (Davenport et al., 1998). Information is the flow of messages or meanings that might add to, restructure, or change knowledge according to Machlup (1980). Coyne (1997) refers to these and related 7
representations of knowledge as the system theoretical view according to which the essential knowledge is contained in the information content and the subjective, inter-subjective, and spatial aspects are largely ignored.
An alternative view has been championed by Dretske (1981), Nonaka (1991; 1994), among others. This perspective acknowledges the importance of subjective factors such as beliefs and their links to actions as well as the relatively tacit dimension of knowledge. Knowledge is associated with beliefs produced and sustained by information (Dretske, 1981). Information represents a flow of messages but knowledge is created and organized from it, anchored by the commitments and beliefs of the concerned individuals. There is also a connection between such knowledge and the subject’s ability to plan and act.
The more implicit and tacit dimension of knowledge has also been highlighted. For Polanyi (1967), explicit or codified knowledge is what is transmittable through formal and systematic languages. Tacit knowledge is more personal and subjective, making it difficult to be formalized and communicated. It tends to be deeply rooted in action, commitment, and involvement in a specific context. According to Nonaka, individuals are able to recreate their own systems of knowledge to accommodate ambiguity, noise, and randomness generated in the organization in its interaction with the external environment (Nonaka, 1994). Such knowledge resides in situated communities of interpreters rather than in texts or messages and these make sense only in particular interpretive contexts (Reddy, 1979). These communities emerge not through absorption of abstract knowledge but when members become insiders and acquire the community’s shared vision and speak the same language (Brown and Duguid (1991). The notion of contextualization of knowledge and evolving communities of practice have particular resonance for MNCs given the geographic distances and cultural differences across units around the world. Following Coyne (1997), we refer to this perspective as the pragmatic view.
Table 1 presents a somewhat stylized set of distinctions between the systems-theoretic and pragmatic perspectives on organizational knowledge. It is worth noting that the dichotomy between system theoretical and pragmatic perspectives and their respective epistemological and ontological assumptions pervade most of the writings on organizational knowledge.
8
System Theoretic View Source Documents, databases, systems, prototypes, processes and procedures, manuals etc. Codified or codifiable; explicit Exchange of documents, electronic means, formal training Relatively mechanistic Cartesian; separation of mind and body.
Pragmatic View People, communities
Form Transfer
Tacit; implicit; intuitive Socialization, apprenticeship, osmotic; situated learning Organic More holistic; unity of mind and body
Organization Philosophical Perspective
Table 1: Perspectives on Organizational Knowledge
2.2 Knowledge Sharing and Management As we would expect, the fundamental differences between the system theoretical and pragmatic views in characterising organizational knowledge are reflected in the divergent approaches and perspectives on knowledge creation, sharing and management in organizations. In general, the former tends to focus on structural and systemic approaches while the latter emphasizes human-centred processes such as socialization, self-organizing teams, extended social interactions, personnel rotation etc. Besides the diverse modes of knowledge creation and transformation, the globalized firms are faced with the challenge of mobilizing and integrating fragmented forms of knowledge spread all over the world (Cohendet et al., 1999). As well, developments in new information and communications technologies (NICT) are increasingly making it easier to separate, transport, and trade knowledge (Antonelli, 1996).
Nonaka has described processes for the creation and transformation of knowledge from tacit to more explicit forms. These include sharing language, experiences, mental maps and models through socialization, reconfiguring existing knowledge through sorting, adding, categorizing, explicit knowledge through combination (typically using computer technology), externalization of tacitly held knowledge through the use of appropriate metaphors and other triggering devices, and internalization of explicit knowledge by recontextualizing along the lines of double loop learning proposed by Argyris and Schon (1978). This involves the
9
creation of new knowledge by reconstructing existing perspectives, frameworks, or premises on an ongoing basis (Nonaka, 1994, p.19). Furthermore, Nonaka presents a model of knowledge expansion and growth arising from the dynamic interplay between tacit and explicit knowledge possessed by organizations through the four processes discussed above.
While it is true that knowledge is continuously generated throughout the organization, some of the areas of critical focus have included R&D (Kerssens-Van Drongelen et al., 1996; Jain and Triandis, 1990), project management, especially in distributed project management teams across multiple, geographically work sites (Van Fenema, 1997; Manheim, 1993), new product develoment (Clark and Fujimoto, 1991; Wheelwright and Clark, 1992; Takeuchi and Nonaka, 1986), joint ventures (Inkpen, 1998; Culpan, 1993), technology transfer (Souder et al., 1990), and customer needs and market developments (Ashton and Stacey, 1995). A few of the arguably effective mechanisms commonly prescribed for the sharing and management of knowledge in these and other studies include: • multifunctional project teams (Wheelwirght and Clark, 1992) • personnel rotation (Edstrom and Galbraith, 1992; Inkpen, 1998) • post-implementaton reviews and audits (Kumar, 1990) • best practices documentation ( Davenport et al., 1997) • structured task communication (Brown and Eisenhardt, 1995) • learning histories (Kleiner and Roth, 1997) • a variety of IT-based systems typically involving various combinations of inter/intranet, document management and communication, knowledge and database technologies. Most of these provide a subset of capabilities for searching, sharing and collaborating over documents in electronic repositories. One of the more celebrated among these are the intranet-based cooperative work and knowledge sharing system called virtual team network (VTN) developed by British Petroleum and used extensively. (Prokesch, 1997)
The ‘knowledge base’ of an organization may consist of several sources or media. KerssensVan Drongelen et al. (1996) has classified these sources into four categories based on the work of Zeleny et al. (1990), • brainware: knowledge in the minds of people, such as intuition, experience, specialized knowhow,
10
• hardware : tangible things such embodying knowledge such as material prototypes, products, R&D equipment, production processes, • groupware: unwritten knowledge shared by people such as rules of thumb, heuristics, procedures, stories, myths, • documentware : knowledge documented on paper or in databases/information systems such as CAD/CAM models, intranets, parts databases, patents, white paper, manuals, lab reports, hand books, project post-implementation reviews etc.
3.0 Toward a typology of KSM mechanisms Based on the preceding discussion, we propose a typology of knowledge sharing mechanisms employed by organizations to facilitate knowledge sharing and to effectively manage the knowledge resource. It consists of the following three categories: Type I: Type II: Type III: Static sharing of documents and knowledge bases (knowledge repositories) Structural and human resource-based mechanisms Integrated mechanisms for facilitating ongoing interpretation and contextualization of knowledge.
It should be noted that the categories are not strictly mutually exclusive in that there could be some overlap among them. Also, most organizations attempt to use all three of these and there are substantial synergies and potential for expansionary spirals in knowledge generation and utilization from the interplay across the categories. In this sense, the categories are best viewed as Weberian ‘ideal types’ to the extent that a specific KSM practice can be roughly mapped into any of one of the categories. The question as to whether organizations in general and MNCs in particular exhibit a dominant mode among the three types is an empirically one.
Type I: Static Sharing of Documents and Knowledge Bases (Shared Knowledge Repositories) This mode of knowledge sharing is based on the assumption that all knowledge is or can be made explicit in the form of documents or structured knowledge and databases. The problem of knowledge sharing and management is this reduced to representation, codification, and persistent storage of the documentary knowledge source in physical or electronic repositories, which can be searched, browsed, queried, filtered, and digested by any participant. The intranet, world wide web, and group-oriented document database and communication systems 11
such as Lotus Notes© have given considerable boost to this mode of knowledge sharing in view of the functionalities for globally distributed document sharing and the appealing hypertextual organization of the documents. This approach draws heavily on the system theoretical view of knowledge produced by symbol manipulation and processing.
Type II: Structural and Human Resource Mechanisms
A wide array of knowledge sharing devices fall within this category including crossfunctional teams, transfer and rotation of key personnel, apprenticeship, use of internal consultants, matrix organizational structure, structured face-to-face and electronic meetings, and structured task communication. Many of the processes described by Nonaka (1994) such as socialization, internalization, and externalization as well as brainstorming, and organizational development are also part of this type. Other related means for encouraging collaborative knowledge transfer include the learning organization and practice centers. The chief elements that is common to all these is the goal of contributing to a shared vision and language and the transfer of personal knowledge stored in the brainware to be available to a wider cross section of people. The thrust, however, is on the social processes and the continuous regeneration and expansion of tacit knowledge as a result of ongoing exchange among people and is informed by the pragmatic model of knowledge discussed in section 2.
Type III: Integrated Mechanisms for Facilitating Continuous Interpretation and Contextualization of Knowledge.
This mode of knowledge sharing attempts to synthesize the seemingly exclusive perspectives that underlie the first two categories. It is predicated on the assumption that the boundaries between tacit and explicit knowledge are fluid allowing for the possibility of a dynamic interplay between the two forms of knowledge. As well, these mechanisms underscore the need to reinterpret all knowledge on an ongoing fashion since the application of most knowledge has to be performed in newer contexts which have to be modelled and integrated to facilitate effective knowledge sharing and reuse. This model comes closest to Nonaka’s notions of shared experiences, crystallization (process through which various groups in an organization test the reality of a concept created by a self-organizing team), and context12
switching hypertext organization. This can potentially result in a dynamic knowledge cycle to continuously create, exploit, and accumulate organizational knowledge. The ndim system developed and tested by the ndim group at the Institute for Complex Engineered Systems at Carnegie Mellon University is a computer system designed with a similar vision of creating living repositories for sharing knowledge in the domains of engineering design and R&D (ndim group, 1995). Kleiner and Roth (1997) describe ‘learning histories’ as a new managerial tool that captures the lessons of successes and failures in an open and interpretive framework that also fits into this category of knowledge sharing and management. 4.0 Research Methods and Techniques The primary goal of our research is to empirically investigate processes and practices for knowledge sharing and management in large, diversified organizations in order to lead to the development of enhanced NICT-based solutions to support it. Our empirical work was guided by existing theoretical perspectives and previous field studies. It is our contention that systematic and focused field studies seeking to inquire into specific aspects of KSM can serve to advance our shared understanding of this important topic beyond anecdotal evidence and isolated case studies.
The fundamental research questions we addressed were the following: 1. What are the different mechanisms by which large multi-national corporations (MNCs) pool, exchange, and share knowledge? 2. What procedures are employed to contextualize, validate, verify, and authenticate the knowledge generated? 3. What is the role of information technology in supporting KSM? The field study was carried out at Du Pont, a large, diversified, global manufacturing company with its headquarters in the United States. Du Pont has world-wide operations and a strong commitment to understanding and sharing knowledge across the various business functions and units. The industries and markets that Du Pont serves range from traditional and specialized nylons, polymers, polyesters, and fibers to agriculture, nutrition, and pharmaceuticals organized under eight business segments and twenty business units. Through most of its history, Du Pont has seen itself as a science-based company with the mission of bringing the benefits of science to the marketplace in ways that benefit the shareholders and
13
society. Scientific and technological knowledge in chemical and material sciences and biological sciences is the basis for the company’s business portfolio.
The primary factor in the choice of Du Pont for the field study was the global and knowledge intensive nature of its operations. Besides the scientific orientation of the company, it has an explicitly stated strategy of seeking growth through what its current Chairman and CEO Chad Holliday calls “….knowledge-intensive projects that transform industries”. He defines “knowledge-intensive” as involving a complete understanding of “….all aspects of the consumer or business need and the dynamics of the existing value chain that meets this need. This will create higher value solutions for our customers in more productive and less capitalintensive ways” (Holliday, 1998). An equally important factor was the long Du Pont tradition of research cooperation with universities and the receptiveness to our proposal shown by senior Du Pont managers.
In order to keep the investigation focused and within manageable proportions, we restricted our data collection activities to knowledge in the areas of two functional areas (R&D and engineering project management - EPM), two business units (Engineering Polymers and Microcircuit Materials, a part of the iTECH business unit previously named Photopolymers and Electronic Materials), in four countries (USA, UK, Switzerland, and Japan). R&D and EPM are both highly knowledge-intensive domains that are at the core of Du Pont’s global operations. While R&D activities at Du Pont are performed at the corporate level as well as distributed at multiple sites in each of the business units, EPM is largely centralized at the corporate level. Both Microcircuit Materials (MCM) and Engineering Polymers (EP) have manufacturing, marketing, and R&D operations spread out across the Americas, Europe, and Asia.
The data gathering and analysis methods most appropriate to the questions we have raised are primarily qualitative and interpretive. Direct observation is clearly the best approach to investigating these problems where the process dynamics needs to be captured and the potential for multiple, conflicting interpretations can be expected. However, direct observation of the focal phenomenon can be extremely difficult and can impose an inordinate burden on the participating organizations. Accordingly, we settled for in-depth interviews with a cross section of professionals. By asking a common and structured set of questions 14
based on semi-structured questionnaires, we attempted to unravel the complex and sometimes implicit processes of knowledge sharing, their relative efficacy in different contexts, and the bottlenecks to effective sharing. It also enabled us to understand and interpret knowledge sharing in context, which is very important from our perspective. We also collected copies of documents that provided us with a more comprehensive portrait of KSM activities. This makes up, to some extent, for the limitations arising from not being able to investigate knowledge management issues longitudinally. The nature of this kind of interpretive data gathering required that we analyze, interpret, and reanalyze the data as it was collected.
All except eight of the 44 respondents who participated in the study were interviewed at their normal places of work. This was important because it gave them an opportunity to access and refer to documents, to demonstrate the system(s) they use, and to point to additional sources. Each interview lasted between 75 and 90 minutes and the responses were recorded by two of the researchers. One of them used the traditional linear narrative and the other employed the mind map technique by which related ideas and concepts were explicitly linked and organized together. The interview notes were compared, reconciled, and summarized at the end of each day of data gathering. A profile of the participants in the study is presented in Table 2. As can be seen the participants represent a cross section of professionals from corporate and business unit R&D and technical support groups in four countries and from Central Engineering. 57% of all respondents and 64% of all R&D respondents had Ph.D level qualifications. Their average number of years of experience at Du Pont was 17.6 with a range of 1 to 34 years.
Total Micro Circuit Materials Central R&D Total no. interviewed No. and % of PhD’s Average no. Years at Du Pont Range of Years at Du Pont 5 5 (100% 19.2 Central Engg 5 0 (0%) 23.6 Exp. Station 3 3 (100% ) 29.0 RTP Bristol (UK) 7 4 (57%) 11.3 Utsonomiya (Jap.) 5 0 (0%) 11.6 Engineering Polymers Meyrin (Swiss) 8 6 (75%) 13 Parkers burg (WVA) 1 1 (100%) 32 Exp. Station 1 1 (100% ) 21 44 25 (57% ) 17.6
9 5 (56% ) 21.0
11-2
1434
11-33
9-34
1-18
9-14
2-26
32
21
1-34
15
Table 2: Participant Profile
The nature of issues explored in the interviews dealt with: 1. The conceptions of knowledge entertained by the respondents. 2. The nature of the formal and informal mechanisms in place for the sharing and dissemination of knowledge. 3. The technology, architecture, design and implementation details of the IT-based systems in place for knowledge sharing and the respondents’ perceptions on the degree to which they are successful and their limitations. 4. Bottlenecks to effective sharing of knowledge across time and space. 5. Requirements for an (ideal) system that might address aspects of (4) above.
The final phase of the study involved revisiting the case data and gathering additional information from key informants to establish a set of requirements for the design of a knowledge management system. This work was largely based on field data obtained as part of the case study and our observations and understanding of the critical constraints in knowledge sharing. The requirements definition was carried out from a more generic perspective so as to enable us to develop a more general model of such a system with the goal of facilitating a dynamic and context-based support for knowledge sharing and management in contemporary organizations.
4.0 Case Study: Findings and Discussion 4.1 Background on the Business Units and the EPM Function Studied
As stated previously, the choice of Du Pont as the field study site was largely guided by the knowledge intensity and global scale of its operations, and its size and diversity. Hounshell and Smith (1998) have observed in their monumental study of the relationship between science and corporate strategy in Du Pont that the company was one of the handful of older enterprises with a long tradition of successful marriage between scientific knowledge and business, primarily through the institutionalization of industrial research and development. The distinguishing feature of these companies was that they were based on technologies that had emerged from scientific discoveries or developments further refined through a scientific approach to problem solving. Especially in the latter half of the 20th century, Du Pont had 16
become well-diversified in respect of both the range of product-markets it served and geography. By the 1960’s, it operated in 65 countries and nearly 50% of its revenues were generated outside the United States. Though a study of this type deals with certain unique aspects and features of Du Pont, we would expect to find resonance for its findings in similar diversified, knowledge-intensive firms. The issues, patterns, and processes employed in mobilizing, sharing, integrating, and managing knowledge and some of the possible relationships observed are unlikely to be widely disparate.
