This section identifies the conditions under which knowledge transfers across orga- nizational units. That is, the section identifies the conditions under which experi- ence in one organization affects another. The section is organized into examining how knowledge transfer is affected by characteristics of the relationship among the organizations, characteristics of the organizational context, features of the knowl- edge being transferred, and dimensions of the transfer process. Consistent with the conceptualization of the organizational context presented in Chap. 2, our discussion of the organizational context is separated into a discussion of the latent and active components of the context.
1. Characteristics of the Relationship Among Organizations
Several characteristics of the relationship between organizations have been found to affect the extent of knowledge transfer. These factors include whether the organiza- tional units share a superordinate identity, how competitive their relationship is, the similarity of the units, and their proximity.
1.1. Superordinate Identity
As noted previously, one very important factor affecting transfer of knowledge across organizations is whether the organizations are embedded in a superordinate relationship. The situation at Airbus in 2007 illustrates the problems that occur in organizations when a superordinate identity is weak or lacking (Clark, 2007). Airbus is a consortium with establishments in four countries: France, Spain, Germany, and Great Britain. The firm experienced enormous problems coordinating across estab- lishments in different countries during the construction of the A380 superjumbo jet. For example, when rear fuselage sections built in Hamburg arrived in Toulouse, it was discovered that the sections lacked the correct wiring. The computer modeling used in Germany and France was found to be incompatible! This problem resulted in significant delays in the production of the A380 and cost the company over six billion dollars. A practice at Airbus was to display the flag of the country of a pre- senter on power point slides during presentations. When Louis Gallois took over as CEO at Airbus in 2007, he banned the practice of displaying country flags because it reinforced identity at the country subgroup level. Instead, Mr. Gallois’ goal was to foster identity at the superordinate level of the firm.
As noted previously, my colleagues and I found that being embedded in a super- ordinate relationship, such as a franchise, increased the likelihood of knowledge transfer (Darr et al., 1995). Similarly, Baum and Ingram (1998) found that Manhattan hotels benefited from their own direct experience and the local experience of hotels that were related to them through belonging to the same chain. Along similar lines, Ingram and Simons (2002) found that kibbutzim benefited from the experience of other kibbutzim that belonged to the same federation, whereas they did not benefit from the experience of those in different federations.
Powell, Koput, and Smith-Doerr (1996) found that biotechnology firms that were linked together in a research and development alliance were more likely to have access to critical information and resource flows that facilitated their growth than firms not engaged in such collaborative relationships. McEvily and Zaheer (1999) found that participation in regional institutions, MEP centers, improved the com- petitive capabilities of small manufacturers. Uzzi (1996) found that clothing apparel firms that were embedded in networks had a greater chance of survival than firms connected to other firms only through “arms-length” ties. Uzzi suggested that learn- ing from one another contributed to the superior performance of firms in the net- work. Burns and Wholey (1993) found that regional and local hospital networks influenced the diffusion of an administrative innovation, matrix management. Thus, being embedded in a franchise, chain, or network relationship facilitates the transfer of knowledge.
Being embedded in a superordinate relationship may affect both the motivation and the communication of participants in ways that facilitate knowledge transfer. Incentives of firms involved in superordinate relationships are typically more favor- able to knowledge transfer than incentives at independent firms. For example, com- petition is usually minimized among organizations that belong to franchises, chains, or networks. These organizations cooperate in certain arenas. The organizations generally trust each other to a greater degree than those not embedded in a network or superordinate relationship (Granovetter, 1985; Uzzi, 1996).
Being embedded in a franchise, chain, or network also provides more opportuni- ties to communicate than are afforded independent organizations (who may even be proscribed from direct communication with each other). Meetings, informal inter- actions, and opportunities to observe each other’s organizations are all more likely to occur among organizations embedded in such a relationship than among indepen- dent organizations. Personnel rotation can occur among the organizations, which often have access to each other’s documents and databases. These communication opportunities provide mechanisms for transferring both explicit and tacit knowl- edge across organizations. For example, Uzzi (1996) noted that more tacit knowl- edge flowed across the firms embedded in a network than across independent organizations. Thus, organizations embedded in superordinate relationships have more opportunities to share information and learn from one another.
