What’s In an Organizational Routine?
“Unpacking” Knowledge Management
Daniel Z. Levin
Organization Management Department
Faculty of Management
Rutgers University
111 Washington Street
Newark, NJ 07102
(973) 353-5983
Fax (973) 353-1664
August, 1999
Under review, Organization Science
Presented at the 1999 annual Academy of Management meetings, Chicago, IL
What’s In aN Organizational Routine?
“Unpacking” knowledge Management
ABSTRACT
This paper looks inside the “black box” of organizational learning and organizational knowledge flows to show how two critical, knowledge-intensive organizational routines—the technology roadmapping (planning) process and the (intra-company) technology transfer process—operate in practice in multiple companies. Whereas much of the literature depicts organizational learning as a kind of sterile, anonymous process of building organizational skills and competencies, this research finds organizations that have the knowledge and skills in place, yet are still not fully successful at implementing a knowledge-intensive routine. Based on a cross-case comparison at multiple firms, this theory-building study suggests that favorable outcomes for a knowledge-intensive routine (or, process) require not just an organizational ability to absorb knowledge (e.g., via technical competence, having a common language, and an intelligent reduction of complexity), but also a high degree of comfort with the situation (e.g., comfort with the knowledge, the people, the process itself) and the use of power (including both hierarchical and structural power) within the organization to make the knowledge flow happen. For managers and scholars alike, these three theoretical components—ability, comfort, and power—and their sub-components may prove useful as leading indicators of success, rather than the usual lagged measures. This framework is also designed to yield a better understanding, from a theoretical perspective, of how knowledge management works in practice.
Both scholars and practitioners have lately taken great interest in the notion of knowledge management and organizational learning (Crossan, Lane, & White, 1999; Hansen, Nohria, & Tierney, 1999; Miner & Mezias, 1996), an area with a long history (e.g., March & Simon, 1958), but where renewed interest and new approaches have added to our understanding. Especially in today’s faster-paced and information-intensive business environment, the ability of a firm to manage and improve the flow of knowledge is critical to all aspects of performance. In particular, a strong organizational capability for organizational learning has been identified as perhaps the only long-term source of dynamic core competencies (Lei, Hitt, & Bettis, 1996) and sustainable strategic advantage for firms (Collis, 1994; Grant, 1996; Levitt & March, 1988; Myers, 1996; Stata, 1989). The importance of organizational learning and knowledge is especially critical in the research and development (R&D) domain, which is perhaps the most knowledge-intensive area in large corporations. R&D, after all, is in the business of absorbing, creating, adapting, and transferring new ideas, knowledge, and technology to the rest of the company (Cohen & Levinthal, 1990). In short, there is no R&D without the flow of knowledge. For “R&D produces one product only—knowledge” (Roussel, Saad, & Erickson, 1991: 67). This study examines two critical knowledge-intensive processes (or, routines) in the R&D domain. By focusing on specific, and crucial, work processes—in this case, the technology roadmapping (planning) process and the technology transfer process, defined and described in more detail below—this research is able to ground the theories and concepts of knowledge and learning in what actually goes on in organizations.
The broad research question addressed by this paper is why do some organizations fail to manage the flow of knowledge and learn, while others succeed? In the past, when attempting to answer this question, organizational learning theory has been geared largely towards issues such as organizational information processing, decision making, and learning skills (e.g., Huber, 1991). In short, the field’s main focus—both in its theoretical and empirical work—has been on the ability of organizations to learn. This paper builds on this work but goes a step further to answer the question: why is it that some organizations have the knowledge and skills in place, yet they still fail to learn?