Our focus of the study was not on the knowledge content of R&D and EPM functions per se but on the mechanisms and processes for generating, sharing, and utilizing it. Given the distributed nature of R&D activities at Du Pont, personnel from the central R&D facility (Experimental Station, Wilmington, Delaware) were also included as participants in the study. Engineering is a centralized function at Du Pont and all the participating engineering personnel were from Central Engineering, one of Du Pont’s large, global service departments. We provide below a brief background on the activities of MCM and EP business units and the EPM function.
4.11 Microcircuit Materials (MCM)
MCM is one part of the iTECH business unit belonging to the Specialty Polymers business segment. Its products include thick film pastes, inks, and other high performance films for electronic circuit boards, display panels etc. It is considered a ‘high technology’ business and the customer base is technologically demanding and volatile. The industry groups served by MCM include some of the more technologically advanced sectors such as computer hardware, mobile telecommunications, avionics, military, and automotive. The rate of new product development is high with nearly 50% of the products having been introduced in the last 4-5 years. Much of the new product development work is driven by the customers whose needs and levels of sophistication tend to evolve rapidly. R&D personnel find themselves having to integrate knowledge concerning new materials and processes with customer- and applicationspecific knowledge to generate new products. Part of this knowledge has immediate or future relevance for the other regions served by MCM.
17
The organization of R&D in MCM consists of multiple layers. The more fundamental and longer range research activities such as the development of a new series of resistors or new metallic powders and other intermediates are performed at the central R&D facility (Experimental Station). Detailed development and scale-up for manufacturing is carried out at the business unit headquarters in Research Triangle Park (RTP), NC. The research facilities at Bristol, UK and Utsonomiya, Japan along with RTP conduct detailed product innovations and modifications typically by working closely with major customers. The primary manufacturing facility is in Puerto Rico with additional manufacturing for regional needs in Europe and the Asia-Pacific being catered to by facilities in Bristol and near Taipei, Taiwan respectively.
4.12 Engineering Polymers (EP)
EP business unit is part of the Performance Coating and Polymers business segment. Its product portfolio consists of a broad range of engineering materials for automotive, electrical/electronic, and industrial applications. Much of the R&D focus is on substitution of metal and rubber and finding new product applications using families of engineering resins such as Zytel© and Minlon© nylons, Delrin© acetal, Hytrel© thermoplastic elastomer etc. The headquarters of EP is at Parkersburg, WVA where much of the business and R&D coordination takes place. Basic research on new and existing resins is largely performed at the Experimental Station and Parkersburg. Development centers and marketing offices in different parts of the world work closely with other equipment manufacturers (OEMs) to improve existing and develop new applications of Du Pont products.
4.13 Engineering Project Management
Engineering is a centralized global service. In addition to building new plants and other facilities and renovating and retrofitting existing ones, Central Engineering works closely with R&D to design and install new and more efficient manufacturing processes. It also has staff groups responsible for all aspects of environmental and safety engineering. It develops and maintains procedures and standards for chemical process design, project implementation, and project safety management.
18
Du Pont has enjoyed a reputation for good project management capabilities in carrying out all aspects of project implementation from preliminary design and costing through detailed design and installation. This is supported by the favourable references in project performance reports produced by independent consultants benchmarked against industry standards. In recent years, however, Central Engineering has been subject to significant downsizing. A wide range of project life cycle activities including detailed design and implementation of projects are being outsourced to full service design and construction firms. Central Engineering’s role has, as a result, been confined to front-end loading, monitoring of project progress against standards, and staff functions. Du Pont has also undertaken several joint venture projects all over the world with other large companies. The number of employees in Central Engineering has declined from nearly 5000 to just below 1000 during the 1990’s.
4.2 Knowledge Sharing and Management (KSM): Corporate Level Initiatives
Given the knowledge-intensive nature of Du Pont’s operations and the critical role of proprietary technical information, development of appropriate procedures and systems for document management and codification have received significant attention during most of the last one hundred years of the company’s history. Many of Du Pont’s business units have their own information centers and libraries. Corporate Information Science is responsible for the most prominent of these, the Lavoisier Library at the Experimental Station and the information centers that serve the Corporate R&D (CRD). Information centers provide access to copies of all relevant documents, memoranda, reports, and items of correspondence could be accessed. The advent of computer systems and electronic storage and communication has resulted in the provision of online access to the information in the library and the information centers through the use of Du Pont’s online library catalog and the databases for retrieving Du Pont’s research and engineering documentation. As a consequence, some of the distributed information centers have been merged or eliminated.
Initiatives aimed at the design, building, and implementation of integrated database systems for the storage, retrieval, and distribution of Du Pont scientific and technical information go back to the 1960’s. These resulted in the development of some of the fundamental principles for efficient representation and easy search of complex molecular structures, a system for polymer nomenclature, and a hierarchical thesaurus for indexing technical reports, engineering 19
drawings, and patent documents. The development of the DuPont SCION (Scientific Information Online) proprietary database in the 1990’s was greatly facilitated by these initiatives (Schultz and Wilks, 1997; Kvalnes, 1997). Technology Conferences sponsored by the Vice President of Research provide useful forums for active networking and sharing of tacit knowledge.
Central Engineering has a long tradition of establishing standards and codification of useful knowledge and practices. A vast array of manuals and guides on topics ranging from project implementation, safety practices, operating policies, training records, and plant maintenance etc. have been developed over the years. Some of these have become partly obsolete as a result of outsourcing and joint venture arrangements that present new challenges for codification and updating of procedural knowledge. Du Pont’s Chemical Process Design Standards cover a wide range spectrum including fluid flow, heat transfer, dust and mist collection, agitation, filtration, distillation, reactor design, construction materials, physical properties etc. They have been developed over 50 years of research and practice and run into four volumes of more than 1000 pages. Sharing hard copies of these documents along with their associated mathematical models and materials properties databases to assist Du Pont engineers, contractors, vendors, and joint venture partners has always been problematic. Considerable progress has been made in Du Pont’s effort to migrate much of this to the World Wide Web (corporate intranet). This is beginning to ease the difficulties in sharing the resources and permits the relatively straightforward updating of the information (Scrivner, 1997).
At the corporate level, the KSM-related responsibilities fall under the purview of Corporate Information Science (CIS) augmented by the activities of the Communication, Collaboration, and Coordination Program Office (C3PO) and the distributed libraries and information centers. The former was founded in 1987 and is responsible for managing Du Pont’s proprietary information, competitive intelligence function, and the Lavoisier Library. It also provides a range of information services to the businesses including ongoing codification of enduring knowledge in the businesses. The distributed libraries and information centers are managed by the individual business units. C3Po spearheads the Communication, Collaboration, Coordination project which includes the company-wide rollout of Lotus Notes© and related software with the objective of providing the architecture for collaboration in Du Pont. This is 20
a 4-5 year project that started in 1998, at the end of which the group hopes to standardize much of the communication and collaboration work around Lotus Notes tools for email, document databases, workflow systems, screen sharing, and calendaring. All knowledge workers are being encouraged to run Lotus Notes clients on their workstations to connect to Du Pont’s network of 90 Lotus Notes servers and 450 other intranet servers. Currently there are nearly 60,000 Lotus Notes clients globally but this number is expected to grow as support for other legacy tools and systems are withdrawn. It is planned that the server network will be gradually extended to an extranet to include relevant external entities such as vendors, contractors, and joint venture partners once the security concerns have been resolved. The C3PO project also facilitates the nurturing of virtual communities of practice with shared interests and expertise, workflow systems, electronic laboratory notebooks, codification of tacit knowledge and prowess, and information technology (IT)-assisted learning.
4.3 Specific Observations and Findings 4.31 Type I Mechanisms and Explicit Knowledge
Type I Mechanisms are the central focus of the ‘knowledge management’ movement and the projects associated with it. The underlying strategy is one of explicit documentation, filtration, and the attempt to systematize, codify, and reuse knowledge to the extent possible. Computer technology in its various forms is used to facilitate selective sharing of the documents and repositories.
As indicated previously, Du Pont has a long tradition of relatively successful deployment of such mechanisms. The migration of many of the document and knowledge bases to the web is progressing steadily. The C3PO and CIS initiatives discussed earlier is aimed at ensuring greater standardization and company-wide dispersion of such mechanisms. We present below some of the significant issues related to their use and perceived effectiveness.
Information overload is a constant refrain, especially among the R&D personnel we interviewed. It appears that the amount of information that is ‘pushed’ at people through electronic mail, document attachments and databases in addition to the physical means is much higher than what can be meaningfully processed by most in the time available. The filtration and search capabilities of Lotus Notes document databases and systems are far too primitive at 21
this stage in relation to the need. One of the respondents compared these to dumpsters in which one had to forage hard to find something useful, if any. The time spent scanning large volumes of marginally relevant information also came in the way of the critical reflection needed to summarize and present in succinct form, their own work for the benefit of others.
Diversity of media and the lack of integration present challenges for effective knowledge processing. Much of the information needed by the respondents is already in electronic form but it is fragmented across a variety of incompatible computer systems and databases. In addition, each knowledge professional typically inherits one or two filing cabinets containing hard copies of correspondence, memos, reports etc. when they begin a new assignment. The differences in the indexing schemes and search mechanisms for the different document bases impede routine activities such as finding useful fragments of information and synthesizing them in the context of particular tasks.
There are critical issues with respect to archival practices that bear on knowledge sharing and management. The traditional, paper-based systems appear to be withering while a robust and company-wide electronic regime is yet to emerge. The ongoing changes in computer hardware, operating systems platforms, data formats, and software versions tend to limit the archival value of electronic documents. The old adage of “if only a company knew everything it knows” could be paraphrased as “ if only a company ‘knew’ everything that is buried in inaccessible computer systems”.
Repositories are available at different levels in the Du Pont units we studied. There are individual repositories and databases that hold much of the local knowledge that is generated. There are also repositories at the level of small groups of people who work together on the same project or with one or a group of customers. The groups need not always be co-located or working synchronously. Finally there are business unit or corporate-level information bases that are maintained. It is inevitable that there will be some duplication and inconsistency across the three levels of repositories. More importantly, the limited general access and the terminological and linguistic differences across these levels hinder the potential for knowledge sharing. As we shall see later, critical human interlocutors emerge at the interfaces between the levels to ameliorate the problem.
22
Several commentators have emphasized the importance of naturally occurring local learning and innovation that takes place at all levels of an organization (for example, Brown, 1991). A related issue is one of facilitating the wider sharing of relevant aspects of such local knowledge through appropriate organizational processes and systems. We found in the course our study that in many situations, such knowledge is embedded in organically evolving and bootlegged systems quite independent of the formal mechanisms available for systematizing and sharing. The reasons for the emergence of such local mechanisms are varied and include: • • The need for specialized structures and capabilities to represent the knowledge. Lack of timely support and assistance for the use of standardized tools and mechanisms. • Idiosyncratic attachment to particular computer tools and legacy systems by some of the employees. In any case, the need to track, evaluate, support, and selectively integrate such local and bottom-up systems cannot be overstated. These are useful initiatives arising from the need to formalize and disseminate localized knowledge with potentially wider implications that arise as a by-product of routine operations. We were able to document a number of these at both MCM and EP though the preponderance of such initiatives at MCM is worth noting.
The Central Engineering group at Du Pont has traditionally been the pioneers in the use of electronic collaboration technology. This is not particularly surprising given the global distribution of project work and the need for rapid, multi-directional knowledge flow among Du Pont engineers, vendors, external contractors, and joint venture partners. As one of the respondents put it, it is a trade-off between getting to work remotely with the collaboration tools or being away from family for long durations. Also, a great deal of effort has been expended in the past aimed at codifying and standardizing a large part of the process knowledge generated over the past 50 years of project work. We observed a preponderance of type I mechanisms being successfully deployed and used in EPM. More importantly, systematized knowledge in the areas of systems for plant safety (design and monitoring) and environmental pollution prevention have become saleable commodities in the form of consulting packages. There is growing demand for such packaged know-how especially in the Asia-Pacific region.
There are differences between the Du Pont units in the modes of structuring and organizing knowledge. For instance, MCM operates in a highly dynamic market place in which rapid 23
learning from the customers and adapting to their changing needs are normative. This has led to the organization of MCM repositories to be primarily based on products, customer segments, and application areas and only secondarily on materials. This is despite the fact that the training and specialization of the scientists and technical support personnel is by materials. In contrast, the business strategy of EP is driven by material push and the knowledge organization at EP is primarily along the material dimension (such as Hytrel, Minlon, Delrin etc.) and within each by major customer groups and OEMs. The need for multiple and divergent classification schemes and indexing systems is a routine aspect of knowledge processing. This has important implications for the design, implementation, use, and maintenance of company-wide KSM solutions with their implicit assumptions regarding the relative homogeneity of knowledge representation schemes.
The resources needed to carry out the difficult and unrewarding task of combining tacit and often fragmented forms of explicit knowledge into systematized and reusable knowledge and to make it publicly available is considerable. While Central Engineering has been successful at this for some time, the relationship between the availability of organizational slack and the group’s ability to standardize and deliver such knowledge is worth exploring. One of the issues arising from the massive outsourcing and downsizing at Central Engineering relates to the reduction in slack and its effect on codification work. There are also concerns about continual erosion of the knowledge base with the department having to rehire ex-Du Ponters and to hire employees from contracting firms to stem the erosion.
4.32 Type II Mechanisms and Tacit Knowledge
It is certainly the case that knowledge workers consult a variety of sources of explicit knowledge and fall back on their learning from formal study and training. However, in the course of actually doing their jobs, most of the learning arises from engaging with real problems, gathering a range of relevant information from diverse sources, and discussing key issues with colleagues and other professionals. Most of the R&D respondents in particular have a strong network of people spread out all over the world that they can turn to for consultation and guidance. A surprisingly large number of such contacts are from outside their own units in other business units to departments in Du Pont or from external entities such as vendors, customers, contractors, universities, and joint venture partners. Some of the 24
respondents can count upwards of 300 people in their network and the intensity of interaction varies over time depending on the tasks on hand. These are typically bottom-up formations and successful R&D scientists spend considerable time and energy developing and nurturing such networks. The diffusion of tacit knowledge in particular through these social networks is quick. The track record of the informants in terms of both the reliability of the acquired knowledge and trust largely determines the length of such associations. Informal protocols of reciprocity also feature in the assessment of track record.
Du Pont has long recognized and promoted mechanisms for such networking. In-company technology conferences and periodic seminars involving people from diverse businesses and technology areas provide facilitative forums for making initial contacts and extending one’s networks. Other processes include apprenticeship training for new scientists under wellestablished R&D personnel as well as the rotation of the latter through newer research sites. The Utsonomiya (Japan) research facility in MCM was established and developed primarily through people rotation and active apprentice-type training programs. In addition, short visits by scientists either to the Experimental Station or to other research sites within the business units are seen as useful mechanisms for the sharing of less tangible aspects of knowledge. The reliance on these mechanisms based on “high bandwidth” channels was more pronounced in R&D as compared to EPM and within R&D, and greater in MCM in comparison with EP. This is consistent with the general observation that higher the role of tacit knowledge, greater the reliance of type II mechanisms.
There is some concern in the R&D groups that the increasing dependence on electronic communication and coordination and the reduced opportunities for face-to-face contact are causing a slow erosion in this mode of knowledge sharing. One of the requirements for any social network to remain active and to be effective in knowledge sharing is the periodic opportunity to ‘catch-up’ through face-to-face meetings, albeit infrequently. Such meetings enable the establishment and/or reinforcement of a common ground needed to cultivate and develop the trust and reciprocity implicit in such relationships. When the operations are globally dispersed as they are in MCM or EP, some of the faster and less expensive mechanisms such as telephone, email, net meeting and even videoconferencing are constrained by linguistic, cultural, and time-zone related barriers. Sustaining such networks involves expensive travel and movement. The curtailment of resources for this purpose in recent years 25
and the increasingly rigorous quarterly financial reporting systems are putting pressure the quality and vitality of such networks. This is particularly hard on the younger scientists who are in the process of developing their personal networks and have to rely on remote, electronic mechanisms to a very large extent.