Aimee Kane, John Levine, and I built on these field studies as well as studies on the effects of recategorization on intergroup bias (Gaertner & Dovidio, 1989) to examine the effect of a superordinate social identity on knowledge transfer between groups (Kane, Argote, & Levine, 2005). Two groups were induced to feel either that they belonged to the same organization and shared a superordinate social identity or that they did not in the controlled setting of the laboratory. Groups produced origami sailboats in interde- pendent assembly lines in different rooms. Groups were informed that the best-perform- ing group would win a financial prize. Midway through the study, the experimenter informed groups that they would experience personnel rotation and the member who occupied the second position in each group was switched. Unbeknownst to group mem- bers, the two groups had been trained in somewhat different production routines. Both production routines resulted in sailboats that met product specifications but one routine involved fewer steps. Using the routine with fewer steps could increase group productiv- ity. The benefits of the better routine, however, were not immediately obvious.
We hypothesized and found an interaction between whether groups shared a superordinate social identity and the quality of the routine possessed by the new- comer who rotated into the group (Kane et al., 2005). When the groups shared a superordinate social identity, they adopted the superior routine and rejected the infe- rior routine introduced by the newcomer. By contrast, when groups did not share a superordinate social identity, they rejected both the superior and the inferior routines. Thus, sharing a superordinate identity led to a discerning adoption of the routine that had the potential to improve their performance. By contrast, groups that did not share an identity rejected performance-improving routines along with those that would not improve their performance. Thus, we found evidence in the controlled setting of the laboratory that a superordinate social identity affected knowledge transfer.
1.2. Quality of Relationship
The quality of the relationship among organizations also affects knowledge transfer. Szulanski (1996) examined barriers to the transfer of best practices within organizations. One of the factors found to contribute to the degree of knowledge transfer was the quality of the relationship between the donor and the recipient. A poor relationship made it more difficult to transfer best practices.
Competition between units is an even stronger barrier to knowledge transfer than a poor relationship. Drawing on social identity theory (e.g., see Tajfel, 1981; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987), Ashforth and Mael (1989) argued that much of the conflict that occurs between groups in organizations derives from the very existence of different groups. According to social identity theory, individuals’ desires for positive self-evaluation lead them to differentiate between social groups by perceiving one’s own group more favorably than other groups (Abrams & Hogg, 1990). Thus, the desire to enhance one’s social identity leads to more positive per- ceptions of the “in-group” and more negative perceptions of the “out-group.” Although there is some disagreement as to the precise mechanism through which in-group biases occur (Gaertner & Insko, 2000), the in-group bias is very robust (see Brewer, 1979; Dasgupta, 2004; and Hewstone, Rubin, & Willis, 2002, for reviews). Building on social identity theory, Kramer (1991) argued that categorizing indi- viduals into distinctive groups produces intergroup competition in organizations (Messick & Mackie, 1989). Thus, attempts to promote group identity in organiza- tions can also produce intergroup competition. Giving groups distinct names, pro- viding opportunities for members to interact, publicizing the performance of different groups, providing rewards based on the performance of different groups, and other techniques designed to increase group identity are also likely to increase intergroup competition. Intergroup competition, in turn, impairs sharing of informa-tion and transfer of knowledge across groups.
Kramer (1991) further noted that organizational settings are particularly condu- cive to the development of competition between groups. In organizational settings, it is often difficult or impossible to assess performance in absolute terms. Instead, orga- nizations rely on relative comparisons. Being perceived as a high-performing depart- ment thus means being perceived as performing better than some other department. This emphasis on relative performance exacerbates competition between units, because often the only way a unit can be perceived as high performing and thereby enhance members’ social identity is to be seen as being better than other units.
