Levels of Knowledge Transfer in Organizations

Knowledge transfer in organizations has been studied at different levels of analysis. This section provides an overview of whether knowledge has been found to transfer at these different levels. Subsequent sections identify the conditions under which knowledge transfer is most likely to occur. Several researchers have examined whether knowledge transfer occurs across products or models of the same product within an organizational context (e.g., Udayajiri & Balakrishnan, 1993). Others have studied whether knowledge transfers between units of the same organization, such as shifts within a manufacturing plant (Epple, Argote, & Devadas, 1991) or product development and marketing departments (Adler, 1990). Others have examined whether knowledge transfers across organizations embedded in a superordinate rela- tionship, such as a franchise or chain (e.g., Baum & Ingram, 1998; Darr et al., 1995). Still others have examined whether knowledge transfers or “spills over” across com- petitors (e.g., Henderson & Cockburn, 1996; Irwin & Klenow, 1994). The extent of knowledge transfer found to occur at each of these levels is reviewed: across prod- ucts, across units of the same organization, across organizations embedded in a superordinate relationship, and across independent organizations.

Knowledge transfer from one model to another or from one product to another can contribute significantly to a firm’s performance. Ghemawat (1985) described a dramatic example of knowledge transfer across products in the motorcycle industry. British motorcycle manufacturers apparently did not understand that achieving a viable cost position on one product depended on their manufacture of other prod- ucts, because different products shared parts and manufacturing processes. British manufacturers stopped producing small motor bikes when they were threatened with competition from the Japanese on that end of the market. Although getting out of the small bike market improved their performance initially, it destroyed their performance in the long run. The small bikes shared many parts and processes with their larger counterparts. When the British stopped production of the small bikes, they dramatically reduced their experience base for learning about the parts and processes shared with the large bikes. According to Ghemawat (1985), the British share of their own home market fell from 34 % in 1968 to only 3 % in 1974.

The Lockheed case previously described in Chap. 3 also suggests the importance of the ability to transfer knowledge across products for firm performance and sur- vival. A competitor of Lockheed, McDonnell Douglas, produced a military KC-10 tanker that was very similar to its commercial DC-10 counterpart. The ability to transfer knowledge across the KC-10 and the DC-10 may have contributed to the more favorable cost position McDonnell Douglas enjoyed relative to Lockheed.

Several studies have empirically examined the extent of knowledge transfer across models or products. These studies have generally found that knowledge transfer occurs—but to an incomplete degree. In his detailed analysis of the Lockheed L-1011 production program, Benkard (2000) found that transfer of knowledge occurred from one model of the L-1011 to another. The transfer, however, was incomplete: not all of the knowledge acquired on the production of the first model transferred to the second. Lockheed began producing one model and produced it with minor variations throughout the entire production program. Several years into the production pro- gram, Lockheed began producing a different model with a shorter fuselage and dif- ferent cargo and galley configurations. Benkard (2000) found considerable—but incomplete—transfer from the first to the second model. Benkard estimated that the first unit of the second model required approximately 25 % more labor than another unit of the original model would have required.

Henderson and Cockburn (1996) also found considerable knowledge transfer across products. In their analysis of the determinants of research productivity in drug discovery, Henderson and Cockburn (1996) found evidence of internal spill- overs of knowledge between related research programs within the same firm.

Irwin and Klenow (1994) found weak evidence for learning spillovers across dif- ferent generations of products in the semiconductor industry. Irwin and Klenow (1994) did not have access to firm-level data on production cost but rather used price as a proxy for unit cost data. Using firm-level price data, Irwin and Klenow (1994) concluded that knowledge transferred or spilled over only in two of the seven product generations. By contrast, using industry-level price data, Udayajiri and Balakrishnan (1993) presented evidence that knowledge transferred over generations for the five semiconductor products they analyzed. Further, the researchers found that experi- ence producing Dynamic Random Access Memory chips benefited the production of other memory products. The difference in findings across these two studies of knowl- edge transfer in the semiconductor industry might be due to their differing focus on firm- versus industry-level data. Both studies used price as a proxy variable for unit costs. It is hard to disentangle the price effects that arise as a consequence of market dynamics from the effects that derive from knowledge transfer.

