Organizational Memory: Consequences of Where Knowledge Is Embedded

Although empirical work on organizational learning and memory has increased in recent years (Miner & Mezias, 1996), there is little empirical evidence about the consequences of where knowledge is embedded for aspects of organizational per- formance. We are just beginning to understand the effect of organizational memory on organizational outcomes. An empirical study that directly examined the effect of memory on dimensions of organizational performance found that organizational memory affected the new product development process by influencing both the interpretation of new information and the performance of new routines (Moorman & Miner, 1997). Results indicated that higher organizational memory levels enhanced the short-term financial performance of new products, while greater mem- ory dispersion increased both the performance and creativity of new products.

Results also indicated that high memory dispersion could detract from creativity under conditions of environmental turbulence.

In this section, I draw on related literature and qualitative work we have done in franchise organizations (Argote & Darr, 2000) to suggest how where knowledge is embedded affects its persistence and transfer in organizations. Research on knowl- edge transfer is briefly discussed here because the transfer of knowledge from one unit to another provides evidence of its persistence. The following sections suggest how embedding knowledge in individuals, in tools, in routines, and in transactive memory systems affects its persistence over time and its transfer to other organiza- tional units.

1. Knowledge Embedded in Individual Members

Individuals provide both a sensitive and a precarious way of storing, maintaining, and transferring knowledge. Individuals are capable of capturing subtle nuances that other repositories are not able to store as readily. For example, in a series of ingenuous experiments, Berry and Broadbent (1984, 1987) showed that although individuals improved their performance as they gained experience with a task, they were not able to articulate what strategies they had used or why their performance had improved. Thus, as they gained experience with the task, individuals acquired tacit knowledge that they were not able to articulate to others. Individuals were able, however, to transfer their tacit knowledge to another similar task. When participants in the experiments performed a second task, the performance of those with previous experience on a similar task was significantly better than that of participants without any previous experience. Thus, even though participants could not articulate why their performance improved, they were able to transfer the knowledge that enabled them to improve their performance to a similar task.

These results suggest that moving personnel is a very effective way to transfer knowledge in organizations because individuals can transfer their tacit knowledge to other tasks and contexts. Thus, by transferring personnel, one transfers the tacit knowledge that individuals carry with them. Most studies of technology transfer find that moving personnel is a powerful facilitator of knowledge transfer (e.g., see Galbraith, 1990; Rothwell, 1978). A benefit of personnel movement is that it allows individuals to transfer their tacit knowledge to new contexts.

An alternative way of transferring tacit knowledge is to convert it to explicit knowledge. Nonaka (1991) described a fascinating example of how an engineer apprenticed herself to a bread maker to acquire the bread maker’s tacit knowledge. Through a long period of observation of the bread maker, the engineer captured the bread maker’s tacit knowledge and converted it to explicit knowledge. This explicit knowledge served as the base for Matsushita’s bread-making machine.

Individuals are the most effective media for acquiring and storing tacit knowl- edge. They are also an effective media for transferring tacit knowledge. Individuals can apply their tacit knowledge to a new task or a new context without converting their tacit knowledge to explicit knowledge. Alternatively, through a lengthy period of observation and apprenticeship, others may be able to capture an expert’s tacit knowledge and convert it to explicit knowledge that others can access.

Without moving personnel or explicitly attempting to capture their knowledge, knowledge embedded in individuals will generally not transfer. Qualitative results from our study of fast-food franchises illustrate how knowledge embedded in indi- viduals generally does not transfer to new sites. In our study of fast-food franchises, we observed 14 innovations at the stores (Argote & Darr, 2000). Of the 14 innovations that occurred in the fast-food franchises, 6 were embedded in individuals. For exam- ple, knowledge about how to hand-toss pizza remained embedded in individual work- ers. Knowledge about how to prioritize pizzas so as to take advantage of cooking time differences across pizza types and sizes and, thus, make better use of the oven also remained embedded in a few individual order-takers. Of the six innovations that were embedded in individuals, only two transferred outside the store of origin. By contrast, both of the innovations embedded in technology transferred outside the store of origin and five of the six innovations embedded in routines transferred outside the store of origin. Thus, knowledge embedded in individuals does not transfer as readily outside the organization of origin as knowledge embedded in technology or routines does.

Several pitfalls are associated with relying on individuals as a knowledge repository for organizations. Knowledge embedded in individuals may decay or depreciate faster than knowledge embedded in social systems. The results of an interesting laboratory study that compared individual and group recall suggest that knowledge embedded in groups is more stable than knowledge embedded in indi- viduals (Weldon & Bellinger, 1997). The researchers found a tendency for groups to exhibit less forgetting than individuals. Further, the organization of group recall was more consistent over trials than the organization of individual recall. Thus, an important difference between collective and individual memories may be the rela- tive stability of collective memories. Knowledge embedded in the group or social system seems to be more stable than knowledge embedded in individuals, even when there is no turnover of those individuals.