Consistent with the competitive nature of relationships in many firms, Menon and Pfeffer (2003) found that individuals preferred to seek knowledge from indi- viduals outside a focal firm than from individuals inside the firm. The researchers attributed this phenomenon to the tendency to denigrate knowledge generated by others within the firm because one saw all the problems developing the knowledge and because accepting knowledge from individuals within the firm could elevate their status.
1.3. Similarity of Contexts
Another key factor that affects transfer of knowledge across groups in organizations is the degree of similarity of contexts. Transfer of knowledge from one situation to another requires some degree of similarity between the two situations for transfer to occur. Much research has been done on transfer at the individual level of analysis. A recurring theme in this research is how difficult it is for an individual to transfer knowledge acquired in one situation to another (Singley & Anderson, 1989). Singley and Anderson (1989) proposed a model of transfer built on Thorndike’s earlier iden- tical elements theory of transfer. According to these models, transfer is based upon the elements shared between tasks. The more elements that are shared, the greater the transfer will be.
Groups typically develop their own idiosyncratic ways of doing work (Levine & Moreland, 1991). Groups may divide tasks in different ways, use technologies somewhat differently, develop different ways of coordinating and communicating, and develop their own unique cultures. Differences across groups are likely to be accentuated by providing groups autonomy in deciding how to accomplish their work and by encouraging them to develop their own culture and task-performance strategies. Differences in how groups accomplish their tasks make it hard to transfer knowledge from one group to another because knowledge acquired in one group may not be applicable to another.
We saw examples of differences developing in how work was done in our study of truck assembly plants (Argote & Epple, 1990). Two of the plants started out with the identical technology. Over time, one of the plants drastically changed how it used the technology to override key features. After these changes were made, man- agers at the other plant argued that much of the knowledge acquired at the first plant was not relevant for their operation because the technologies of the two plants were so different.
Research on situated cognition also implies that knowledge may be difficult to transfer across work groups. In this tradition, knowledge is seen as situated—highly dependent upon the particular constellation of people, machines, and conditions that exist at the work site (e.g., see Hutchins, 1991; Lave, 1991; Suchman, 1987). Individuals learn through intensive apprenticeships that expose them to the idiosyn- cratic conditions that exist in their work settings. Brown and Duguid (1991) pro- vided compelling examples of situated cognition in their study of service technicians. Training manuals were not very helpful for these technicians. Instead they learned their jobs through interacting with other service technicians, customers, and the machines. Stories that other technicians told were often more helpful in diagnosing what was wrong with a machine than the machine’s error code because the stories captured idiosyncratic local conditions while the error codes did not. Idiosyncratic local conditions make it very difficult to transfer knowledge across groups.
We have seen instances where local conditions limited the potential of knowl- edge transfer in our research. For example, one of the organizations we studied was fortunate to have a particularly gifted manager. Over time, the manager assumed more and more responsibilities. Micro changes in the distribution of responsibilities at the organization accumulated to the point where its macro structure started to dif- fer from its sister organizations. These differences in expertise and in the division of labor at the two organizations decreased the relevance of some of the knowledge acquired in one organization for another.
Other studies have found similarity to be an important predictor of the extent to which firms monitor and imitate each other. For example, Porac and Thomas (1994) found that similar retail organizations were more likely to monitor each other than dissimilar organizations. In a longitudinal simulation of student groups, Baum and Berta (1998) found that groups were more likely to imitate other student groups that occupied a similar market position.
1.4. Geographic Proximity
Being close geographically has also been found to facilitate knowledge transfer. Galbraith (1990) analyzed 32 different attempts to transfer manufacturing technol- ogy from one facility of an organization to another. He examined the amount of time it took to increase productivity at the recipient facility to the level achieved at the donor site prior to transfer. Galbraith found that the time it took for productivity to recover was slower when the organizations were geographically far apart.
The results of our research on inter-shift transfer of knowledge also suggest that geographic proximity facilitates knowledge transfer. Results from our study of transfer of knowledge across shifts in a manufacturing facility indicated that knowl- edge acquired during the period of one-shift operation carried forward quite rapidly to the period of two-shift operation (Epple et al., 1996). The carryforward of knowl- edge was almost complete within 2 weeks of the second shift’s start-up. We found a more rapid and more complete degree of knowledge transfer in this study of inter- shift transfer than in studies of knowledge transfer across geographically dispersed organizations. It seems likely that the greater proximity of the two shifts facilitated knowledge transfer between them.