Human agents were sought for a variety of reasons. Probably the strongest one that came up repeatedly in our interviews was the perception that it resulted in significant efficiencies and time savings. An illustration from an Engineering Polymers R&D center at Meyrin, Switzerland is illustrative. One of the laboratories there provides analytical support and carries out a wide range of routine and specialized tests on properties of new materials, identification of additives etc. When a scientist approaches the supervisor who has managed this laboratory for the past 15 years, he can give good and immediate answers to questions such as whether the same test has been carried out previously and if not whether the results of a closely matched previous test can be extrapolated for a particular problem that a scientist is trying to solve. The information needed to extract the knowledge is available in the form of a database of previous tests carried out and the results but this may not have the format and flexibility requirements that the scientist needs. Searching through this database is also timeconsuming. Besides, the effort and difficulty involved in externalizing the tacit dimension of subtle matches with previous tests is non-trivial. In performing even routine research tasks such as interpreting the results of certain complex analytical tests, sitting down with experts who can help with the interpretation saves much time and reduces some of the guesswork.
Arising from a combination of historical factors, Du Pont R&D is faced with a bimodal distribution age distribution of its personnel. The age cohort in the 50-plus range is very large, in part due to the expansionary spiral of the 1950s through early 70’s. The 40-50 age cohort is small but there is a small increase in numbers in the 25-40 age group. This introduces vulnerabilities in the form of potential erosion of tacit knowledge base in the absence of tacitto-tacit transfer and tacit-to-explicit conversion of knowledge. Such programs are resourceintensive and can only succeed in a climate of trust and security. Also, they need to be guided by sound judgment on the future value of different bodies of tacit knowledge in relation to the company’s strategic direction and the emerging trends in the relevant product-market environments.
26
The phased implementation of the integrated electronic coordination architecture and the C3PO project has generated incentives and pressures to make a transition to new ways of collaborating and knowledge sharing. For some aspects of knowledge work this is only a matter of adapting to a new set of computer tools and they complement other modes. However, the extent to which this regime can or need to supplant the existing, organic and informal mechanisms of tacit knowledge sharing is unclear at this stage. Many of the respondents do make a clear distinction between electronic collaboration and the sharing of knowledge.
4.33 Type III Mechanisms and Knowledge
Type III mechanisms refer to organizational processes and system capabilities that facilitate an active and dynamic interplay between tacit and explicit forms of knowledge. They attempt to integrate the diverse forms of available knowledge and highlight the need for ongoing and context-based interpretation of knowledge. Further, these mechanisms explicitly accommodate emergent and serendipitous knowledge that is at the root of innovative product and process development work in organizations. Creating the conditions and designing processes and systems to support such synthetic knowledge creation and sharing activities pose a major challenge.
There is a clear understanding in Du Pont of the limits of the electronic communication, collaboration, and coordination systems from a knowledge processing perspective. These systems have laid the foundation for communication, information sharing, and some amount of collaborative work across distances. It has also provided the means for more efficient and wider yet selective dissemination of documents and codified knowledge. However, the recognition that people and their networks have to be effectively and seamlessly interwoven with the computer networks and systems has triggered the search for new approaches.
Even the ardent proponents of electronic communication and coordination concede that knowledge management can only be roughly 30% based on the systems being implemented and the rest has to be based on people. Accordingly, a campaign to facilitate the evolution of communities of practice through the use of the communication and collaboration systems has been launched. These communities are largely organic formations comprising members with 27
specialized skills or know-how. They may rely on technology to identify potential members and establish communities to by sharing useful information, offering guidance and critique, recounting related experiences etc. There are currently over a dozen such communities (such as winding experts, people with extrusion skills) in existence at Du Pont and they function like virtual guilds. A project to inventory and assess the effectiveness of such on-line communities is currently in progress.
There seems to be roughly three stages these communities go through as they evolve to the point when they can function without the need for coordinators or champions. The first phase consists essentially of the community members coming together and establishing a common ground. Potential members get to know each other personally and professionally, and engage in limited information sharing. The tools used are mainly e-mail, telephone, net meeting, and screen or document sharing but face-to-face meeting can serve to speed up this process. Some of the communities do not progress beyond this stage in the absence of common problems or projects or in some cases, inspired leadership. Under favorable conditions, a community might move to second phase of being able to engage in effective collaborative work. In this stage, the members have worked out their semantic and terminological differences and have some knowledge regarding the capabilities and competences of each other. The group identity begins to get forged by working collaboratively on a common project or problem. The passage through these two phases is a precondition for the knowledge sharing phase in which adequate levels of trust and confidence in each other have been established and free and open sharing of all forms of knowledge begins to take place.
Innovation-oriented knowledge processing involving new product and process development necessitates the integration of vast amounts of explicit knowledge with certain unique and deeply personal insights acquired through direct practical experiences. The validity and veracity of aspects of the new knowledge being synthesized need to be tested and established through ongoing experimentation and consultations with domain experts. Much of the explicit knowledge available needs to interpreted in the new context. We have documented a number of cases in which both routine and radical innovations were produced by combining different types of knowledge and making leaps that cut across traditional boundaries drawn by existing technologies, business units, and even disciplines. Once the feasibility of the initial idea is established, detailed experiments and trials, and tests can follow. 28
The Bio3GT program that led to the development of a biological process for the manufacture of 3GT fibre is one such example. It involved considerable cross-fertilization of knowledge between the polymeric materials group and the Life Sciences business segment. This has the potential to radically lower the unit cost of production of the particular commodity polymer. The project idea came together at a time when the cost reduction using the more traditional methods like discovering new catalysts could not possibly attain the level needed to compete globally with low cost producers in the newly industrialized economies. Some of the tacit knowledge that is brought to bear in such knowledge processing is acquired through years of variegated experiences in functions as diverse as plant operations, product management, and venture business management. These experiences provide an appreciation for making technological advances and choices that adapt to and anticipate market conditions. Another example of recontextualization of previously acquired knowledge in MCM involves the development of a new process for flat panel displays on LCD screens that leverages available knowledge from the textile printing business unit.
One of the main aspects of the integrated knowledge processing described in the foregoing is the pivotal contributions by key knowledge practitioners who typically combine extensive amount of tacit knowledge with other knowledge. We were able to identify four such knowledge practitioner roles that typically operate at the interfaces between disciplines, technologies, business units, functions, and businesses and customers. Each is described below. 1. High Level Synthesizer: These are typically high level technology or R&D managers with a wide range of experiences across several business units and in several operational and functional areas. They are adept at environmental scanning and closely monitor the trajectories of various technologies. They operate at the interfaces between different technology areas and business units. They develop and maintain a vast network of formal and informal knowledge sources. They look for opportunities arising from various combinations of different types of knowledge.
2. Librarian: Librarians operate at the interfaces between materials, processes and product markets. By virtue of their ability to painstakingly gather, assimilate, index, and store copious amounts of information pertaining to material properties, costs, customer 29
requirements, and changes to the requirements over time, they are able to provide timely assistance and service to other knowledge workers. They deal primarily with explicit knowledge. It takes nearly 10-15 years’ experience to grow into this knowledge practitioner role.
3. Knowledge Engineer: Knowledge engineers operate at the classic interface between R&D, marketing, and customers. They work closely with the customers to adapt existing products and to develop new products and applications. Some of the work might appear rather routine but the tacit dimension of knowledge is in their assessment of customers’ current and future needs. This knowledge has to be integrated with detailed and explicit knowledge concerning the company’s products, processes, and technologies.
4. Knowledge Operators: Knowledge operators are typically technical or customer support personnel who are usually front-line employees located close to the operational realities of the business and the market place. They accumulate and transmit tacit knowledge in the form of embodied skills (Nonaka and Takeuchi, 1995). They tend to work at the interfaces between R&D and manufacturing or R&D and customer operations. They work very closely with knowledge engineers.
Table 3 provides an outline of the different knowledge practitioner roles, the interfaces they operate at, and the types of knowledge dealt with.
30
Roles
Interfaces
Knowledge Operation
High Level Synthesizer Librarian
Cross Discipline/ SBUs/functions Materials/ Product Customer Marketing/R & D
Explicit and Tacit
Mostly Explicit
Knowledge Engineers
Explicit and Tacit
Knowledge Customer-Product/ Mostly Tacit Operators Operations Table 3: Knowledge Practitioner Roles, Relevant Interfaces and Operations
5.0 Toward System Requirements for Enhanced Support for KSM
Based on the foregoing discussion and the interviewee responses, we sketch below a set of systems and corporate policy requirements that can contribute to the goal of more integrated knowledge processing in organizations. The term ‘system’ in this context has to be interpreted broadly to accommodate the symbiotic relationship between physical repositories, computer tools and databases at one end and the cognitive and social processes involving human actors at the other.
a. Organization of Documents Documents of all types and stored on different media are used in knowledge work. However, there is a critical need for structuring and viewing the information content in such document spaces from different perspectives provided by disciplines, business units, market conditions, and individual preferences. No single classification structure can accommodate this diversity of needs over space and time. Moreover, such structures need to evolve in a bottom-up fashion. This was very evident at every level of information and knowledge processing by the participants in the study. This requirement can be addressed through the provision of the following features: 31
-
Development of an integrated system for classification and indexing of documentation. Ease of creation of personal but shareable repositories that use multiple classifications. Ability to search across multiple repositories.
These features are interdependent as the evolution of a rich classification system is facilitated by the ease with which personal yet shareable repositories can be created. Such a classification system that can handle the diversity of perspectives across the organization is needed to address the problem of searching across multiple fragmented document repositories. b. Ameliorating the effects of information overload Information overload is one of the most commonly cited problems. Its causes are many fold, including multiple, disparate databases and the absence of personal control over much of the information that is ‘pushed’ at people. The disorientation that arises from the difficulties people experience as a result of such over-abundance of information, much of which may be irrelevant is well documented. The following suggestions are offered in this connection: Documents to be made available at different levels of detail with the option to view any or all (title, cover page with meta-data (author, date, subject, keywords etc.), abstract only, abstract plus body of the document) Facilitating optimal awareness: This can be achieved through the design of solutions that experimentally model the information needs of users and notifies each one as newer and relevant documents become available, say on an intranet server or a document database. The users are thus in a position to ‘pull’ relevant information and view it at the required level of detail.
c. Dealing with the problem of information seek-provide asymmetry Seek-provide asymmetry refers to the commonly observed problem in knowledge sharing and management systems of the disparity between a person’s willingness to access and use the information or knowledge contributed by others and the contributions that s/he is prepared to make to the shared pool. Such asymmetry occurs at the level of contributing keywords for documents contributed as well as for generating the documents or databases themselves. There is a clear need to develop effective metrics for measuring the knowledge contributions of individuals. Furthermore, organizational policies need to evolve to encourage provision and sharing of good content that can in turn trigger further knowledge accretion.
32
d. Management of Language Management of language is an often overlooked issue. However, management of language is critical to organizing, searching, systematizing, and sharing of knowledge in any organization. By language we refer to all lexical, terminological, semantic, and linguistic differences that can be a major impediment to good communication and effective knowledge sharing in all large, global companies. Language represents the common ground that is defined by a set of people in the context of their interactions. This is a shifting common ground depending on people, the nature of the market, and technological changes. Documenting, clarifying, and establishing correspondences in the diverse uses of the language on an ongoing basis and the design of tools for supporting this need are key activities that deserve consideration.
e. Supporting the formation and growth of people networks and the evolution of communities of practice. The key requirement for building such networks is the ability to identify and learn more about people who can provide particular kinds of information or assistance. Matching the needs with the profiles of people is not an easy task. With the increasing reliance on electronic means of collaboration, organizations have initiated electronic people finders and other databases to supplement the paper-based directories. However, the benefits from such resources are often stymied by the absence of critical information such as history of previous project and country experiences. As well, these resources tend to become obsolete over a short period. User ownership and control over the information in the people finder databases can go a long way to maintaining its currency and value. An important requirement to facilitate the development of people networks is the fostering of user-owned and user-updated people finders and profiles backed by organizational policies for its maintenance.
f. Methods and policies for the integration of organically evolving systems. An important observation in our study was the emergence of organically evolving systems. This runs counter to the traditional systems development model based on architectures that are essentially top-down and based on global priorities of the business unit or the firm. However, bottom-up systems and tools that address typically immediate and local needs may embed knowledge that extend beyond the location of creation. When the knowledge pertaining to such an effort is not shared, there is always the risk of it being reinvented at other locations or business units. One potentially useful direction is provided by the distributed, open-source 33
software model augmented with a community tools repository with which people can develop, access, and share such knowledge. In any case, a clear policy for dealing with organically grown tools is critical to leverage effort and knowledge across geographically distributed centers.
g. Long term electronic archiving. As noted earlier, the transition from traditional paper based systems to digital systems has led to the disintegration of a well-structured archival process. In Du Pont this practice was well structured and organized in the paper world which allowed for the systematization of a number of distinct bodies of knowledge. However, no new systematic methods have been developed to address the archival problem in the electronic arena. The design of systems for knowledge sharing and management needs to address this issue from the outset. Failure to do so risks losing large amounts of unrecoverable information and knowledge.
h. Individual and group rewards The design of reward systems for contributing to and utilizing the shared knowledge repositories is one of the more difficult issues that organizations are faced with. Such systems need to explicitly acknowledge the group nature of much of the knowledge production. Both individuals and groups (including communities of practice) will have to be treated as distinct entities and rewarded accordingly. Also, both the amount and the degree of use of the
contributed knowledge have to be factored into assessments of knowledge professionals’ performance.
6.0 Conclusion
This paper contributes to the growing literature on the organizational processes and mechanisms for knowledge sharing and management especially in large, global companies. By drawing on and synthesizing related bodies of writing, we have attempted to further the debate on what constitutes organizational knowledge. The diverse forms in which such knowledge manifests leads to a review of some of mechanisms and systems for knowledge sharing and management in contemporary organizations and the development and elaboration of a typology of such mechanisms. A detailed field study exploring the major knowledge sharing issues, practices, constraints, and mechanisms was carried out in selected departments 34
and business units of a large, multinational company with knowledge-intensive operations. The field study lends qualified support to the proposition that exclusive focus on certain types of knowledge can be counter-productive. Knowledge sharing and management systems and processes in large global companies need to be integrative and flexible enough to facilitate the dynamic interplay between different forms of knowledge. A set of system requirements targeted to enhance these capabilities is outlined.
References Antonelli, C. (1996), “Localized Knowledge Production Processes and Information Networks”, Journal of Evolutionary Economics, 6, pp. 281-296. Antonelli, C. (1999), “The Evolution of the Industrial Organization of the Production of Knowledge”, Cambridge Journal of Economics, 23, pp. 243-260. Argyris, A. and Schon, D. (1978), Organizational Learning, Reading, MA: Addison-Wesley. Arrow, K. (1969), “Classificatory Notes on the Production and Transmission of Technical Knowledge”, American Economic Review, 59, pp.29-35. Arrow, K. (1994), “Methodological Individualism and Social Knowledge”, American Economic Review, 84, pp.1-9. Ashton, W.B. and Stacey G.S.(1995), “Technical Intelligence in Business: Understanding Technology Threats and Opportunities”, Int. J. of Technology Management, 10(1). Bartlett, A.B., and Ghoshal, S. (1989), Managing Across Borders: The Transnational Solution, Boston, MA : Harvard Business School Press. Brown, J.S. (1991), “Research that Reinvents the Corporation”, Harvard Business Review, Jan.-Feb. Brown J.S. and Duguid, P. (1991), “Organizational Learning and Communities of Practice: Toward a Unified View of Working, Learning, and Organization”, Organization Science, 2(1), pp.40-57. Brown S.L. and Eisenhardt K.M.(1995), “Product Development: Past Research, Present Findings, and Future Directions”, Academy of Management Review, 20(2), pp.343-378. Clark, K.B. and Fujimoto, T. (1991), Product Development Performance, Boston, MA: Harvard Business School Press. Cohendet, P., Kern, F., Mehmanpazir, B., and Munier, F. (1999), “Knowledge Coordination, Competence Creation, and Integrated Networks in Globalized Firms”, Cambridge Journal of Economics, 23, pp.223-241.