Another downside of relying on individuals as a knowledge repository is that individuals may not be motivated to share their knowledge. The Engeström et al. (1990) example discussed earlier in which an individual hoarded knowledge and did not share it with others is an example of this phenomenon. Many studies have shown that individuals typically do not share information that they uniquely hold (e.g., see Stasser & Titus, 1985).

A third downside of relying on individuals as a knowledge repository for organi- zations is that individuals can leave and take their knowledge with them. Conditions under which individual turnover will be especially harmful for organizations were discussed earlier in this chapter. Organizations can use a variety of strategies for capturing individual knowledge. Embedding individual knowledge in organiza- tional structures and routines is a productive way to mitigate the effect of individual turnover (Rao & Argote, 2006). Similarly, organizations may try to capture the knowledge of individuals and embed it in technology such as information systems and knowledge networks (Moreland, 1999; Stewart, 1995a, 1995b).

A significant component of individual knowledge, however, such as tacit knowl- edge, may be less amenable to being embedded in organizational structures and technologies. For organizations with a large component of tacit knowledge, attempts to bond the individual to the organization may be more fruitful than attempts to embed the knowledge in structures and technologies. Starbuck (1992) described the strategies organizations such as law firms and consulting firms use to prevent individuals from leaving. In these organizations, where much of the knowledge is embedded in individuals, their turnover would be very harmful for the organiza- tion’s performance. Hence, contracts are written and incentives are developed to motivate key individuals to remain with the firm.

A fourth disadvantage of relying on individuals to transfer knowledge is that it is hard for individuals to reach a large number of people without some degradation in the communication. Thus, for large organizations where reliability is important, it will not be efficient or effective to rely on individuals as the primary means for transferring knowledge. Individuals can be used effectively to complement other repositories, but relying solely on individuals to transfer knowledge in these settings will not be effective.

2. Knowledge Embedded in Tools and the Tool–Tool Network

Technology is a very effective repository for retaining explicit knowledge. As noted in the discussion of organizational forgetting in Chap. 3, we observed the least depreciation of knowledge in technologically sophisticated organizations. While more research is needed to determine whether it is technological sophistication that drives these organizations’ ability to retain knowledge, the depreciation rates observed across a variety of settings are consistent with the hypothesis that embed- ding knowledge in technology is an effective way to mitigate its depreciation. Similarly, Smunt (1987) suggested that embedding knowledge in technology is an effective way to prevent organizational forgetting. While embedding knowledge in technology does not guarantee its persistence, it makes persistence more likely.

Embedding knowledge in technology is also an effective way of transferring knowledge to other sites. Two of the innovations we observed in our study of fast- food franchises were embedded in technology. The “cheese spreader” example dis- cussed earlier in this chapter is an example of one of these innovations. Both of the innovations embedded in technology transferred outside the store of origin. While the number of innovations embedded in technology was, of course, too small to permit firm conclusions, the results are suggestive of the effectiveness of technol- ogy as a medium for transferring knowledge.

The results of our study of knowledge transfer across shifts in a manufacturing facility also illustrate the effectiveness of technology as a mechanism for transfer- ring organizational knowledge. The technology-transfer literature provides further evidence that embedding knowledge in technology and transferring it to another site can result in substantial savings for the recipient organization.

Interestingly, transferring knowledge by embedding it in technology is often most successful when it is accompanied by transferring a few individuals as well (e.g., see Galbraith, 1990; Rothwell, 1978). The advantages of individuals as knowl- edge repositories complement those of technology. Individuals capture the tacit knowledge, the subtlety, and the understanding behind the technology. By contrast, technology provides consistency and reliability and reaches a large scale.

A cost of embedding knowledge in technology is that the knowledge may become obsolete yet be more resistant to change because it is embedded in “hard” form. Abernathy and Wayne’s (1974) analysis of Ford’s production of the Model T suggested that Ford’s investment in “hard” automation to produce the Model T made it more difficult for Ford to change to meet customer preferences and offer a more varied product line. This example illustrates potential disadvantages of embedding knowl- edge in technology: increased rigidity and resistance to change. Today’s technologies are generally more flexible than they were in the 1920s, so the downside potential of embedding knowledge in technology might now be somewhat less. Nonetheless, rigidity associated with embedding knowledge in hard form is an important potential cost of embedding knowledge in technology that should be considered.