Research on transfer of knowledge in nuclear power plant operation also under- scores the importance of proximity for knowledge transfer. Lester and McCabe (1993) examined how the performance of nuclear reactors in the USA and France varied as a function of industry structure. The researchers found that transfer of knowledge across reactors was greatest when the reactors were located at the same site. This finding explained much of the performance advantage of France, which typically builds four reactors at the same site, relative to the USA, where the pre- ponderance of units were built at sites where there is at most one other reactor.
By contrast, Darr and Kurtzberg (2000) found that geographical proximity was not a predictor of knowledge transfer in pizza stores in Great Britain, once similarity of store strategy was taken into account (cf. Burt, 1987). Darr and Kurtzberg (2000) found that stores were more likely to learn from other stores following a similar than a dis- similar strategy. Geographic similarity and customer similarity were insignificant pre- dictors of knowledge transfer when strategic similarity was taken into account. Along similar lines, Borgatti and Cross (2003) found that the effect of geographic distance on knowledge sharing was mediated by transactive memory. Once transactive memory was taken into account, the effect of geographic distance was reduced significantly.
Further research is needed on the role of geographic proximity and similarity in knowledge transfer. This research would benefit from specifying (and testing) the underlying processes through which the factors affect knowledge transfer. For example, does proximity facilitate communication and, thereby, improve knowl- edge transfer? Alternatively, are proximate organizations more likely to be similar and, thus, have more relevant knowledge to share? A greater understanding of the processes that mediate the relationship between proximity and knowledge transfer is needed.
2. Characteristics of the Latent Context of Organizations
Characteristics of an organization, such as its size or success or status, also affect the likelihood that it will be imitated by other organizations. In a study of imitation among firms in their choice of investment banker, Haunschild and Miner (1997) found that organizations were more likely to imitate firms with exceptional perfor- mance than to imitate firms with average performance. Somewhat surprisingly, firms imitated both the best and the worst performers. The latter effect was particu- larly pronounced for deals completed during the most recent year, when publicity might have the greatest impact (Haunschild & Miner, 1997). Thus, imitation of firms with the high premiums (i.e., the worst deals) could reflect the salience of these investment bankers rather than the quality of their deals. Haunschild and Miner (1997) also found that firms were more likely to imitate large than small firms and to follow a strategy when it was used by many rather than few other firms in the industry (see also Burns & Wholey, 1993). Baum and Berta (1998) found that stu- dent groups were more likely to copy successful than unsuccessful groups. Sine, Shane, and Di Gregorio (2003) found that knowledge developed by high-status organizations was more likely to be used by other organizations than knowledge developed by a low-status organization.
An organization’s structure and practices also affect knowledge transfer across units within the organization. Tsai (2002) found that social interaction fostered knowledge transfer across units, while centralization impaired transfer. Collins and Smith (2006) found that human resource practices such as group incentives and performance based on growth were positively related to an organization’s social climate, which in turn was positively related to organizational performance.
Characteristics of the recipient organization, such as its “absorptive capacity,” can also affect the extent of knowledge transfer. Cohen and Levinthal (1990) defined absorptive capacity as the ability of a firm to recognize the value of external infor- mation, assimilate it, and apply it. Cohen and Levinthal argued that absorptive capacity, which is largely a function of the firm’s level of prior related knowledge, is critical to innovation. In an empirical study of the transfer of best practices within a firm, Szulanski (1996) found that high absorptive capacity on the part of recipients facilitated the transfer of best practices. Volberda, Foss, and Lyles (2010) reviewed and integrated the vast literature on absorptive capacity.