35
Coyne, R. (1997), “Language, Space, and Information”, In Droege, P. (ed.), Intelligent Environments, Amsterdam: Elsevier, pp.495-516. Culpan, R, (1983), Multinational Strategic Alliances, Binghampton, NY: Howarth Press Inc. David, P.A., (1993), “Knowledge Property and the System Dynamics of Technological Change”, Proceedings of the World Bank Annual Conference on Development Economics, Washington DC: The World Bank. Davenport, T.H., De Long D.H., and Beers, M.C. (1998), “Successful Knowledge Management Projects”, Sloan Management Review, 39(2), pp.43-57. Dewey, J. (1929), The Quest for Certainty, New York: G. P. Putnam. Doz, Y., and Prahalad, C.K.(1991), “Managing MNCs: A Search for a New Paradigm”, Strategic Management Journal, 12(5), pp.145-164. Dretske, F. (1981), Knowledge and the Flow of Information, Cambridge, MA: MIT Press. Edstrom, A., and Galbraith, J. (1977), “Transfer of Managers as a Coordination and Control Strategy in Multinational Organizations”, ASQ, 22(2), pp.248-263. Eliasson,G., Foster, S., Lindberg,T., Pousette,T., and Taymaz, E. (1990), The Knowledge Based Information Economy, Stockholm : The Industrial Institute for Economic and Social Research. Galbraith, J.R. (1974), “Organization Design: An Information Processing View”, Interfaces, 4(3), pp.30-36. Goldman, A.I. (1999), Knowledge in a Social World, Oxford : Oxford University Press. Holliday, C. (1998), “To Our Shareholders”, Du Pont 1998 Annual Report, p.3. Hounshell, D.A. and Smith J.K, Jr. (1988), Science and Corporate Strategy: Du Pont R&D, 1902-1980, Cambridge: Cambridge University Press. Inkpen, A.C., (1998), “The Management of Knowledge in International Alliances: The Role of Collaborative Process”, Carnegie Bosch Institute Working Paper, available athttp://cbi.gsia.cmu.edu/work/inkpen.htm. Jain,R.K., and Triandis, H.C.(1990), Management of R&D Organizations : Managing the Unmanageable, New York: Wiley and Sons. Jorde, T.M., and Teece, D.J.,(1990), “Innovation and Cooperation: Implications for AntiTrust”, Journal of Economic Perspectives, 4, pp.75-96. Kerssens-Van Drongelen, I.C., de Weerde-Nederhof, P.C., Fissher, O.A.M.. (1996), “Describing the Issues of Knowledge Management in R&D: Toward a Communications and Analysis Tool”, R&D Management, 26(3), pp.213-230. 36
Kleiner, A. and Roth, G. (1997), “How to Make Experience Your Company’s Best Teacher”, HBR, Sept-Oct., p.172-178. Kogut, B. and Zander, U. (1992a), Knowledge of the Firm, Combinative Capabilities, and the Replication of Technology, Organizational Science, (3), pp. 383-397. Kogut, B. and Zander, U. (1992b), Knowledge of the Firm, Technology Transfer, and the Theory of the Multinational Corporation, Working Paper, Institute of International Business, Stockholm School of Economics, December. Kumar, K. (1990), “Post-Implementation Evaluation of Computer-Based Information Systems: Current Practices”, Communications of the ACM, 34(2). Kvalnes, F.H. (1997), “History of Managing Technical Information at Du Pont”, Internal Du Pont Note. Machlup, F. (1980), Knowledge: Its Creation, Distribution, and Economic Significance, Princeton, NJ: Princeton University Press. Manheim, M.L. (1992), “Global Information Technology: Issues and Opportunities”, International Information Systems, 1: pp.38-67. March, J.G. and Simon, H.A. (1959), Organizations, New York: John Wiley. Ndim group (1995), “An Environment for Realizing Computer Supported Collaborative Work”, Technical Report Number 05-93-95, Engineering Design Research Center, Carnegie Mellon University. Nelson R.R. and Winter S.G. (1982), An Evolutionary Theory of Economic Change, Cambridge, MA: Harvard University Press. Nonaka, I. (1991), “The Knowledge Creating Company”, HBR, Nov-Dec., pp.96-104. Nonaka, I.(1994), “Dynamic Theory of Organizational Knowledge Creation”, Organization Science, 5(1), pp.14-35. Nonaka, I. And Takeuchi, H.(1995), The Knowledge Creating Company, New York: Oxford University Press. Polanyi, M. (1967), The Tacit Dimension, Garden City, NY: Doubleday Anchor. Prokesch, S.E., “Unleashing the Power of Learning: An Interview with British Petroleum’s John Browne”, HBR, Sept-Oct., pp.146-172. Quine, W.O. (1953), “Two Dogmas of Empiricism”, in, From a Logical Point of View, Cambridge, MA: Harvard University Press.
37
Reddy, M. (1979), “The Conduit Metaphor: A Case of Frame Conflict in our Language About Language”, In Ortony, A.(ed.), Metaphor and Thought, Cambridge, UK: Cambridge University Press, pp. 284-324. Schultz, J.L. and Wilks, E.S. (1997), “Multiple-Level Polymer Registration in the Du Pont SCION Database, 37(2), pp. 165-170. Scrivner, N.C. (1997), Du Pont Makes Engineering Standards Interactive and Intranet-Ready, Scientific Computing and Automation, October, pp. 31-33. Simon, H.A. (1997), The New Science of Management Decision (Revised ed.), Englewood Cliffs, NJ: Prentice Hall. Simon, H.A. (1996), “Knowledge and the Time to Attend to it”, Paper presented at the Carnegie Bosch Institute International Conference on High Performance Global Companies, April 21, 1996, Boca Raton, FL, available athttp://cbi.gsia.cmu.edu/work/96-2.htm Souder, W.E., Nashar, A.S., and Padmanabhan, V. (1990), ‘A Guide to the Best Technology Transfer Practices”, Technology Transfer, Winter-Spring. Takeuchi, H., and Nonaka, I. (1986), “The New Product Development Game”, HBR, Jan-Feb. Van Fenema , P. (1997), “Management of Globally Distributed Projects : Experiences form the GOLDD Team”, Proceedings of the International Conference on Information Systems, Atlanta, GA, pp.474-475. Wheelwright S.C., and Clark, K.B.(1992), Revolutionizing Product Development, New York: Free Press. Zeleny, M., Cornet, R.J., and Stoner, J.A.F. (1990), “Moving form the Age of Specialization to the Era of Integration”, Human Systems Management, Vol.9, pp.153-170.
Associate Professor Joseph G. Davis Joseph Davis is Associate Professor of Information Systems and Coordinator of the Decision Systems Laboratory at the University of Wollongong, Wollongong, NSW, Australia and a Visiting Research Scientist at the Institute for Complex Engineered Systems, Carnegie Mellon University, Pittsburgh. He has previously served on the Information Systems faculties at The University of Auckland, Auckland, New Zealand and Indiana University, Bloomington, IN. He has also been a Visiting Professor at the Czech Management Centre, Celakovice, Czech Republic (1996), Katz Graduate School of Business, University of Pittsburgh (1994-95) and Visiting Researcher at the Engineering Design Research Center in 1990-91, 1994-95. He has a Ph.D in Information Systems and Strategic Management from the Graduate School of Business, University of Pittsburgh and an MBA from the Indian Institute of Management, Ahmedabad, India, prior to which he completed his Bachelor of Science in Mathematics.
38
Joseph’s research interests span information modelling and management and the design and implementation complex application systems to support managerial decisions and engineering design. His research has been published in some of the major journals and presented in the important conferences in the field. He has served on the Program Committees of several important international conferences including the Internationl Conference on Information Systems (ICIS). He has actively collaborated with the ndim group on research into the design, implementation, deployment, and testing of computer-based environments to support design work. His research is currently supported by grants from Carnegie Bosch Institue for International Management and the Australian Research Council under the Strategic Partnership: Industry, Research, and Training (SPIRT) scheme in collaboration with BHP Information Technology Division, Sydney.
Dr Eswaran Subrahmanian Eswaran Subrahmanain is a Principal Research Scientist at the Institute for Complex Engineered Systems (ICES)at Carnegie Mellon. He has a Ph.D. in Urban and Public Affairs with emphasis on Information Systems from CMU. He also has a B.Tech. (Hons) in Chemical Engineering and a M.S. in Computer Science. He has been at ICES and its precursor EDRC (the NSF center for engineering design) since 1987. His research interests are in the area of computer-supported decision making and engineering design. He was the head of the team from Carnegie Mellon to develop a Technical Data Manager that integrated he views of information at the ALCOA Technical Center across divisions. He was the principal investigator in the Asea-Brown Boveri Engineering Design System project for the design of power transformers (1991-1994). Since, he has worked on information flow analysis of the engineering of many of the products with ABB Corporate Research, Norway. He was the lead investigator of the information systems for Product development project funded by Robert Bosch Gmbh (1996-1998). He is currently the principal investigator of the Engineering Process Information Systems in the design of transit systems for ABB-Daimler Benz Transportation(ADTRANZ).In an ARPA project collaborative systems for assembly based design he is the co-principal investigator with Profs. Talukdar and Siewiorek (1996-1999). He is a co-principal investigator of a project to build a team-based design environment, n-dim, to support design by aiding in the formulation, capturing and sharing diverse views which has been used in the creation of prototypes for different work contexts including our own. He is the author of over 30 technical papers in the areas of design research, use and applications of information technology in design. He spent 6 weeks at the invitation of Danish National Research lab (RISOE) in Denmark to study the state of the art in Computer Supported Collaborative Work in Europe. He served on the editorial board of AIEDAM . He is currently in the editorial board of Research in Engineering Design Journal.
Professor Art Westerberg Arthur Westerberg is currently University Professor and the Swearingen Professor of Chemical Engineering at Carnegie Mellon University, Pittsburgh. He received his B.S. from Minnesota, M.S. from Princeton, and Ph.D. from Imperial College, University of London, all in Chemical Engineering. He has served as Director of the Design Research Center (78-80), Chemical Engineering Department Head (80-83), and Director of the EDRC, an NSF funded Center(86-89). His research interests are in engineering design, specifically in design 39
synthesis, analysis, optimization, environments, and applications of expert systems. He is the recipient of awards from AIChE (CAST Division Award, 83; Walker Award, 87; Institute Lecture, 89; McAfee Award, Pittsburgh Section, 90, AICHE Fellow, 1993, Founders Award, 1995), E.V. Murphee Award from ACS, 97 and from ASEE (Chemical Engineering Division Award, 81). He is a member of the National Academy of Engineering. His membership in societies and organizations include Sigma Xi; Member of National Research Council Committee on Chemical Engineering Frontiers, head of Panel on Computer Assisted Process and Control Engineering, 1984 - 1988; American Institute of Chemical Engineers, past Program Co-Chairman (1975-77) and chairman (1977-79) of area 15a; member, co-chair, chair of National Awards Committee, 1988-94; American Society of Engineering Education, session chairman 1970; Chemical Engineering Program Chairman summer meeting, New Orleans (9 events/sessions) 1979; CACHE, 1971-1980, secretary 1971-73, FOCAPD-II Conference Chairman, Snowmass, CO, June 19-24, 1983; Member National Academy of Engineering since 1987. He is on the journal advisory boards of Computers and Chemical Engineering Journal, 1982-present; Chemical Engineering Reviews, 1982-present; AI EDAM, 1987-present; Research in Engineering Design, 1988-present; Journal of Optimization Theory and Applications (JOTA), 1989-present; Industrial & Engineering Chemistry Research, 1993-1996.
40
doc_779346490.pdf
Organizations have recognized that knowledge constitutes a valuable intangible asset for creating and sustaining competitive advantages
Knowledge Sharing and Management in Large, MultiNational Corporations
Joseph G. Davis Associate Professor of Information Systems University of Wollongong Wollongong, NSW 2522 AUSTRALIA Email: [email protected] Phone: 61 2 4221 3648 Fax: 61 2 4221 4474 Eswaran Subrahmanian Principal Research Scientist Institute for Complex Engineered Systems (ICES) Carnegie Mellon University Pittsburgh, PA 15217 Email: [email protected] Arthur W. Westerberg University Professor and Swearingan Professor of Chemical Engineering Carnegie Mellon University Pittsburgh, PA 15217 Email: [email protected] (Please address all correspondence to Joseph G. Davis) This research was funded by a grant from the Carnegie Bosch Institute, Pittsburgh, PA under the ‘Impact of Global Information Revolution on International Management’ program. Previous versions of this paper have been presented at the bi-annual Carnegie Bosch Institute conference held in San Francisco, October 21-23, 1999, at the Prototyping Research Laboratory, Department of Mechanical Engineering, Stanford University, Palo Alto, CA on October 19, 1999, at the Du Pont Experimental Station, Wilmington, DE on December 16, 1999, and at the Du Pont Photopolymers & Electronic Materials headquarters at Research Triangle Park, NC on December 17, 1999. We have benefited from comments and suggestions from several participants at these presentations and seminars. We are indebted to Joe Miller, Du Pont’s Vice President, Research and Rita Seelig Ayers and Margaret Isselmann of Corporate Information Science for their cooperation and assistance without which the empirical segment of this study could not have been completed. We also thank the 44 respondents and Steven Miller (Leader of the Communication, Collaboration, and Coordination Program Office at Du Pont) for being generous with their knowledge and time. We acknowledge Professors V.S. Arunachalam and David Hounshell of Carnegie Mellon University for their guidance.
1
Knowledge Sharing and Management in Large, MultiNational Corporations
Abstract
Knowledge sharing and management (KSM) has emerged as an important issue for international management. However, there is considerable confusion as to what constitutes organizational knowledge, whether and how it can be systematically managed, and what are some of the effective organizational and technological mechanisms for facilitating knowledge management. This study seeks to unravel the complexities associated with it, initially by developing a typology of such mechanisms. This will provide a starting point for a detailed field study carried out in a large, multinational company (Du Pont) focusing on the critical issues, concrete practices, bottlenecks, and constraints in knowledge sharing and management in two functional areas, two business units, and four countries. Using the field data, we also elucidate a dynamic model of knowledge with emphasis on ongoing interpretation and contextualization of previously generated knowledge. Design guidelines for implementing computer-based systems to support KSM are also presented.
2
Knowledge Sharing and Management in Large, MultiNational Corporations
1.0 Introduction The notion of contemporary organizations as knowledge producing, sharing, and disseminating entities is gaining rapid currency among researchers in Organizational Studies and International Management, among others. While the critical role played by the stock and application of knowledge in economic development at the macro-economic level is relatively well understood (Machlup, 1980; Nelson and Winter 1982; Eliasson et al. 1990), its centrality in the management of individual firms is more of a recent concern. This interest is perhaps a response to the challenges posed by an increasingly complex business environment characterized by intensified competition, greater globalization, and compressed product life cycles and the consequent information overload for senior management. Concurrently, advances in new information and communications technologies (NICT) in the form computersupported cooperative work (CSCW) systems, groupware, knowledge-based systems, internet, intranet, and the world-wide web (WWW) promise capabilities for developing effective solutions to the KSM problem.
The issues referred to above have been faced in a more acute form by large, multinational corporations (MNCs) for which the forces of global integration, local differentiation, and worldwide innovation have become stronger and more compelling. Several scholars have argued that such firms have had to devise means to enhance their global flexibility and learning levels in order to stay competitive (Bartlett and Ghoshal, 1991; Doz and Prahalad, 1991, among others). This is increasingly achieved through the adoption of new organizational capabilities for pooling world-wide knowledge and to transfer and adapt innovative product and process technologies and project management knowhow to international markets. We investigate empirically the organisational and technological mechanisms employed by MNCs to promote knowledge sharing and to develop and manage their intellectual resources. The implicit assumptions that underpin the category of ‘knowledge’ in this context as reflected in
3
the relevant literature and contemporary organizational practices will be explored and critiqued. We also address the barriers to effective knowledge management and discuss approaches to overcoming them through an alternative conceptualization of the dynamics of organizational knowledge and the design of appropriate NICT-based systems to facilitate its sharing and management.
This study was carried out in three phases. In the first, we developed a comprehensive typology of organizational mechanisms for knowledge sharing and management. An exhaustive review of the growing academic and practitioner literature provided the basis for this typology. In the second phase, we completed a detailed field-based, case study of the conceptualization of knowledge and KSM practices in R&D and project management in a large MNC - Du Pont, a diversified, manufacturing company with headquarters in the United States. The company has extensive R&D and project management operations in a number of countries. The data gathering was carried out through in-depth interviews with key R&D and engineering project management (EPM) executives in four different countries in two separate business units. The typology developed in the first phase and a preliminary set of interviews with senior Du Pont managers were used to develop semi-structured questionnaires for the data gathering. In the third and final phase, set of system requirements for the design of an integrated knowledge management system was generated based on our analysis of the field data.