Knowledge management systems have been advocated as a knowledge repository that captures knowledge from the past to inform future decisions. Evidence on the effectiveness of these systems, however, is mixed. Haas and Hansen (2005) studied a consulting firm that used a knowledge management system consisting of docu- ment libraries linked by a search engine. The researchers found that the perfor- mance of consulting teams was negatively associated with the number of documents they used from the system. Further, using documents from a knowledge manage- ment system was especially harmful for experienced teams and teams facing very competitive environments.

By contrast, Kim (2008) found that the effect of using a knowledge management system in a chain of retail grocery stores was generally positive. The impact of the knowledge management system on performance was more positive for managers with fewer alternative sources of knowledge, for managers in remote locations, and for managers of products that did not become obsolete quickly.

The difference in the effectiveness of these two knowledge management systems seems likely due to task differences between the two contexts. Arguably, the prob- lems encountered by managers in a retail grocery team were more similar and less novel than those faced by consulting teams. The document repository form of knowledge management system appears to be more effective in organizations facing recurring tasks that are similar than in those facing varied and novel tasks.

3. Knowledge Embedded in the Task–Task Network

Routines, elements of the task–task network, are effective mechanisms for storing and maintaining knowledge. For example, in our study of fast-food franchises, we saw an example of a very efficient and effective routine for placing pepperoni on pizza. As described previously, it was discovered that placing pepperoni in a pattern that resembled spokes on a wheel before deep-dish pizza was cooked resulted in an even distribution of pepperoni on the cooked pizza. The discovery was embodied in a routine that could be used easily by all pizza makers. Embedding the knowledge in a routine made it more resistant to employee turnover. If the individual who made the discovery of how to achieve an even distribution departed from the store, the knowledge would remain in the organization. Embedding knowledge in a routine enhances persistence.

Routines, elements of the task–task network, are also an effective mechanism for transferring knowledge to other organizations. The routine for placing pepperoni was discovered in a store in Southwestern Pennsylvania. The routine transferred very quickly to other stores in the same franchise. A consultant from the parent corporation who saw the routine on a visit to one of the stores was impressed by the routine’s effectiveness and diffused it widely to stores throughout the corporation. The routine is now used in nearly all of the stores of the parent corporation.

Indeed, other routines we observed in the fast-food franchise study also trans- ferred outside the store of origin. Of the 14 innovations we identified in our fast- food franchise study, 6 were embedded in routines. Of these six, five transferred outside the store of origin. And three of those transferred to stores in different fran- chises. Thus, embedding knowledge in a routine is an effective way to facilitate knowledge transfer.

The results of a study by Zander and Kogut (1995) are consistent with our quali- tative results regarding knowledge transfer. Zander and Kogut (1995) examined fac- tors affecting the speed of transfer of manufacturing capabilities. The researchers found that capabilities that could be codified (e.g., in documents or software) trans- ferred more readily than capabilities not easily codified. In order to be embedded in a routine, capabilities must be codified.

A downside of relying on routines is that they may be used inappropriately. Researchers have written about the importance of “unlearning” in organizations— forgetting the old and developing a better, more appropriate routine as a way of adapt- ing to changed circumstances (e.g., Hedberg, 1981). Unlearning is arguably an example of learning—of developing a more elaborate response repertoire that specifies the conditions under which various responses are appropriate. So rather than “forget” a routine used in the past, it would be preferable to remember the routine, the condi- tions under which it worked, and why it is no longer successful. Thus, lessons of the past can be applied to the present to facilitate organizational performance.

4. Knowledge Embedded in the Member–Task Network

Several studies have examined the persistence of knowledge in transactive memory systems. Research has shown that knowledge embedded in transactive memory systems persists more than knowledge embedded in individual members. In a series of studies, researchers compared the recall 1 week after the training of groups whose members were trained together to the recall of groups who members were trained apart (Liang et al., 1995; Moreland et al., 1996). Groups trained together recalled more 1 week later than groups whose members were trained apart. The greater recall of group-trained groups relative to individually trained groups could be explained by their more developed transactive memory system (Liang et al., 1995). Studies have examined how resilient knowledge embedded in transactive mem- ory systems is to member turnover. If turnover from one period to the next is com- plete such that no individual worked with someone they had previously, then the transactive memory formed from working together is no longer relevant and confers no benefits for task performance (Moreland et al., 1996). If turnover is not complete and a subset of members continue to work together, transactive memory systems can retain value (Lewis, Belliveau, Herndon, & Keller, 2007). The effectiveness of transactive memory systems depends on contingencies, such as the extent to which the skills of new members are similar to those of departed members.

Source: Argote Linda (2013), Organizational Learning: Creating, Retaining and Transferring Knowledge, Springer; 2nd ed. 2013 edition.

Leave a Reply

Your email address will not be published. Required fields are marked *