In a related vein, Rothwell (1978) described the many problems that develop when the technology to be transferred is beyond the understanding of the recipient organization. Similarly, Galbraith (1990) found that previous experience with technology transfers minimized the initial productivity loss associated with the transfer of manufacturing technology to new establishments. Furthermore, Hamel (1991) reported that a wide gap in skills between partners in a strategic alliance impaired transfer of knowledge between them. In order to replicate a partner’s skills, a firm must understand the steps between its current capability and that of its partner.
Allen’s (1977) work on technology transfer also underscores the importance of absorptive capacity on the part of recipients and suggests a structure for facilitating it. Allen found that for applied research problems whose solution could not be com- pletely codified in scientific principles, having a “gatekeeper” at the boundary of a group who could communicate with internal and external constituencies facilitated performance. The gatekeeper absorbed knowledge from outside and interpreted it for internal constituencies.
Motivation also matters in knowledge transfer across groups. Zander and Kogut (1995) found that the more competitors were perceived as engaging in developing a similar product, the faster the speed of internal technology transfer. Thus, the fear of being surpassed by competitors enhanced the transfer of capabilities within the firm.
3. Characteristics of the Active Context
The active context includes the members and tools that perform the organization’s task and the networks formed by crossing members, tools, and tasks. Each compo- nent of the active context is now discussed.
3.1. Members
Moving members is a very effective means of transferring knowledge across orga- nizations (Almeida & Kogut, 1999) or organizational units (Kane et al., 2005). Allen (1977) argued that individuals are the most effective carriers of information because they are able to restructure information so that it applies to new contexts. In addition to being able to restructure knowledge, individuals are able transfer tacit as well as explicit knowledge when they move. Berry and Broadbent (1987) found that although experienced individuals were not able to articulate their knowledge, they were able to transfer it to a similar task. That is, the finding that experience on one task improved performance on another task even though indi- viduals were not able to articulate why their performance had improved on the first task suggested that the knowledge was tacit. Individuals were able to transfer tacit knowledge from one task to another. This ability to transfer tacit knowledge across different contexts makes personnel movement a powerful transfer mechanism.
3.2. Member–Member or Social Network
Research on how social networks affect knowledge transfer has increased dramati- cally in recent years. Hansen (1999) found that weak ties were well-suited for trans- ferring explicit knowledge and strong ties were appropriate for transferring tacit knowledge. Focusing on network structure, Reagans and McEvily (2003) found that social cohesion and range facilitated knowledge transfer over and above the effect of tie strength. Phelps, Heidl, and Wadhwa (2012) reviewed the large literature on networks.
3.3. Tasks and the Task–Task Network
Moving task networks or routines from one organizational unit to another is a mech- anism for transferring knowledge. As described in our discussion of the franchise study, we found several examples of routines transferring across stores (Argote & Darr, 2000). Routines are especially likely to be adopted by recipient units when the units share a superordinate identity (Kane et al., 2005).
3.4. Tools and the Tool–Tool Network
Tools are also a mechanism for transferring knowledge. For example, Ashworth et al. (2004) found that the implementation of an information tool facilitated knowl- edge transfer across the geographically distributed units of a financial services firm. Our work on inter-shift transfer of knowledge in an automotive plant suggests that embedding knowledge in tools is a powerful and effective way to transfer knowl- edge (Epple et al., 1996). Moving members along with tools has been found to be more effective than moving tools alone (Galbraith, 1990). Members transfer tacit knowledge to the new context when they move, which complements the knowledge embedded in tools, which is typically explicit.
Organizations have invested in knowledge management systems to facilitate the transfer of knowledge from one unit to another and the retention of knowledge over time. Evidence on the effectiveness of knowledge management systems in the form of document repositories is mixed. Haas and Hansen (2005) found a negative asso- ciation between the number of documents downloaded from a knowledge manage- ment system and the performance of consulting teams. Further, using documents from a knowledge management system was especially harmful for experienced teams and teams facing very competitive environments. By contrast, Kim (2008) found a generally positive effect of a knowledge management on the performance of stores in a retail grocery chain. Further, the positive effect was stronger for man- agers with fewer alternative sources of knowledge, for those remotely located and for those dealing with products that did not become obsolete. The difference between the negative effect of using material from a document repository in the consulting teams and the positive effect in the retail grocery stores could be due to differences in the tasks the two organizations perform. Retail grocery stores seem likely to encounter more routine tasks than consulting firms and therefore might find more useful knowledge in document repositories.