This paper is organised as follows: in the second section we present an overview of the extensive literature on knowledge management and the related topic of organizational learning. This review constitutes the basis for the typology of KSM mechanisms proposed in section 3. Section 4 is devoted to the research methods and techniques employed in gathering the data for the case study. The fifth section includes the details of the case study followed by an outline of some of the generic system requirements for developing systemic solutions to the KSM problem.
4
2.0 Overview of the Relevant Literature 2.1 Organizational Knowledge There is emerging consensus that perhaps the most important source of sustainable competitive advantage in an increasingly turbulent global business environment is knowledge. The organizational capability to create, recognize, disseminate widely, and embody knowledge in new products and technologies is critical when faced with shifting markets, rapid product obsolescence, hyper-competition, and financial upheavals (Nonaka, 1991).
However, despite extensive discussions of concepts such as ‘knowledge management’, ‘intellectual capital’ and related ones like ‘organizational learning’ and ‘organizational memory’, there appears to be a lack of conceptual clarity in the minds of researchers and practising managers on their specific meanings. Davenport et al. (1998) has suggested that it is not useful to worry about the deeper ontological and epistemological dimensions of knowledge; they instead focus on specific knowledge management project as the unit of analysis. This, in our view, is a risky strategy since it is hard to study knowledge management without a coherent conceptualisation of organizational knowledge. We present below, a review of the divergent perspectives on organizational knowledge and the latent assumptions that underpin them.
2.11 Philosophical Issues From an epistemological standpoint, most of what is characterized as organizational knowledge falls within the purview of ‘weak knowledge’ (Goldman, 1999). This is sharply contrasted with ‘strong knowledge’ which is the traditional focus of philosophical inquiry. Strong knowledge imposes very stringent criteria for the use of the term ‘knowledge’ which includes true belief backed up by valid justification or warrant for the belief, and the systematic exclusion of alternative possibilities. Such knowledge is rarely attainable in organizational domains. Besides, this form of skeptical rationalism bears little resemblance to the action-centered processes of generating, sharing, and utilizing knowledge in organizations. It is unlikely that the ‘strong knowledge’ epistemology is particularly helpful1 in view of the inter-subjective and social nature of much of organizational knowledge. The latter tends to
1
The role and influence of strong knowledge in the activities of the R&D department cannot be ruled out.
5
privilege consensus, immediacy, and perceived use value over strong-verifiability, superobjectivity, and transcendence.
Weak knowledge implies true belief without the iron-clad backing of strong warrant or justifiability for the belief. It does, however, imply something more than just information and involves sustainable belief produced by the information that provides some kind of evidence pointing to the validity of the propositions and away from its rivals (Dretske, 1981; Goldman, 1999).
The justification for organizational knowledge is typically reliant on the following bases: i. Coherence: knowledge fragments obtain their justification by virtue or cohering or meshing with the rest of the knowledge base so as to produce a holistic effect (Quine, 1952; Goldman, 1999). ii. Pragmatism: pragmatist epistemology emphasizes the relationship between knowledge and action. This position was initially developed by William James and subsequently expanded on by Dewey who endeavored to bridge the chasm between theory and practice, knowledge and action. The validity and use of knowledge can go only so far as it can “…..pass into actions which rearrange or reconstruct in some way, be it little or large, the world in which we live” (Dewey, 1929). Barring an extreme instrumentalist view, pragmatism attempts to develop an interactive model of knowledge based on engagement with the world through action, experimentation, and experience (Nonaka and Takeuchi, 1995).
2.12 Economics and Organizational Knowledge
Traditional microeconomic theory depicts (technical) knowledge as a quasi-public good. It is characterized by high levels of indivisibility and non-excludability. Its generation is the result of scientific research and general methodological procedures. Its transfer is largely unproblematic and is viewed as a spontaneous aspect of the economic system. The ability to appropriate the knowledge by the innovator is low even though patenting and intellectual property rights can reduce the scope for societal benefits from the knowledge (Arrow, 1969; 1994; Antonelli, 1999).
6
This perspective has been challenged by a number of researchers. The distinction between technological information and technological knowledge is sharply drawn with the latter conceptualised as incorporating a set of capabilities and competences needed to utilize the knowledge which in turn can be leveraged to generate new knowledge. Such knowledge is generated by a process characterized by cumulativeness and path dependence (Jorde and Teece, 1990; David, 1993, among others). Knowledge, according to this view, is highly localized and embedded in the previous background and experience of individual firms. It is the result of a learning process and involves highly specific and relatively ‘tacit’ knowledge processing (Antonelli, 1999).
2.13 Other Representations of Organizational Knowledge
The central role of knowledge in the firm and the organizational processes and mechanisms for its integration and sharing across national borders is the primary basis of Kogut and Zander’s theory of the multinational corporation. They have also highlighted the need for the mechanisms to be sensitive to the degree of tacitness or codifiability of the knowledge (Kogut and Zander, 1992a; 1992b)
A range of definitions and perspectives on knowledge has been presented in the organizational literature. Kerssens Van-Drongelen et al. (1996) defines knowledge primarily in the context of R&D as “… information internalized by means of research, study, or experience, that has value for the organization” (Kressens Van-Drongelen et al., 1996 p.214). Similar conceptualizations of knowledge as the result of processing and refining of information have been implicitly or explicitly employed by a number of authors. This view is an extension of the information-processing paradigm popularized by March and Simon (1959), Simon (1977), and Galbraith (1974). More recently Simon (1996) has argued that the challenge for today’s managers is to filter and extract relevant knowledge from the vast amounts of potential and actual information available from internal and external sources. Effective systems need to be designed to intelligently filter information. In a similar vein, Davenport et al. describe knowledge as information combined with experience, context, interpretation, and reflection and knowledge production as comprising value addition to information (Davenport et al., 1998). Information is the flow of messages or meanings that might add to, restructure, or change knowledge according to Machlup (1980). Coyne (1997) refers to these and related 7
representations of knowledge as the system theoretical view according to which the essential knowledge is contained in the information content and the subjective, inter-subjective, and spatial aspects are largely ignored.
An alternative view has been championed by Dretske (1981), Nonaka (1991; 1994), among others. This perspective acknowledges the importance of subjective factors such as beliefs and their links to actions as well as the relatively tacit dimension of knowledge. Knowledge is associated with beliefs produced and sustained by information (Dretske, 1981). Information represents a flow of messages but knowledge is created and organized from it, anchored by the commitments and beliefs of the concerned individuals. There is also a connection between such knowledge and the subject’s ability to plan and act.
The more implicit and tacit dimension of knowledge has also been highlighted. For Polanyi (1967), explicit or codified knowledge is what is transmittable through formal and systematic languages. Tacit knowledge is more personal and subjective, making it difficult to be formalized and communicated. It tends to be deeply rooted in action, commitment, and involvement in a specific context. According to Nonaka, individuals are able to recreate their own systems of knowledge to accommodate ambiguity, noise, and randomness generated in the organization in its interaction with the external environment (Nonaka, 1994). Such knowledge resides in situated communities of interpreters rather than in texts or messages and these make sense only in particular interpretive contexts (Reddy, 1979). These communities emerge not through absorption of abstract knowledge but when members become insiders and acquire the community’s shared vision and speak the same language (Brown and Duguid (1991). The notion of contextualization of knowledge and evolving communities of practice have particular resonance for MNCs given the geographic distances and cultural differences across units around the world. Following Coyne (1997), we refer to this perspective as the pragmatic view.
Table 1 presents a somewhat stylized set of distinctions between the systems-theoretic and pragmatic perspectives on organizational knowledge. It is worth noting that the dichotomy between system theoretical and pragmatic perspectives and their respective epistemological and ontological assumptions pervade most of the writings on organizational knowledge.
8
System Theoretic View Source Documents, databases, systems, prototypes, processes and procedures, manuals etc. Codified or codifiable; explicit Exchange of documents, electronic means, formal training Relatively mechanistic Cartesian; separation of mind and body.
Pragmatic View People, communities
Form Transfer
Tacit; implicit; intuitive Socialization, apprenticeship, osmotic; situated learning Organic More holistic; unity of mind and body
Organization Philosophical Perspective
Table 1: Perspectives on Organizational Knowledge
2.2 Knowledge Sharing and Management As we would expect, the fundamental differences between the system theoretical and pragmatic views in characterising organizational knowledge are reflected in the divergent approaches and perspectives on knowledge creation, sharing and management in organizations. In general, the former tends to focus on structural and systemic approaches while the latter emphasizes human-centred processes such as socialization, self-organizing teams, extended social interactions, personnel rotation etc. Besides the diverse modes of knowledge creation and transformation, the globalized firms are faced with the challenge of mobilizing and integrating fragmented forms of knowledge spread all over the world (Cohendet et al., 1999). As well, developments in new information and communications technologies (NICT) are increasingly making it easier to separate, transport, and trade knowledge (Antonelli, 1996).
Nonaka has described processes for the creation and transformation of knowledge from tacit to more explicit forms. These include sharing language, experiences, mental maps and models through socialization, reconfiguring existing knowledge through sorting, adding, categorizing, explicit knowledge through combination (typically using computer technology), externalization of tacitly held knowledge through the use of appropriate metaphors and other triggering devices, and internalization of explicit knowledge by recontextualizing along the lines of double loop learning proposed by Argyris and Schon (1978). This involves the
9
creation of new knowledge by reconstructing existing perspectives, frameworks, or premises on an ongoing basis (Nonaka, 1994, p.19). Furthermore, Nonaka presents a model of knowledge expansion and growth arising from the dynamic interplay between tacit and explicit knowledge possessed by organizations through the four processes discussed above.
While it is true that knowledge is continuously generated throughout the organization, some of the areas of critical focus have included R&D (Kerssens-Van Drongelen et al., 1996; Jain and Triandis, 1990), project management, especially in distributed project management teams across multiple, geographically work sites (Van Fenema, 1997; Manheim, 1993), new product develoment (Clark and Fujimoto, 1991; Wheelwright and Clark, 1992; Takeuchi and Nonaka, 1986), joint ventures (Inkpen, 1998; Culpan, 1993), technology transfer (Souder et al., 1990), and customer needs and market developments (Ashton and Stacey, 1995). A few of the arguably effective mechanisms commonly prescribed for the sharing and management of knowledge in these and other studies include: • multifunctional project teams (Wheelwirght and Clark, 1992) • personnel rotation (Edstrom and Galbraith, 1992; Inkpen, 1998) • post-implementaton reviews and audits (Kumar, 1990) • best practices documentation ( Davenport et al., 1997) • structured task communication (Brown and Eisenhardt, 1995) • learning histories (Kleiner and Roth, 1997) • a variety of IT-based systems typically involving various combinations of inter/intranet, document management and communication, knowledge and database technologies. Most of these provide a subset of capabilities for searching, sharing and collaborating over documents in electronic repositories. One of the more celebrated among these are the intranet-based cooperative work and knowledge sharing system called virtual team network (VTN) developed by British Petroleum and used extensively. (Prokesch, 1997)
The ‘knowledge base’ of an organization may consist of several sources or media. KerssensVan Drongelen et al. (1996) has classified these sources into four categories based on the work of Zeleny et al. (1990), • brainware: knowledge in the minds of people, such as intuition, experience, specialized knowhow,
10
• hardware : tangible things such embodying knowledge such as material prototypes, products, R&D equipment, production processes, • groupware: unwritten knowledge shared by people such as rules of thumb, heuristics, procedures, stories, myths, • documentware : knowledge documented on paper or in databases/information systems such as CAD/CAM models, intranets, parts databases, patents, white paper, manuals, lab reports, hand books, project post-implementation reviews etc.
3.0 Toward a typology of KSM mechanisms Based on the preceding discussion, we propose a typology of knowledge sharing mechanisms employed by organizations to facilitate knowledge sharing and to effectively manage the knowledge resource. It consists of the following three categories: Type I: Type II: Type III: Static sharing of documents and knowledge bases (knowledge repositories) Structural and human resource-based mechanisms Integrated mechanisms for facilitating ongoing interpretation and contextualization of knowledge.
It should be noted that the categories are not strictly mutually exclusive in that there could be some overlap among them. Also, most organizations attempt to use all three of these and there are substantial synergies and potential for expansionary spirals in knowledge generation and utilization from the interplay across the categories. In this sense, the categories are best viewed as Weberian ‘ideal types’ to the extent that a specific KSM practice can be roughly mapped into any of one of the categories. The question as to whether organizations in general and MNCs in particular exhibit a dominant mode among the three types is an empirically one.
Type I: Static Sharing of Documents and Knowledge Bases (Shared Knowledge Repositories) This mode of knowledge sharing is based on the assumption that all knowledge is or can be made explicit in the form of documents or structured knowledge and databases. The problem of knowledge sharing and management is this reduced to representation, codification, and persistent storage of the documentary knowledge source in physical or electronic repositories, which can be searched, browsed, queried, filtered, and digested by any participant. The intranet, world wide web, and group-oriented document database and communication systems 11
such as Lotus Notes© have given considerable boost to this mode of knowledge sharing in view of the functionalities for globally distributed document sharing and the appealing hypertextual organization of the documents. This approach draws heavily on the system theoretical view of knowledge produced by symbol manipulation and processing.
Type II: Structural and Human Resource Mechanisms
A wide array of knowledge sharing devices fall within this category including crossfunctional teams, transfer and rotation of key personnel, apprenticeship, use of internal consultants, matrix organizational structure, structured face-to-face and electronic meetings, and structured task communication. Many of the processes described by Nonaka (1994) such as socialization, internalization, and externalization as well as brainstorming, and organizational development are also part of this type. Other related means for encouraging collaborative knowledge transfer include the learning organization and practice centers. The chief elements that is common to all these is the goal of contributing to a shared vision and language and the transfer of personal knowledge stored in the brainware to be available to a wider cross section of people. The thrust, however, is on the social processes and the continuous regeneration and expansion of tacit knowledge as a result of ongoing exchange among people and is informed by the pragmatic model of knowledge discussed in section 2.
Type III: Integrated Mechanisms for Facilitating Continuous Interpretation and Contextualization of Knowledge.
This mode of knowledge sharing attempts to synthesize the seemingly exclusive perspectives that underlie the first two categories. It is predicated on the assumption that the boundaries between tacit and explicit knowledge are fluid allowing for the possibility of a dynamic interplay between the two forms of knowledge. As well, these mechanisms underscore the need to reinterpret all knowledge on an ongoing fashion since the application of most knowledge has to be performed in newer contexts which have to be modelled and integrated to facilitate effective knowledge sharing and reuse. This model comes closest to Nonaka’s notions of shared experiences, crystallization (process through which various groups in an organization test the reality of a concept created by a self-organizing team), and context12
switching hypertext organization. This can potentially result in a dynamic knowledge cycle to continuously create, exploit, and accumulate organizational knowledge. The ndim system developed and tested by the ndim group at the Institute for Complex Engineered Systems at Carnegie Mellon University is a computer system designed with a similar vision of creating living repositories for sharing knowledge in the domains of engineering design and R&D (ndim group, 1995). Kleiner and Roth (1997) describe ‘learning histories’ as a new managerial tool that captures the lessons of successes and failures in an open and interpretive framework that also fits into this category of knowledge sharing and management. 4.0 Research Methods and Techniques The primary goal of our research is to empirically investigate processes and practices for knowledge sharing and management in large, diversified organizations in order to lead to the development of enhanced NICT-based solutions to support it. Our empirical work was guided by existing theoretical perspectives and previous field studies. It is our contention that systematic and focused field studies seeking to inquire into specific aspects of KSM can serve to advance our shared understanding of this important topic beyond anecdotal evidence and isolated case studies.
The fundamental research questions we addressed were the following: 1. What are the different mechanisms by which large multi-national corporations (MNCs) pool, exchange, and share knowledge? 2. What procedures are employed to contextualize, validate, verify, and authenticate the knowledge generated? 3. What is the role of information technology in supporting KSM? The field study was carried out at Du Pont, a large, diversified, global manufacturing company with its headquarters in the United States. Du Pont has world-wide operations and a strong commitment to understanding and sharing knowledge across the various business functions and units. The industries and markets that Du Pont serves range from traditional and specialized nylons, polymers, polyesters, and fibers to agriculture, nutrition, and pharmaceuticals organized under eight business segments and twenty business units. Through most of its history, Du Pont has seen itself as a science-based company with the mission of bringing the benefits of science to the marketplace in ways that benefit the shareholders and
13
society. Scientific and technological knowledge in chemical and material sciences and biological sciences is the basis for the company’s business portfolio.