The answer, this paper argues, based on in-depth field research of 14 cases at four large companies, is that organizational learning is not a sterile, anonymous process of just skills development. That is, while important, the ability to absorb knowledge—what researchers have also called “absorptive capacity” (Cohen & Levinthal, 1990)—is not always enough for a knowledge-intensive process to work well if project participants are too uncomfortable or if project supporters have too little power. To make these three concepts of ability, comfort and power more concrete, we will explore some of the nuances and dynamics of the two or three most critical sub-categories within each of these three areas. As we will see, in some cases, people in organizations had the ability to manage the flow of knowledge, but they did not have enough comfort with the situation to do so. This lack of comfort could be with the knowledge in question, with other departments, or with the particulars of a work process. In all three instances, however, this discomfort led people in the organization, in effect, to not want to learn. Indeed a useful way to understand any organizational action, including the flow of knowledge, is to focus on both the desire and ability of the organization or of its members to perform that action (Levin & Shortell, 1996; Zajac & Kraatz, 1993; Zajac & Shortell, 1989). When we say that organizational learning and the flow of knowledge can only occur if the people involved want to and are able to do so, it seems, in retrospect, almost self-evident. Nevertheless, theorists and practitioners—Szulanski (1996) is a notable exception—typically emphasize the ability to learn (e.g., Levitt & March, 1988; Nonaka, 1994; Teece, 1977; Walsh & Ungson, 1991), avoiding the question of whether or not people even want to learn in the first place. In addition, this study’s findings suggest that a third factor also comes into play. For it may not be enough for organizational learning to take place if there is only an ability to absorb new knowledge and a high degree of comfort in doing so. Rather, supporters had to have the power and influence in the organization to make the knowledge flow happen, to overcome inertia (Hannan & Freeman, 1984) and resistance (Day, 1994; Dougherty & Hardy, 1996; Maidique, 1980).
A Process-based Approach
One exciting and useful approach to studying how learning takes place in organizations is to study systematically the generic routines (or, processes) that organizations use (Levin, Radnor, & Thouati, 1997; Malone et al., 1999; Pentland & Rueter, 1994). Routines in the technology management domain, for example, would include the technology strategy-making process, the R&D portfolio management process, and the technology transfer process. This approach is especially fruitful because organizational routines form the basis for organizational learning (Levitt & March, 1988), yet “much remains to be done at both the empirical and the theoretical level” (Grant, 1996: 384). Note that the concept of a routine is not used here, as it is by some theorists elsewhere, to indicate a skill or capability for action (Nelson & Winter, 1982); rather, it is the actual action itself. Focusing on the routine, or process—defined here as a repeated set of related activities—results in research that (1) better reflects the reality of organizational life, (2) better understands the terminology, jargon, and thought patterns (i.e., the “cognitive categories”) that organizational members use to describe their organization and how work gets done, (3) better discerns how firms adapt in a dynamic way to high-velocity environments, (4) better connects strategic (macro) and project (micro) levels of analysis (Burgelman, 1983), and (5) better translates rigorous research into findings that are useful to practitioners, a goal recently endorsed by several organization science leaders (Galunic, 1997; Hambrick, 1994; Mowday, 1997).
The issue of which processes are particularly critical in the technology management domain, and how, has been the subject of an ongoing research partnership of academic scholars and of practitioners in a half-dozen large technology-oriented companies—under the auspices of the National Center for Manufacturing Sciences (NCMS) consortium. As a result, the focus of this research has been on large companies, which are responsible for most industrial R&D (Raber, 1997). This partnership is in accord with recent calls by academics for greater practitioner/scholar collaboration in the area of organizational learning (Miner & Mezias, 1996). While some preliminary work focused on documenting the two dozen or more major technology management processes and their inter-relationships, the research reported here centers on investigating the underlying mechanisms and factors—from a theoretical perspective—that are at work in a small number of key processes (routines). Note that the goal here is not to take a reengineering approach and provide a step-by-step “process map” of how technology roadmapping or technology transfer does or should take place; that is perhaps a job best left to consultants and practitioners themselves. Rather, the research goal is to look for central, underlying, theoretical elements of a process; i.e., what makes the process “tick.”