Although document repositories typically capture explicit rather than tacit knowledge, newer generations of tools have the potential to facilitate the transfer of tacit as well as explicit knowledge. For example, online communities and discus- sion groups can facilitate connections among individuals and, thereby, have the potential to enable the transfer of tacit as well as explicit knowledge (Alavi & Leidner, 2001). Hwang, Singh, and Argote (2012) found that as participants gained experience using an organization’s online forum, they shared knowledge with other employees who were located in increasingly distant sites. Further research is needed on new social media tools such as blogs and forums to determine whether, how, and when they are effective in transferring knowledge.
3.5. Member–Task and Member–Tool Networks
Transactive memory systems can facilitate knowledge transfer. Borgatti and Cross (2003) found that a transactive memory system enabled organizations to overcome the effect of geographic distance on knowledge sharing. Lewis, Lange, and Gillis (2005) found that a transactive memory system developed on one task transferred to another and improved subsequent performance, if the tasks had common elements.
4. Characteristics of the Knowledge Transferred
Characteristics of the information being transferred also affect the ease and success of knowledge transfer. In their study of the assimilation of innovations in hospitals, Meyer and Goes (1988) found that characteristics of the innovation itself, such as its observability, were more important predictors of the innovation’s assimilation than characteristics of the organization, its leadership, or the environment in which the organization was embedded. Thus, characteristics of the knowledge being trans- ferred can be particularly important in determining the degree of transfer.
Tacit knowledge or knowledge that is not well understood is more difficult to transfer than explicit knowledge. In their study of factors affecting the speed of transfer of manufacturing capabilities, Zander and Kogut (1995) found that knowl- edge that was codified in documents and software and that could be readily taught to new workers transferred more easily than capabilities not codified or easily taught. Similarly, Szulanski (1996) found that knowledge that was high in “causal ambiguity” was harder to transfer than well-understood knowledge.
The complexity of the information being transferred is also likely to influence the success of the transfer. Galbraith (1990) found that attempts to transfer complex manufacturing technology were associated with higher initial losses in productivity at the recipient organization than attempts to transfer simpler technology. Similarly, Ounjian and Carne (1987) and Rothwell (1978) found that increased complexity reduced the rate of diffusion of innovation. By contrast, Meyer and Goes (1988) found that an innovation was more likely to be assimilated into hospitals when it was complex.
The observability of knowledge also affects its ease of transfer. Meyer and Goes (1988) found that the ease of observing an innovation and seeing its effect influenced its rate of assimilation. Observable innovations were assembled more easily than ones that were more difficult to observe.
The demonstrability of knowledge also affects its transfer. Knowledge that is high in demonstrability is easy to explain while knowledge that is low in demonstra- bility is harder to explain and convince others of its appropriateness (Laughlin & Ellis, 1986). Kane (2010) found an interaction between knowledge demonstrability and the extent to which groups shared a superordinate. When knowledge demon- strability was high, knowledge transferred whether or not members shared an iden- tity. By contrast, when knowledge demonstrability was low, knowledge transfer was greater when members shared than when they did not share an identity.
The information features of the innovation from our study of fast-food franchises that transferred most widely are consistent with these features found to facilitate knowledge transfer. As described in Chap. 3, an innovative method for placing pep- peroni was developed at one of the stores in Southwestern Pennsylvania. The method of distributing pepperoni evenly on a pizza before it was cooked that had worked well for regular pizza did not work well for deep-dish pizza. When the method was used on deep-dish pizza, distributing pepperoni evenly before cooking resulted in a cooked pie with an unappealing clump of pepperoni at the center. A store in Southwestern Pennsylvania discovered an effective method for achieving an even distribution of pepperoni on deep-dish pizza after it was cooked. The pepperoni was placed on the pie in a pattern that resembled spokes on a wheel. As the pizza cooked and the cheese flowed, the pepperoni would be distributed (more or less) evenly on the pizza. This method for distributing pepperoni was not complex. It was observ- able and codified in a routine that could be easily taught to new employees. Thus, the method scored high on features of information found to facilitate knowledge transfer. The pepperoni placement method transferred initially to other stores in the same franchise. A consultant from the parent corporation recognized the effective- ness of the method and promoted it in visits to other franchises and at national meet- ings. The method is now used by almost every store in the corporation. This example illustrates that knowledge that is observable, explicit, and not overly complex trans- fers readily to other organizations.