The primary factor in the choice of Du Pont for the field study was the global and knowledge intensive nature of its operations. Besides the scientific orientation of the company, it has an explicitly stated strategy of seeking growth through what its current Chairman and CEO Chad Holliday calls “….knowledge-intensive projects that transform industries”. He defines “knowledge-intensive” as involving a complete understanding of “….all aspects of the consumer or business need and the dynamics of the existing value chain that meets this need. This will create higher value solutions for our customers in more productive and less capitalintensive ways” (Holliday, 1998). An equally important factor was the long Du Pont tradition of research cooperation with universities and the receptiveness to our proposal shown by senior Du Pont managers.
In order to keep the investigation focused and within manageable proportions, we restricted our data collection activities to knowledge in the areas of two functional areas (R&D and engineering project management - EPM), two business units (Engineering Polymers and Microcircuit Materials, a part of the iTECH business unit previously named Photopolymers and Electronic Materials), in four countries (USA, UK, Switzerland, and Japan). R&D and EPM are both highly knowledge-intensive domains that are at the core of Du Pont’s global operations. While R&D activities at Du Pont are performed at the corporate level as well as distributed at multiple sites in each of the business units, EPM is largely centralized at the corporate level. Both Microcircuit Materials (MCM) and Engineering Polymers (EP) have manufacturing, marketing, and R&D operations spread out across the Americas, Europe, and Asia.
The data gathering and analysis methods most appropriate to the questions we have raised are primarily qualitative and interpretive. Direct observation is clearly the best approach to investigating these problems where the process dynamics needs to be captured and the potential for multiple, conflicting interpretations can be expected. However, direct observation of the focal phenomenon can be extremely difficult and can impose an inordinate burden on the participating organizations. Accordingly, we settled for in-depth interviews with a cross section of professionals. By asking a common and structured set of questions 14
based on semi-structured questionnaires, we attempted to unravel the complex and sometimes implicit processes of knowledge sharing, their relative efficacy in different contexts, and the bottlenecks to effective sharing. It also enabled us to understand and interpret knowledge sharing in context, which is very important from our perspective. We also collected copies of documents that provided us with a more comprehensive portrait of KSM activities. This makes up, to some extent, for the limitations arising from not being able to investigate knowledge management issues longitudinally. The nature of this kind of interpretive data gathering required that we analyze, interpret, and reanalyze the data as it was collected.
All except eight of the 44 respondents who participated in the study were interviewed at their normal places of work. This was important because it gave them an opportunity to access and refer to documents, to demonstrate the system(s) they use, and to point to additional sources. Each interview lasted between 75 and 90 minutes and the responses were recorded by two of the researchers. One of them used the traditional linear narrative and the other employed the mind map technique by which related ideas and concepts were explicitly linked and organized together. The interview notes were compared, reconciled, and summarized at the end of each day of data gathering. A profile of the participants in the study is presented in Table 2. As can be seen the participants represent a cross section of professionals from corporate and business unit R&D and technical support groups in four countries and from Central Engineering. 57% of all respondents and 64% of all R&D respondents had Ph.D level qualifications. Their average number of years of experience at Du Pont was 17.6 with a range of 1 to 34 years.
Total Micro Circuit Materials Central R&D Total no. interviewed No. and % of PhD’s Average no. Years at Du Pont Range of Years at Du Pont 5 5 (100% 19.2 Central Engg 5 0 (0%) 23.6 Exp. Station 3 3 (100% ) 29.0 RTP Bristol (UK) 7 4 (57%) 11.3 Utsonomiya (Jap.) 5 0 (0%) 11.6 Engineering Polymers Meyrin (Swiss) 8 6 (75%) 13 Parkers burg (WVA) 1 1 (100%) 32 Exp. Station 1 1 (100% ) 21 44 25 (57% ) 17.6
9 5 (56% ) 21.0
11-2
1434
11-33
9-34
1-18
9-14
2-26
32
21
1-34
15
Table 2: Participant Profile
The nature of issues explored in the interviews dealt with: 1. The conceptions of knowledge entertained by the respondents. 2. The nature of the formal and informal mechanisms in place for the sharing and dissemination of knowledge. 3. The technology, architecture, design and implementation details of the IT-based systems in place for knowledge sharing and the respondents’ perceptions on the degree to which they are successful and their limitations. 4. Bottlenecks to effective sharing of knowledge across time and space. 5. Requirements for an (ideal) system that might address aspects of (4) above.
The final phase of the study involved revisiting the case data and gathering additional information from key informants to establish a set of requirements for the design of a knowledge management system. This work was largely based on field data obtained as part of the case study and our observations and understanding of the critical constraints in knowledge sharing. The requirements definition was carried out from a more generic perspective so as to enable us to develop a more general model of such a system with the goal of facilitating a dynamic and context-based support for knowledge sharing and management in contemporary organizations.
4.0 Case Study: Findings and Discussion 4.1 Background on the Business Units and the EPM Function Studied
As stated previously, the choice of Du Pont as the field study site was largely guided by the knowledge intensity and global scale of its operations, and its size and diversity. Hounshell and Smith (1998) have observed in their monumental study of the relationship between science and corporate strategy in Du Pont that the company was one of the handful of older enterprises with a long tradition of successful marriage between scientific knowledge and business, primarily through the institutionalization of industrial research and development. The distinguishing feature of these companies was that they were based on technologies that had emerged from scientific discoveries or developments further refined through a scientific approach to problem solving. Especially in the latter half of the 20th century, Du Pont had 16
become well-diversified in respect of both the range of product-markets it served and geography. By the 1960’s, it operated in 65 countries and nearly 50% of its revenues were generated outside the United States. Though a study of this type deals with certain unique aspects and features of Du Pont, we would expect to find resonance for its findings in similar diversified, knowledge-intensive firms. The issues, patterns, and processes employed in mobilizing, sharing, integrating, and managing knowledge and some of the possible relationships observed are unlikely to be widely disparate.
Our focus of the study was not on the knowledge content of R&D and EPM functions per se but on the mechanisms and processes for generating, sharing, and utilizing it. Given the distributed nature of R&D activities at Du Pont, personnel from the central R&D facility (Experimental Station, Wilmington, Delaware) were also included as participants in the study. Engineering is a centralized function at Du Pont and all the participating engineering personnel were from Central Engineering, one of Du Pont’s large, global service departments. We provide below a brief background on the activities of MCM and EP business units and the EPM function.
4.11 Microcircuit Materials (MCM)
MCM is one part of the iTECH business unit belonging to the Specialty Polymers business segment. Its products include thick film pastes, inks, and other high performance films for electronic circuit boards, display panels etc. It is considered a ‘high technology’ business and the customer base is technologically demanding and volatile. The industry groups served by MCM include some of the more technologically advanced sectors such as computer hardware, mobile telecommunications, avionics, military, and automotive. The rate of new product development is high with nearly 50% of the products having been introduced in the last 4-5 years. Much of the new product development work is driven by the customers whose needs and levels of sophistication tend to evolve rapidly. R&D personnel find themselves having to integrate knowledge concerning new materials and processes with customer- and applicationspecific knowledge to generate new products. Part of this knowledge has immediate or future relevance for the other regions served by MCM.
17
The organization of R&D in MCM consists of multiple layers. The more fundamental and longer range research activities such as the development of a new series of resistors or new metallic powders and other intermediates are performed at the central R&D facility (Experimental Station). Detailed development and scale-up for manufacturing is carried out at the business unit headquarters in Research Triangle Park (RTP), NC. The research facilities at Bristol, UK and Utsonomiya, Japan along with RTP conduct detailed product innovations and modifications typically by working closely with major customers. The primary manufacturing facility is in Puerto Rico with additional manufacturing for regional needs in Europe and the Asia-Pacific being catered to by facilities in Bristol and near Taipei, Taiwan respectively.
4.12 Engineering Polymers (EP)
EP business unit is part of the Performance Coating and Polymers business segment. Its product portfolio consists of a broad range of engineering materials for automotive, electrical/electronic, and industrial applications. Much of the R&D focus is on substitution of metal and rubber and finding new product applications using families of engineering resins such as Zytel© and Minlon© nylons, Delrin© acetal, Hytrel© thermoplastic elastomer etc. The headquarters of EP is at Parkersburg, WVA where much of the business and R&D coordination takes place. Basic research on new and existing resins is largely performed at the Experimental Station and Parkersburg. Development centers and marketing offices in different parts of the world work closely with other equipment manufacturers (OEMs) to improve existing and develop new applications of Du Pont products.
4.13 Engineering Project Management
Engineering is a centralized global service. In addition to building new plants and other facilities and renovating and retrofitting existing ones, Central Engineering works closely with R&D to design and install new and more efficient manufacturing processes. It also has staff groups responsible for all aspects of environmental and safety engineering. It develops and maintains procedures and standards for chemical process design, project implementation, and project safety management.
18
Du Pont has enjoyed a reputation for good project management capabilities in carrying out all aspects of project implementation from preliminary design and costing through detailed design and installation. This is supported by the favourable references in project performance reports produced by independent consultants benchmarked against industry standards. In recent years, however, Central Engineering has been subject to significant downsizing. A wide range of project life cycle activities including detailed design and implementation of projects are being outsourced to full service design and construction firms. Central Engineering’s role has, as a result, been confined to front-end loading, monitoring of project progress against standards, and staff functions. Du Pont has also undertaken several joint venture projects all over the world with other large companies. The number of employees in Central Engineering has declined from nearly 5000 to just below 1000 during the 1990’s.
4.2 Knowledge Sharing and Management (KSM): Corporate Level Initiatives
Given the knowledge-intensive nature of Du Pont’s operations and the critical role of proprietary technical information, development of appropriate procedures and systems for document management and codification have received significant attention during most of the last one hundred years of the company’s history. Many of Du Pont’s business units have their own information centers and libraries. Corporate Information Science is responsible for the most prominent of these, the Lavoisier Library at the Experimental Station and the information centers that serve the Corporate R&D (CRD). Information centers provide access to copies of all relevant documents, memoranda, reports, and items of correspondence could be accessed. The advent of computer systems and electronic storage and communication has resulted in the provision of online access to the information in the library and the information centers through the use of Du Pont’s online library catalog and the databases for retrieving Du Pont’s research and engineering documentation. As a consequence, some of the distributed information centers have been merged or eliminated.
Initiatives aimed at the design, building, and implementation of integrated database systems for the storage, retrieval, and distribution of Du Pont scientific and technical information go back to the 1960’s. These resulted in the development of some of the fundamental principles for efficient representation and easy search of complex molecular structures, a system for polymer nomenclature, and a hierarchical thesaurus for indexing technical reports, engineering 19
drawings, and patent documents. The development of the DuPont SCION (Scientific Information Online) proprietary database in the 1990’s was greatly facilitated by these initiatives (Schultz and Wilks, 1997; Kvalnes, 1997). Technology Conferences sponsored by the Vice President of Research provide useful forums for active networking and sharing of tacit knowledge.
Central Engineering has a long tradition of establishing standards and codification of useful knowledge and practices. A vast array of manuals and guides on topics ranging from project implementation, safety practices, operating policies, training records, and plant maintenance etc. have been developed over the years. Some of these have become partly obsolete as a result of outsourcing and joint venture arrangements that present new challenges for codification and updating of procedural knowledge. Du Pont’s Chemical Process Design Standards cover a wide range spectrum including fluid flow, heat transfer, dust and mist collection, agitation, filtration, distillation, reactor design, construction materials, physical properties etc. They have been developed over 50 years of research and practice and run into four volumes of more than 1000 pages. Sharing hard copies of these documents along with their associated mathematical models and materials properties databases to assist Du Pont engineers, contractors, vendors, and joint venture partners has always been problematic. Considerable progress has been made in Du Pont’s effort to migrate much of this to the World Wide Web (corporate intranet). This is beginning to ease the difficulties in sharing the resources and permits the relatively straightforward updating of the information (Scrivner, 1997).
At the corporate level, the KSM-related responsibilities fall under the purview of Corporate Information Science (CIS) augmented by the activities of the Communication, Collaboration, and Coordination Program Office (C3PO) and the distributed libraries and information centers. The former was founded in 1987 and is responsible for managing Du Pont’s proprietary information, competitive intelligence function, and the Lavoisier Library. It also provides a range of information services to the businesses including ongoing codification of enduring knowledge in the businesses. The distributed libraries and information centers are managed by the individual business units. C3Po spearheads the Communication, Collaboration, Coordination project which includes the company-wide rollout of Lotus Notes© and related software with the objective of providing the architecture for collaboration in Du Pont. This is 20
a 4-5 year project that started in 1998, at the end of which the group hopes to standardize much of the communication and collaboration work around Lotus Notes tools for email, document databases, workflow systems, screen sharing, and calendaring. All knowledge workers are being encouraged to run Lotus Notes clients on their workstations to connect to Du Pont’s network of 90 Lotus Notes servers and 450 other intranet servers. Currently there are nearly 60,000 Lotus Notes clients globally but this number is expected to grow as support for other legacy tools and systems are withdrawn. It is planned that the server network will be gradually extended to an extranet to include relevant external entities such as vendors, contractors, and joint venture partners once the security concerns have been resolved. The C3PO project also facilitates the nurturing of virtual communities of practice with shared interests and expertise, workflow systems, electronic laboratory notebooks, codification of tacit knowledge and prowess, and information technology (IT)-assisted learning.
4.3 Specific Observations and Findings 4.31 Type I Mechanisms and Explicit Knowledge
Type I Mechanisms are the central focus of the ‘knowledge management’ movement and the projects associated with it. The underlying strategy is one of explicit documentation, filtration, and the attempt to systematize, codify, and reuse knowledge to the extent possible. Computer technology in its various forms is used to facilitate selective sharing of the documents and repositories.
As indicated previously, Du Pont has a long tradition of relatively successful deployment of such mechanisms. The migration of many of the document and knowledge bases to the web is progressing steadily. The C3PO and CIS initiatives discussed earlier is aimed at ensuring greater standardization and company-wide dispersion of such mechanisms. We present below some of the significant issues related to their use and perceived effectiveness.
Information overload is a constant refrain, especially among the R&D personnel we interviewed. It appears that the amount of information that is ‘pushed’ at people through electronic mail, document attachments and databases in addition to the physical means is much higher than what can be meaningfully processed by most in the time available. The filtration and search capabilities of Lotus Notes document databases and systems are far too primitive at 21
this stage in relation to the need. One of the respondents compared these to dumpsters in which one had to forage hard to find something useful, if any. The time spent scanning large volumes of marginally relevant information also came in the way of the critical reflection needed to summarize and present in succinct form, their own work for the benefit of others.
Diversity of media and the lack of integration present challenges for effective knowledge processing. Much of the information needed by the respondents is already in electronic form but it is fragmented across a variety of incompatible computer systems and databases. In addition, each knowledge professional typically inherits one or two filing cabinets containing hard copies of correspondence, memos, reports etc. when they begin a new assignment. The differences in the indexing schemes and search mechanisms for the different document bases impede routine activities such as finding useful fragments of information and synthesizing them in the context of particular tasks.
There are critical issues with respect to archival practices that bear on knowledge sharing and management. The traditional, paper-based systems appear to be withering while a robust and company-wide electronic regime is yet to emerge. The ongoing changes in computer hardware, operating systems platforms, data formats, and software versions tend to limit the archival value of electronic documents. The old adage of “if only a company knew everything it knows” could be paraphrased as “ if only a company ‘knew’ everything that is buried in inaccessible computer systems”.
Repositories are available at different levels in the Du Pont units we studied. There are individual repositories and databases that hold much of the local knowledge that is generated. There are also repositories at the level of small groups of people who work together on the same project or with one or a group of customers. The groups need not always be co-located or working synchronously. Finally there are business unit or corporate-level information bases that are maintained. It is inevitable that there will be some duplication and inconsistency across the three levels of repositories. More importantly, the limited general access and the terminological and linguistic differences across these levels hinder the potential for knowledge sharing. As we shall see later, critical human interlocutors emerge at the interfaces between the levels to ameliorate the problem.
22
Several commentators have emphasized the importance of naturally occurring local learning and innovation that takes place at all levels of an organization (for example, Brown, 1991). A related issue is one of facilitating the wider sharing of relevant aspects of such local knowledge through appropriate organizational processes and systems. We found in the course our study that in many situations, such knowledge is embedded in organically evolving and bootlegged systems quite independent of the formal mechanisms available for systematizing and sharing. The reasons for the emergence of such local mechanisms are varied and include: • • The need for specialized structures and capabilities to represent the knowledge. Lack of timely support and assistance for the use of standardized tools and mechanisms. • Idiosyncratic attachment to particular computer tools and legacy systems by some of the employees. In any case, the need to track, evaluate, support, and selectively integrate such local and bottom-up systems cannot be overstated. These are useful initiatives arising from the need to formalize and disseminate localized knowledge with potentially wider implications that arise as a by-product of routine operations. We were able to document a number of these at both MCM and EP though the preponderance of such initiatives at MCM is worth noting.