The two processes chosen for this research study were technology roadmapping and technology transfer. For this study we define technology roadmapping as the process which produces a “roadmap” documenting the technological intentions of an organization and communicating them to interested and affected parties (Galvin, 1998; Willyard and McClees 1987). In practice, roadmapping participants typically met, as part of a cross-functional effort, over a period of weeks or months to develop a plan for what technologies would be needed for a given product line in the next few years. Although people had multiple goals for this knowledge-intensive process, the main one was to help overcome the age-old problem in organizations of “not [being able to] link technological and market issues” (Dougherty, 1992). The technology transfer process is defined for this study as the transfer of a technical artifact within the R&D domain (e.g., from central R&D to a business unit) and the accompanying flow of knowledge among those involved. The focus here is primarily on intra-company transfer (Leonard-Barton & Sinha, 1993), and not, as the term is sometimes used elsewhere, the transfer of a mature technology from developed to developing countries or between companies (Teece, 1977). Note, too, that this study is more in the tradition of research on commercialization of new technologies, especially point-to-point transfer, than in the diffusion-of-innovation research tradition from economics, sociology, and marketing (Leonard-Barton, 1990; Rogers, 1995).
Both processes were chosen, in part, because, after qualitative discussions with participating practitioners, these areas seemed to be of great interest and importance yet did not always function well. More importantly, from a theoretical standpoint, both processes stood out as particularly knowledge-intensive and critical to the overall flow of knowledge in what is already a knowledge-intensive domain. That is, both processes require the transfer of knowledge across functional boundaries and both are firmly connected, as both inputs and outputs, to other routines in the technology management domain. For example, technology roadmapping is intimately connected to routines for determining strategy, for managing the overall R&D portfolio, for selecting individual R&D projects, for assessing customers’ technology needs, for planing product changes, for monitoring the external environment, etc. Because of all these inter-connections, technology transfer and roadmapping appear to be “linchpin” processes, critical to the overall functioning of the system. So while other processes might also have been interesting to study in depth (and would make fertile ground for future study), these two appeared to be particularly knowledge-intensive, and thus a fruitful area for studying the theoretical underpinnings of knowledge-related processes. By studying these processes, we delve into the “heart” of an organization’s efforts to manage the flow of knowledge.
Methods
This study’s findings are based primarily on in-depth fieldwork at four multi-billion dollar companies, identified here by pseudonyms, participating in the research program. The companies, which all belong to the National Center for Manufacturing Sciences (NCMS) consortium, are in different industries but all have a connection to technology and manufacturing. Note that the sample of companies is not random because the research goal is not to draw statistical inferences but rather to generate grounded theory (Eisenhardt, 1989; Glaser & Strauss, 1967). Two researchers jointly conducted interviews of 79 managers and engineers in R&D labs and in the business units of these large multinational companies. These in-person interviews (in addition to a small number of international phone interviews) took place in 1997 and early 1998.
For each process, the research team developed a detailed 5-7-page position paper and interview guide. This document, which included a description of the process and a list of open-ended questions (e.g., what are the goals for this process? how are you involved? what happened during implementation?), was derived from some preliminary research on technology management processes in general. The company contact within each of the four companies initially directed the research team to participants in specific cases and to people familiar with the process in general. An effort was made to interview both successful and struggling (or failed) efforts within each company and to speak with both the knowledge senders and receivers. To enhance reliability in the cross-case comparisons, the case material in Tables 2 and 3 reflects only those case studies for which there were multiple participants interviewed and an ample amount of interview data. The interviewers also gathered background material and documentation (e.g., organization charts, reports, presentation slides) from the contact people and interviewees.
During the 12-month period of fieldwork, 79 people (split 60/40% between R&D and business units) were interviewed, totaling 90 hours and 206 pages of single-spaced typed interview notes. Interviews took place on eight visits to company sites in California, Florida, Pennsylvania, Illinois, Ohio, Michigan, New Jersey, New York, and Wisconsin. The number of cases per company varied based on availability of data and practitioner interest in that topic. In all, for technology transfer, there were eight full cases at three divisions of two companies; for technology roadmapping, six cases at four companies. Technology roadmapping was discussed in 74% of all interviews and technology transfer in 69%. About half of the interviewees were managers. The average interview lasted about one hour and 15 minutes. All interview notes were typed separately by each interviewer and then merged into a single document. This document, typically about 50 single-spaced typed pages per company-visit was then summarized in a 5-7 page trip report. To supplement the formal visits, this research also used non-traditional approaches such as roundtable discussions among academics and practitioners from core and outside companies, as well as informal academic/practitioner data analysis discussion meetings.