Although several features of knowledge may be immutable, others may be ame- nable to change. In our study of pizza franchises, some tacit knowledge seemed inherently tacit. Knowledge about how to hand toss pizza, for example, was a kind of tacit knowledge not easily made explicit. By contrast, other knowledge that was tacit could have been codified in more explicit terms that would enable the knowl- edge to transfer more readily. For example, one order taker at a store developed heuristics for sequencing pizza preparation in a way that took advantage of differ- ences in cooking times for different types and sizes of pizza. The heuristic enabled the ovens to be used more efficiently and the pizzas to be prepared more quickly. Although the heuristic remained in the mind of one employee, it could have been codified in a procedure that others could use.
5. Characteristics of the Transfer Process
Characteristics of the transfer process, such as its timing, also affect the extent of knowledge transfer. Organizations seem particularly open to learning from the experience of others early in their life cycle. Thus, there may be “windows of oppor- tunity” (Tyre & Orlikowski, 1994) for transferring knowledge from others.
The results of our shipyard study are consistent with the prediction that organiza- tions are more likely to learn from others at the start of their operation (Argote et al., 1990). We examined transfer of knowledge across 13 World War II shipyards that went into production at different points in time. Shipyards that began production later were found to be more productive initially than those with earlier start dates. Once shipyards began production, however, they did not benefit further from pro- duction experience at other yards. Thus, transfer of knowledge occurred at the start of production but not thereafter.
The results of Baum and Ingram’s work on hotel chains are also consistent with the prediction that organizations are more open to learning at the start of operation. Baum and Ingram (1998) found that Manhattan hotels benefited from the experi- ence of hotels in different chains in the industry up to the time of the focal hotel’s founding but not thereafter. Similarly, in their study of the adoption of farming prac- tices, Foster and Rosenzweig (1995) found that farmers learned from the experience of other farmers and that farmers’ learning from others diminished over time. In a study of success rates at angioplasty surgery, Kelsey et al. (1984) found that organi- zations were most likely to learn from the experience of others when they first began to perform the new surgical procedure.
By contrast, in a study of imitation of which investment banker to use on an acquisition, Haunschild and Miner (1997) found that firms learned from each other on a continuing basis. Along similar lines, we found that pizza stores learned from other stores in the same franchise on an ongoing basis (Darr et al., 1995). The dif- ferent results regarding whether learning is confined to the start of operation or occurs on a continuing basis may be due to differences in prevailing knowledge repositories across the settings. For both the shipyards and the hotels, a significant component of the knowledge was embedded in the physical equipment, layout, and technology of the establishments. By contrast, less knowledge regarding the invest- ment banker decision or pizza production was embedded in “hard” form. Changing physical equipment, layout, and technology may be more costly and disruptive than changing softer forms of knowledge. Future research is needed to examine factors affecting when firms are most open to learn from other firms. The extent to which knowledge is embedded in technology versus softer forms is likely to be a key factor.
The learning mechanism also affects the extent of knowledge transfer. Rulke, Zaheer, and Anderson (2000) contrasted the extent to which three learning channels affected an organization’s knowledge of its own capabilities as well as the capabili- ties of other firms. Three types of learning channels were identified empirically by factor analysis: purposive (learning through deliberate attempts to transfer knowl- edge through company newsletters, formal training programs, and the like), rela- tional (learning from personal contacts both inside and outside the firm), and external arm’s length (learning from trade association publications and newsletters). Rulke et al. (2000) found that purposive and relational (but not external) channels contrib- uted to greater self-knowledge of an organization’s own capabilities. Somewhat sur- prisingly, an organization’s knowledge of the capabilities of other organizations was not affected by any of these learning channels.