The Central Engineering group at Du Pont has traditionally been the pioneers in the use of electronic collaboration technology. This is not particularly surprising given the global distribution of project work and the need for rapid, multi-directional knowledge flow among Du Pont engineers, vendors, external contractors, and joint venture partners. As one of the respondents put it, it is a trade-off between getting to work remotely with the collaboration tools or being away from family for long durations. Also, a great deal of effort has been expended in the past aimed at codifying and standardizing a large part of the process knowledge generated over the past 50 years of project work. We observed a preponderance of type I mechanisms being successfully deployed and used in EPM. More importantly, systematized knowledge in the areas of systems for plant safety (design and monitoring) and environmental pollution prevention have become saleable commodities in the form of consulting packages. There is growing demand for such packaged know-how especially in the Asia-Pacific region.
There are differences between the Du Pont units in the modes of structuring and organizing knowledge. For instance, MCM operates in a highly dynamic market place in which rapid 23
learning from the customers and adapting to their changing needs are normative. This has led to the organization of MCM repositories to be primarily based on products, customer segments, and application areas and only secondarily on materials. This is despite the fact that the training and specialization of the scientists and technical support personnel is by materials. In contrast, the business strategy of EP is driven by material push and the knowledge organization at EP is primarily along the material dimension (such as Hytrel, Minlon, Delrin etc.) and within each by major customer groups and OEMs. The need for multiple and divergent classification schemes and indexing systems is a routine aspect of knowledge processing. This has important implications for the design, implementation, use, and maintenance of company-wide KSM solutions with their implicit assumptions regarding the relative homogeneity of knowledge representation schemes.
The resources needed to carry out the difficult and unrewarding task of combining tacit and often fragmented forms of explicit knowledge into systematized and reusable knowledge and to make it publicly available is considerable. While Central Engineering has been successful at this for some time, the relationship between the availability of organizational slack and the group’s ability to standardize and deliver such knowledge is worth exploring. One of the issues arising from the massive outsourcing and downsizing at Central Engineering relates to the reduction in slack and its effect on codification work. There are also concerns about continual erosion of the knowledge base with the department having to rehire ex-Du Ponters and to hire employees from contracting firms to stem the erosion.
4.32 Type II Mechanisms and Tacit Knowledge
It is certainly the case that knowledge workers consult a variety of sources of explicit knowledge and fall back on their learning from formal study and training. However, in the course of actually doing their jobs, most of the learning arises from engaging with real problems, gathering a range of relevant information from diverse sources, and discussing key issues with colleagues and other professionals. Most of the R&D respondents in particular have a strong network of people spread out all over the world that they can turn to for consultation and guidance. A surprisingly large number of such contacts are from outside their own units in other business units to departments in Du Pont or from external entities such as vendors, customers, contractors, universities, and joint venture partners. Some of the 24
respondents can count upwards of 300 people in their network and the intensity of interaction varies over time depending on the tasks on hand. These are typically bottom-up formations and successful R&D scientists spend considerable time and energy developing and nurturing such networks. The diffusion of tacit knowledge in particular through these social networks is quick. The track record of the informants in terms of both the reliability of the acquired knowledge and trust largely determines the length of such associations. Informal protocols of reciprocity also feature in the assessment of track record.
Du Pont has long recognized and promoted mechanisms for such networking. In-company technology conferences and periodic seminars involving people from diverse businesses and technology areas provide facilitative forums for making initial contacts and extending one’s networks. Other processes include apprenticeship training for new scientists under wellestablished R&D personnel as well as the rotation of the latter through newer research sites. The Utsonomiya (Japan) research facility in MCM was established and developed primarily through people rotation and active apprentice-type training programs. In addition, short visits by scientists either to the Experimental Station or to other research sites within the business units are seen as useful mechanisms for the sharing of less tangible aspects of knowledge. The reliance on these mechanisms based on “high bandwidth” channels was more pronounced in R&D as compared to EPM and within R&D, and greater in MCM in comparison with EP. This is consistent with the general observation that higher the role of tacit knowledge, greater the reliance of type II mechanisms.
There is some concern in the R&D groups that the increasing dependence on electronic communication and coordination and the reduced opportunities for face-to-face contact are causing a slow erosion in this mode of knowledge sharing. One of the requirements for any social network to remain active and to be effective in knowledge sharing is the periodic opportunity to ‘catch-up’ through face-to-face meetings, albeit infrequently. Such meetings enable the establishment and/or reinforcement of a common ground needed to cultivate and develop the trust and reciprocity implicit in such relationships. When the operations are globally dispersed as they are in MCM or EP, some of the faster and less expensive mechanisms such as telephone, email, net meeting and even videoconferencing are constrained by linguistic, cultural, and time-zone related barriers. Sustaining such networks involves expensive travel and movement. The curtailment of resources for this purpose in recent years 25
and the increasingly rigorous quarterly financial reporting systems are putting pressure the quality and vitality of such networks. This is particularly hard on the younger scientists who are in the process of developing their personal networks and have to rely on remote, electronic mechanisms to a very large extent.
Human agents were sought for a variety of reasons. Probably the strongest one that came up repeatedly in our interviews was the perception that it resulted in significant efficiencies and time savings. An illustration from an Engineering Polymers R&D center at Meyrin, Switzerland is illustrative. One of the laboratories there provides analytical support and carries out a wide range of routine and specialized tests on properties of new materials, identification of additives etc. When a scientist approaches the supervisor who has managed this laboratory for the past 15 years, he can give good and immediate answers to questions such as whether the same test has been carried out previously and if not whether the results of a closely matched previous test can be extrapolated for a particular problem that a scientist is trying to solve. The information needed to extract the knowledge is available in the form of a database of previous tests carried out and the results but this may not have the format and flexibility requirements that the scientist needs. Searching through this database is also timeconsuming. Besides, the effort and difficulty involved in externalizing the tacit dimension of subtle matches with previous tests is non-trivial. In performing even routine research tasks such as interpreting the results of certain complex analytical tests, sitting down with experts who can help with the interpretation saves much time and reduces some of the guesswork.
Arising from a combination of historical factors, Du Pont R&D is faced with a bimodal distribution age distribution of its personnel. The age cohort in the 50-plus range is very large, in part due to the expansionary spiral of the 1950s through early 70’s. The 40-50 age cohort is small but there is a small increase in numbers in the 25-40 age group. This introduces vulnerabilities in the form of potential erosion of tacit knowledge base in the absence of tacitto-tacit transfer and tacit-to-explicit conversion of knowledge. Such programs are resourceintensive and can only succeed in a climate of trust and security. Also, they need to be guided by sound judgment on the future value of different bodies of tacit knowledge in relation to the company’s strategic direction and the emerging trends in the relevant product-market environments.
26
The phased implementation of the integrated electronic coordination architecture and the C3PO project has generated incentives and pressures to make a transition to new ways of collaborating and knowledge sharing. For some aspects of knowledge work this is only a matter of adapting to a new set of computer tools and they complement other modes. However, the extent to which this regime can or need to supplant the existing, organic and informal mechanisms of tacit knowledge sharing is unclear at this stage. Many of the respondents do make a clear distinction between electronic collaboration and the sharing of knowledge.
4.33 Type III Mechanisms and Knowledge
Type III mechanisms refer to organizational processes and system capabilities that facilitate an active and dynamic interplay between tacit and explicit forms of knowledge. They attempt to integrate the diverse forms of available knowledge and highlight the need for ongoing and context-based interpretation of knowledge. Further, these mechanisms explicitly accommodate emergent and serendipitous knowledge that is at the root of innovative product and process development work in organizations. Creating the conditions and designing processes and systems to support such synthetic knowledge creation and sharing activities pose a major challenge.
There is a clear understanding in Du Pont of the limits of the electronic communication, collaboration, and coordination systems from a knowledge processing perspective. These systems have laid the foundation for communication, information sharing, and some amount of collaborative work across distances. It has also provided the means for more efficient and wider yet selective dissemination of documents and codified knowledge. However, the recognition that people and their networks have to be effectively and seamlessly interwoven with the computer networks and systems has triggered the search for new approaches.
Even the ardent proponents of electronic communication and coordination concede that knowledge management can only be roughly 30% based on the systems being implemented and the rest has to be based on people. Accordingly, a campaign to facilitate the evolution of communities of practice through the use of the communication and collaboration systems has been launched. These communities are largely organic formations comprising members with 27
specialized skills or know-how. They may rely on technology to identify potential members and establish communities to by sharing useful information, offering guidance and critique, recounting related experiences etc. There are currently over a dozen such communities (such as winding experts, people with extrusion skills) in existence at Du Pont and they function like virtual guilds. A project to inventory and assess the effectiveness of such on-line communities is currently in progress.
There seems to be roughly three stages these communities go through as they evolve to the point when they can function without the need for coordinators or champions. The first phase consists essentially of the community members coming together and establishing a common ground. Potential members get to know each other personally and professionally, and engage in limited information sharing. The tools used are mainly e-mail, telephone, net meeting, and screen or document sharing but face-to-face meeting can serve to speed up this process. Some of the communities do not progress beyond this stage in the absence of common problems or projects or in some cases, inspired leadership. Under favorable conditions, a community might move to second phase of being able to engage in effective collaborative work. In this stage, the members have worked out their semantic and terminological differences and have some knowledge regarding the capabilities and competences of each other. The group identity begins to get forged by working collaboratively on a common project or problem. The passage through these two phases is a precondition for the knowledge sharing phase in which adequate levels of trust and confidence in each other have been established and free and open sharing of all forms of knowledge begins to take place.
Innovation-oriented knowledge processing involving new product and process development necessitates the integration of vast amounts of explicit knowledge with certain unique and deeply personal insights acquired through direct practical experiences. The validity and veracity of aspects of the new knowledge being synthesized need to be tested and established through ongoing experimentation and consultations with domain experts. Much of the explicit knowledge available needs to interpreted in the new context. We have documented a number of cases in which both routine and radical innovations were produced by combining different types of knowledge and making leaps that cut across traditional boundaries drawn by existing technologies, business units, and even disciplines. Once the feasibility of the initial idea is established, detailed experiments and trials, and tests can follow. 28
The Bio3GT program that led to the development of a biological process for the manufacture of 3GT fibre is one such example. It involved considerable cross-fertilization of knowledge between the polymeric materials group and the Life Sciences business segment. This has the potential to radically lower the unit cost of production of the particular commodity polymer. The project idea came together at a time when the cost reduction using the more traditional methods like discovering new catalysts could not possibly attain the level needed to compete globally with low cost producers in the newly industrialized economies. Some of the tacit knowledge that is brought to bear in such knowledge processing is acquired through years of variegated experiences in functions as diverse as plant operations, product management, and venture business management. These experiences provide an appreciation for making technological advances and choices that adapt to and anticipate market conditions. Another example of recontextualization of previously acquired knowledge in MCM involves the development of a new process for flat panel displays on LCD screens that leverages available knowledge from the textile printing business unit.
One of the main aspects of the integrated knowledge processing described in the foregoing is the pivotal contributions by key knowledge practitioners who typically combine extensive amount of tacit knowledge with other knowledge. We were able to identify four such knowledge practitioner roles that typically operate at the interfaces between disciplines, technologies, business units, functions, and businesses and customers. Each is described below. 1. High Level Synthesizer: These are typically high level technology or R&D managers with a wide range of experiences across several business units and in several operational and functional areas. They are adept at environmental scanning and closely monitor the trajectories of various technologies. They operate at the interfaces between different technology areas and business units. They develop and maintain a vast network of formal and informal knowledge sources. They look for opportunities arising from various combinations of different types of knowledge.
2. Librarian: Librarians operate at the interfaces between materials, processes and product markets. By virtue of their ability to painstakingly gather, assimilate, index, and store copious amounts of information pertaining to material properties, costs, customer 29
requirements, and changes to the requirements over time, they are able to provide timely assistance and service to other knowledge workers. They deal primarily with explicit knowledge. It takes nearly 10-15 years’ experience to grow into this knowledge practitioner role.
3. Knowledge Engineer: Knowledge engineers operate at the classic interface between R&D, marketing, and customers. They work closely with the customers to adapt existing products and to develop new products and applications. Some of the work might appear rather routine but the tacit dimension of knowledge is in their assessment of customers’ current and future needs. This knowledge has to be integrated with detailed and explicit knowledge concerning the company’s products, processes, and technologies.
4. Knowledge Operators: Knowledge operators are typically technical or customer support personnel who are usually front-line employees located close to the operational realities of the business and the market place. They accumulate and transmit tacit knowledge in the form of embodied skills (Nonaka and Takeuchi, 1995). They tend to work at the interfaces between R&D and manufacturing or R&D and customer operations. They work very closely with knowledge engineers.
Table 3 provides an outline of the different knowledge practitioner roles, the interfaces they operate at, and the types of knowledge dealt with.
30
Roles
Interfaces
Knowledge Operation
High Level Synthesizer Librarian
Cross Discipline/ SBUs/functions Materials/ Product Customer Marketing/R & D
Explicit and Tacit
Mostly Explicit
Knowledge Engineers
Explicit and Tacit
Knowledge Customer-Product/ Mostly Tacit Operators Operations Table 3: Knowledge Practitioner Roles, Relevant Interfaces and Operations
5.0 Toward System Requirements for Enhanced Support for KSM
Based on the foregoing discussion and the interviewee responses, we sketch below a set of systems and corporate policy requirements that can contribute to the goal of more integrated knowledge processing in organizations. The term ‘system’ in this context has to be interpreted broadly to accommodate the symbiotic relationship between physical repositories, computer tools and databases at one end and the cognitive and social processes involving human actors at the other.
a. Organization of Documents Documents of all types and stored on different media are used in knowledge work. However, there is a critical need for structuring and viewing the information content in such document spaces from different perspectives provided by disciplines, business units, market conditions, and individual preferences. No single classification structure can accommodate this diversity of needs over space and time. Moreover, such structures need to evolve in a bottom-up fashion. This was very evident at every level of information and knowledge processing by the participants in the study. This requirement can be addressed through the provision of the following features: 31
-
Development of an integrated system for classification and indexing of documentation. Ease of creation of personal but shareable repositories that use multiple classifications. Ability to search across multiple repositories.
These features are interdependent as the evolution of a rich classification system is facilitated by the ease with which personal yet shareable repositories can be created. Such a classification system that can handle the diversity of perspectives across the organization is needed to address the problem of searching across multiple fragmented document repositories. b. Ameliorating the effects of information overload Information overload is one of the most commonly cited problems. Its causes are many fold, including multiple, disparate databases and the absence of personal control over much of the information that is ‘pushed’ at people. The disorientation that arises from the difficulties people experience as a result of such over-abundance of information, much of which may be irrelevant is well documented. The following suggestions are offered in this connection: Documents to be made available at different levels of detail with the option to view any or all (title, cover page with meta-data (author, date, subject, keywords etc.), abstract only, abstract plus body of the document) Facilitating optimal awareness: This can be achieved through the design of solutions that experimentally model the information needs of users and notifies each one as newer and relevant documents become available, say on an intranet server or a document database. The users are thus in a position to ‘pull’ relevant information and view it at the required level of detail.
c. Dealing with the problem of information seek-provide asymmetry Seek-provide asymmetry refers to the commonly observed problem in knowledge sharing and management systems of the disparity between a person’s willingness to access and use the information or knowledge contributed by others and the contributions that s/he is prepared to make to the shared pool. Such asymmetry occurs at the level of contributing keywords for documents contributed as well as for generating the documents or databases themselves. There is a clear need to develop effective metrics for measuring the knowledge contributions of individuals. Furthermore, organizational policies need to evolve to encourage provision and sharing of good content that can in turn trigger further knowledge accretion.
32
d. Management of Language Management of language is an often overlooked issue. However, management of language is critical to organizing, searching, systematizing, and sharing of knowledge in any organization. By language we refer to all lexical, terminological, semantic, and linguistic differences that can be a major impediment to good communication and effective knowledge sharing in all large, global companies. Language represents the common ground that is defined by a set of people in the context of their interactions. This is a shifting common ground depending on people, the nature of the market, and technological changes. Documenting, clarifying, and establishing correspondences in the diverse uses of the language on an ongoing basis and the design of tools for supporting this need are key activities that deserve consideration.
e. Supporting the formation and growth of people networks and the evolution of communities of practice. The key requirement for building such networks is the ability to identify and learn more about people who can provide particular kinds of information or assistance. Matching the needs with the profiles of people is not an easy task. With the increasing reliance on electronic means of collaboration, organizations have initiated electronic people finders and other databases to supplement the paper-based directories. However, the benefits from such resources are often stymied by the absence of critical information such as history of previous project and country experiences. As well, these resources tend to become obsolete over a short period. User ownership and control over the information in the people finder databases can go a long way to maintaining its currency and value. An important requirement to facilitate the development of people networks is the fostering of user-owned and user-updated people finders and profiles backed by organizational policies for its maintenance.
f. Methods and policies for the integration of organically evolving systems. An important observation in our study was the emergence of organically evolving systems. This runs counter to the traditional systems development model based on architectures that are essentially top-down and based on global priorities of the business unit or the firm. However, bottom-up systems and tools that address typically immediate and local needs may embed knowledge that extend beyond the location of creation. When the knowledge pertaining to such an effort is not shared, there is always the risk of it being reinvented at other locations or business units. One potentially useful direction is provided by the distributed, open-source 33
software model augmented with a community tools repository with which people can develop, access, and share such knowledge. In any case, a clear policy for dealing with organically grown tools is critical to leverage effort and knowledge across geographically distributed centers.
g. Long term electronic archiving. As noted earlier, the transition from traditional paper based systems to digital systems has led to the disintegration of a well-structured archival process. In Du Pont this practice was well structured and organized in the paper world which allowed for the systematization of a number of distinct bodies of knowledge. However, no new systematic methods have been developed to address the archival problem in the electronic arena. The design of systems for knowledge sharing and management needs to address this issue from the outset. Failure to do so risks losing large amounts of unrecoverable information and knowledge.
h. Individual and group rewards The design of reward systems for contributing to and utilizing the shared knowledge repositories is one of the more difficult issues that organizations are faced with. Such systems need to explicitly acknowledge the group nature of much of the knowledge production. Both individuals and groups (including communities of practice) will have to be treated as distinct entities and rewarded accordingly. Also, both the amount and the degree of use of the
contributed knowledge have to be factored into assessments of knowledge professionals’ performance.
6.0 Conclusion
This paper contributes to the growing literature on the organizational processes and mechanisms for knowledge sharing and management especially in large, global companies. By drawing on and synthesizing related bodies of writing, we have attempted to further the debate on what constitutes organizational knowledge. The diverse forms in which such knowledge manifests leads to a review of some of mechanisms and systems for knowledge sharing and management in contemporary organizations and the development and elaboration of a typology of such mechanisms. A detailed field study exploring the major knowledge sharing issues, practices, constraints, and mechanisms was carried out in selected departments 34
and business units of a large, multinational company with knowledge-intensive operations. The field study lends qualified support to the proposition that exclusive focus on certain types of knowledge can be counter-productive. Knowledge sharing and management systems and processes in large global companies need to be integrative and flexible enough to facilitate the dynamic interplay between different forms of knowledge. A set of system requirements targeted to enhance these capabilities is outlined.
References Antonelli, C. (1996), “Localized Knowledge Production Processes and Information Networks”, Journal of Evolutionary Economics, 6, pp. 281-296. Antonelli, C. (1999), “The Evolution of the Industrial Organization of the Production of Knowledge”, Cambridge Journal of Economics, 23, pp. 243-260. Argyris, A. and Schon, D. (1978), Organizational Learning, Reading, MA: Addison-Wesley. Arrow, K. (1969), “Classificatory Notes on the Production and Transmission of Technical Knowledge”, American Economic Review, 59, pp.29-35. Arrow, K. (1994), “Methodological Individualism and Social Knowledge”, American Economic Review, 84, pp.1-9. Ashton, W.B. and Stacey G.S.(1995), “Technical Intelligence in Business: Understanding Technology Threats and Opportunities”, Int. J. of Technology Management, 10(1). Bartlett, A.B., and Ghoshal, S. (1989), Managing Across Borders: The Transnational Solution, Boston, MA : Harvard Business School Press. Brown, J.S. (1991), “Research that Reinvents the Corporation”, Harvard Business Review, Jan.-Feb. Brown J.S. and Duguid, P. (1991), “Organizational Learning and Communities of Practice: Toward a Unified View of Working, Learning, and Organization”, Organization Science, 2(1), pp.40-57. Brown S.L. and Eisenhardt K.M.(1995), “Product Development: Past Research, Present Findings, and Future Directions”, Academy of Management Review, 20(2), pp.343-378. Clark, K.B. and Fujimoto, T. (1991), Product Development Performance, Boston, MA: Harvard Business School Press. Cohendet, P., Kern, F., Mehmanpazir, B., and Munier, F. (1999), “Knowledge Coordination, Competence Creation, and Integrated Networks in Globalized Firms”, Cambridge Journal of Economics, 23, pp.223-241.
35
Coyne, R. (1997), “Language, Space, and Information”, In Droege, P. (ed.), Intelligent Environments, Amsterdam: Elsevier, pp.495-516. Culpan, R, (1983), Multinational Strategic Alliances, Binghampton, NY: Howarth Press Inc. David, P.A., (1993), “Knowledge Property and the System Dynamics of Technological Change”, Proceedings of the World Bank Annual Conference on Development Economics, Washington DC: The World Bank. Davenport, T.H., De Long D.H., and Beers, M.C. (1998), “Successful Knowledge Management Projects”, Sloan Management Review, 39(2), pp.43-57. Dewey, J. (1929), The Quest for Certainty, New York: G. P. Putnam. Doz, Y., and Prahalad, C.K.(1991), “Managing MNCs: A Search for a New Paradigm”, Strategic Management Journal, 12(5), pp.145-164. Dretske, F. (1981), Knowledge and the Flow of Information, Cambridge, MA: MIT Press. Edstrom, A., and Galbraith, J. (1977), “Transfer of Managers as a Coordination and Control Strategy in Multinational Organizations”, ASQ, 22(2), pp.248-263. Eliasson,G., Foster, S., Lindberg,T., Pousette,T., and Taymaz, E. (1990), The Knowledge Based Information Economy, Stockholm : The Industrial Institute for Economic and Social Research. Galbraith, J.R. (1974), “Organization Design: An Information Processing View”, Interfaces, 4(3), pp.30-36. Goldman, A.I. (1999), Knowledge in a Social World, Oxford : Oxford University Press. Holliday, C. (1998), “To Our Shareholders”, Du Pont 1998 Annual Report, p.3. Hounshell, D.A. and Smith J.K, Jr. (1988), Science and Corporate Strategy: Du Pont R&D, 1902-1980, Cambridge: Cambridge University Press. Inkpen, A.C., (1998), “The Management of Knowledge in International Alliances: The Role of Collaborative Process”, Carnegie Bosch Institute Working Paper, available athttp://cbi.gsia.cmu.edu/work/inkpen.htm. Jain,R.K., and Triandis, H.C.(1990), Management of R&D Organizations : Managing the Unmanageable, New York: Wiley and Sons. Jorde, T.M., and Teece, D.J.,(1990), “Innovation and Cooperation: Implications for AntiTrust”, Journal of Economic Perspectives, 4, pp.75-96. Kerssens-Van Drongelen, I.C., de Weerde-Nederhof, P.C., Fissher, O.A.M.. (1996), “Describing the Issues of Knowledge Management in R&D: Toward a Communications and Analysis Tool”, R&D Management, 26(3), pp.213-230. 36
Kleiner, A. and Roth, G. (1997), “How to Make Experience Your Company’s Best Teacher”, HBR, Sept-Oct., p.172-178. Kogut, B. and Zander, U. (1992a), Knowledge of the Firm, Combinative Capabilities, and the Replication of Technology, Organizational Science, (3), pp. 383-397. Kogut, B. and Zander, U. (1992b), Knowledge of the Firm, Technology Transfer, and the Theory of the Multinational Corporation, Working Paper, Institute of International Business, Stockholm School of Economics, December. Kumar, K. (1990), “Post-Implementation Evaluation of Computer-Based Information Systems: Current Practices”, Communications of the ACM, 34(2). Kvalnes, F.H. (1997), “History of Managing Technical Information at Du Pont”, Internal Du Pont Note. Machlup, F. (1980), Knowledge: Its Creation, Distribution, and Economic Significance, Princeton, NJ: Princeton University Press. Manheim, M.L. (1992), “Global Information Technology: Issues and Opportunities”, International Information Systems, 1: pp.38-67. March, J.G. and Simon, H.A. (1959), Organizations, New York: John Wiley. Ndim group (1995), “An Environment for Realizing Computer Supported Collaborative Work”, Technical Report Number 05-93-95, Engineering Design Research Center, Carnegie Mellon University. Nelson R.R. and Winter S.G. (1982), An Evolutionary Theory of Economic Change, Cambridge, MA: Harvard University Press. Nonaka, I. (1991), “The Knowledge Creating Company”, HBR, Nov-Dec., pp.96-104. Nonaka, I.(1994), “Dynamic Theory of Organizational Knowledge Creation”, Organization Science, 5(1), pp.14-35. Nonaka, I. And Takeuchi, H.(1995), The Knowledge Creating Company, New York: Oxford University Press. Polanyi, M. (1967), The Tacit Dimension, Garden City, NY: Doubleday Anchor. Prokesch, S.E., “Unleashing the Power of Learning: An Interview with British Petroleum’s John Browne”, HBR, Sept-Oct., pp.146-172. Quine, W.O. (1953), “Two Dogmas of Empiricism”, in, From a Logical Point of View, Cambridge, MA: Harvard University Press.
37
Reddy, M. (1979), “The Conduit Metaphor: A Case of Frame Conflict in our Language About Language”, In Ortony, A.(ed.), Metaphor and Thought, Cambridge, UK: Cambridge University Press, pp. 284-324. Schultz, J.L. and Wilks, E.S. (1997), “Multiple-Level Polymer Registration in the Du Pont SCION Database, 37(2), pp. 165-170. Scrivner, N.C. (1997), Du Pont Makes Engineering Standards Interactive and Intranet-Ready, Scientific Computing and Automation, October, pp. 31-33. Simon, H.A. (1997), The New Science of Management Decision (Revised ed.), Englewood Cliffs, NJ: Prentice Hall. Simon, H.A. (1996), “Knowledge and the Time to Attend to it”, Paper presented at the Carnegie Bosch Institute International Conference on High Performance Global Companies, April 21, 1996, Boca Raton, FL, available athttp://cbi.gsia.cmu.edu/work/96-2.htm Souder, W.E., Nashar, A.S., and Padmanabhan, V. (1990), ‘A Guide to the Best Technology Transfer Practices”, Technology Transfer, Winter-Spring. Takeuchi, H., and Nonaka, I. (1986), “The New Product Development Game”, HBR, Jan-Feb. Van Fenema , P. (1997), “Management of Globally Distributed Projects : Experiences form the GOLDD Team”, Proceedings of the International Conference on Information Systems, Atlanta, GA, pp.474-475. Wheelwright S.C., and Clark, K.B.(1992), Revolutionizing Product Development, New York: Free Press. Zeleny, M., Cornet, R.J., and Stoner, J.A.F. (1990), “Moving form the Age of Specialization to the Era of Integration”, Human Systems Management, Vol.9, pp.153-170.
Associate Professor Joseph G. Davis Joseph Davis is Associate Professor of Information Systems and Coordinator of the Decision Systems Laboratory at the University of Wollongong, Wollongong, NSW, Australia and a Visiting Research Scientist at the Institute for Complex Engineered Systems, Carnegie Mellon University, Pittsburgh. He has previously served on the Information Systems faculties at The University of Auckland, Auckland, New Zealand and Indiana University, Bloomington, IN. He has also been a Visiting Professor at the Czech Management Centre, Celakovice, Czech Republic (1996), Katz Graduate School of Business, University of Pittsburgh (1994-95) and Visiting Researcher at the Engineering Design Research Center in 1990-91, 1994-95. He has a Ph.D in Information Systems and Strategic Management from the Graduate School of Business, University of Pittsburgh and an MBA from the Indian Institute of Management, Ahmedabad, India, prior to which he completed his Bachelor of Science in Mathematics.
38
Joseph’s research interests span information modelling and management and the design and implementation complex application systems to support managerial decisions and engineering design. His research has been published in some of the major journals and presented in the important conferences in the field. He has served on the Program Committees of several important international conferences including the Internationl Conference on Information Systems (ICIS). He has actively collaborated with the ndim group on research into the design, implementation, deployment, and testing of computer-based environments to support design work. His research is currently supported by grants from Carnegie Bosch Institue for International Management and the Australian Research Council under the Strategic Partnership: Industry, Research, and Training (SPIRT) scheme in collaboration with BHP Information Technology Division, Sydney.
Dr Eswaran Subrahmanian Eswaran Subrahmanain is a Principal Research Scientist at the Institute for Complex Engineered Systems (ICES)at Carnegie Mellon. He has a Ph.D. in Urban and Public Affairs with emphasis on Information Systems from CMU. He also has a B.Tech. (Hons) in Chemical Engineering and a M.S. in Computer Science. He has been at ICES and its precursor EDRC (the NSF center for engineering design) since 1987. His research interests are in the area of computer-supported decision making and engineering design. He was the head of the team from Carnegie Mellon to develop a Technical Data Manager that integrated he views of information at the ALCOA Technical Center across divisions. He was the principal investigator in the Asea-Brown Boveri Engineering Design System project for the design of power transformers (1991-1994). Since, he has worked on information flow analysis of the engineering of many of the products with ABB Corporate Research, Norway. He was the lead investigator of the information systems for Product development project funded by Robert Bosch Gmbh (1996-1998). He is currently the principal investigator of the Engineering Process Information Systems in the design of transit systems for ABB-Daimler Benz Transportation(ADTRANZ).In an ARPA project collaborative systems for assembly based design he is the co-principal investigator with Profs. Talukdar and Siewiorek (1996-1999). He is a co-principal investigator of a project to build a team-based design environment, n-dim, to support design by aiding in the formulation, capturing and sharing diverse views which has been used in the creation of prototypes for different work contexts including our own. He is the author of over 30 technical papers in the areas of design research, use and applications of information technology in design. He spent 6 weeks at the invitation of Danish National Research lab (RISOE) in Denmark to study the state of the art in Computer Supported Collaborative Work in Europe. He served on the editorial board of AIEDAM . He is currently in the editorial board of Research in Engineering Design Journal.
Professor Art Westerberg Arthur Westerberg is currently University Professor and the Swearingen Professor of Chemical Engineering at Carnegie Mellon University, Pittsburgh. He received his B.S. from Minnesota, M.S. from Princeton, and Ph.D. from Imperial College, University of London, all in Chemical Engineering. He has served as Director of the Design Research Center (78-80), Chemical Engineering Department Head (80-83), and Director of the EDRC, an NSF funded Center(86-89). His research interests are in engineering design, specifically in design 39
synthesis, analysis, optimization, environments, and applications of expert systems. He is the recipient of awards from AIChE (CAST Division Award, 83; Walker Award, 87; Institute Lecture, 89; McAfee Award, Pittsburgh Section, 90, AICHE Fellow, 1993, Founders Award, 1995), E.V. Murphee Award from ACS, 97 and from ASEE (Chemical Engineering Division Award, 81). He is a member of the National Academy of Engineering. His membership in societies and organizations include Sigma Xi; Member of National Research Council Committee on Chemical Engineering Frontiers, head of Panel on Computer Assisted Process and Control Engineering, 1984 - 1988; American Institute of Chemical Engineers, past Program Co-Chairman (1975-77) and chairman (1977-79) of area 15a; member, co-chair, chair of National Awards Committee, 1988-94; American Society of Engineering Education, session chairman 1970; Chemical Engineering Program Chairman summer meeting, New Orleans (9 events/sessions) 1979; CACHE, 1971-1980, secretary 1971-73, FOCAPD-II Conference Chairman, Snowmass, CO, June 19-24, 1983; Member National Academy of Engineering since 1987. He is on the journal advisory boards of Computers and Chemical Engineering Journal, 1982-present; Chemical Engineering Reviews, 1982-present; AI EDAM, 1987-present; Research in Engineering Design, 1988-present; Journal of Optimization Theory and Applications (JOTA), 1989-present; Industrial & Engineering Chemistry Research, 1993-1996.
40
doc_779346490.pdf