Additional aspects of the transfer process affect its success. In his study of attempts to transfer manufacturing technology from one facility to another within a firm, Galbraith (1990) found that the time it took for productivity to recover was faster when coproduction continued at the donor site and when the engineering team from the donor organization was relocated for at least 1 month to the recipient orga- nization. Continuing production at the donor site facilitates the transfer of tacit knowledge that is not written down or embedded in documents and plans. Because production continues at the donor site, the recipient site is able to access the donor’s store of tacit knowledge through observation. Once production is discontinued at the donor site, much of the tacit knowledge would be lost.
Moving personnel to the recipient organization is a powerful way to transfer tacit as well as explicit knowledge. Many studies have found that moving personnel is a very powerful way to facilitate knowledge transfer. For example, based on results from a study of technology transfer in the textile machinery sector, Rothwell (1978) concluded that the most effective way to transfer technology was to move people. As noted previously in our discussion of moving members, people are able to trans- fer tacit as well as explicit knowledge when they move from one context to another.
The richness of information transferred may explain the effectiveness of per- sonal meetings and conferences relative to correspondence, papers, and publica- tions in transferring technology (Daft & Lengel, 1984). In a study of the diffusion of computer simulation technology, Dutton and Starbuck (1979) found that face-to- face meetings and conferences were more effective in diffusing the technology than written media such as papers, proceedings, and correspondence. Face-to-face meet- ings and conferences provide opportunities to transfer a richer set of information, including some tacit knowledge, than written media. The richness afforded by face- to-face communication conferences was especially important early in the diffusion of technology, when common understandings were being developed.
Ancona’s and Caldwell’s (1992) study of the activities groups use to manage relationships with other groups provides insights into knowledge transfer processes. Based on interviews and questionnaires from managers of new product teams, the researchers identified strategies groups took vis-à-vis external constituencies. Ambassadorial activities, which involved protecting the team and garnering resources and support for it, initially had a positive effect on meeting deadlines and budgets, but the effect dissipated over time. Although activities aimed at coordinat- ing technical and design issues did not affect team performance initially, these activ- ities had a positive effect on innovation over the long run. Prolonged scouting activities that involved gathering information and scanning the environment were negatively associated with team performance—both initially and over the long run. These results suggest that task–coordinator activities, which are intimately con- nected to the organization’s workflow, may be more important over the long run than ambassadorial or scouting activities.
Future research is needed to understand more fully the conditions under which knowledge transfers and to determine the effectiveness of various transfer mecha- nisms. Are certain mechanisms more effective knowledge conduits than others? Does the effectiveness of a particular mechanism vary as a function of the type of knowledge being transferred or the stage of the transfer process? More generally, a greater understanding of the micro processes underlying the transfer of knowledge is needed.
A very interesting issue that would benefit from future research is whether the processes and outcomes of knowledge transfer differ from the processes and out- comes of learning from one’s own experience. Results of a study by Walsh and Anderson (2011) at the individual level suggest that the processes and outcomes of learning from instruction differ from those of learning from one’s own experience. Walsh and Anderson (2011) compared a condition where participants received instruction before task performance and feedback after performance to a condition where participants only received feedback after performance. Behavior was affected by experience only in the no-instruction condition; the behavior of those receiving instruction improved immediately to asymptotic levels. By contrast, learning rates estimated from neural data were affected by the individual’s experience and did not differ between the instruction and no-instruction conditions. These results suggest that instruction can immediately affect behavior but that certain neural responses must be learned through experience. Research on whether the processes and out- comes of knowledge transfer and learning from one’s own experience differ at the group and organizational levels is needed.
Source: Argote Linda (2013), Organizational Learning: Creating, Retaining and Transferring Knowledge, Springer; 2nd ed. 2013 edition.