More generally, further research is needed to understand the conditions under which knowledge transfers across products or generations of products. The similar- ity of products and the extent to which they build on a common knowledge base are key factors conditioning the ease of transferring knowledge across them. The role of similarity will be addressed in a subsequent section of this chapter. The extent to which the individuals producing the product and the tools used are the same also are likely to affect the extent of knowledge transfer.

Researchers have also studied knowledge transfer across the components of an organization. For example, as described in Chap. 4, my colleagues and I studied the amount of knowledge transfer that occurred from the first shift to the second shift when the new shift was introduced at a truck assembly plant (Epple et al., 1996). We found that the transfer from the period of one-shift operation to the period of two- shift operation was rapid and almost complete. Within 2 weeks, the second shift achieved a level of productivity that it had taken the first shift almost 2 years to achieve. We suggested that the rapid and complete transfer was due to much of the organization’s knowledge being embedded in its technology (see Mishina, 1992).

A significant amount of knowledge has also been found to transfer across orga- nizations that are embedded in a superordinate relationship. My colleagues and I examined transfer of knowledge across 13 World War II shipyards that went into production at different points in time (Argote et al., 1990). Mechanisms for transfer- ring knowledge across the shipyards existed (Lane, 1951). The organizations pro- duced the same product with a standardized design. A central agency was responsible for purchasing, designing each yard’s layout, approving the technology it used, and supervising its construction. The central agency also stationed engineers, inspec- tors, and auditors at each site to share information about “best practices.” 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 production experience at other yards. Thus, transfer of knowl- edge occurred across the shipyards at the start of production but not thereafter.

The results of our franchise study described earlier in this chapter indicated that fast-food stores benefited from the experience of other stores in the same franchise (Darr et al., 1995). Similarly, Baum and Ingram (1998) found that Manhattan hotels benefited from the experience of other local hotels in the same chain. Ingram and Simons (2002) found that kibbutzim benefited from the experience of other kibbut- zim in the same federation but not from the experience of kibbutzim in different federations.

Several studies have examined whether knowledge transfers or spills over across firms in an industry. Although there is evidence of knowledge transfer from other firms in the industry, firms typically learn more from their own direct experience than from the experience of their competitors. For example, Zimmerman (1982) studied transfer of knowledge in nuclear reactors built between 1953 and 1963. The unit cost of construction was analyzed as a function of a firm’s direct experience constructing power plants and the cumulative experience in the industry construct- ing power plants. Both types of experience were found to be significant predictors of the unit cost of construction. The effect of firm-specific experience, however, was more significant than the effect of industry experience.

Joskow and Rose (1985) took a more refined approach to measuring experience in their study of 411 coal-burning steam-electric generating units built between 1960 and 1980. Joskow and Rose (1985) analyzed the unit cost of construction as a function of firm-specific experience, architect-engineer experience, and industry experience. An architect-engineer team typically designed plants for more than one firm. Joskow and Rose (1985) found that firm-specific experience and architect- engineer experience were significant predictors of the unit cost of construction, whereas industry experience was not significant. Thus, firms learned from their own direct experience and the experience of architect-engineer teams. Transfer of knowl- edge occurred across firms that employed the same architect-engineer team but not across other firms in the industry.

Irwin and Klenow (1994) also studied knowledge transfer across competitors. Although knowledge transfer was found to occur, firms learned three times as much from an additional unit of their own direct experience as from an additional unit of experience at another firm. Further, knowledge appeared to spill over as much between firms in different countries as between firms within a country. In contrast to popular notions about learning in Japanese firms, Japanese semiconductor firms were not found to differ from firms in other countries in their rate of learning.

Henderson and Cockburn (1996) also analyzed knowledge transfer across competitors. In their study of factors affecting research productivity in drug dis- covery, Henderson and Cockburn (1996) found that firms benefited from their own direct experience and from the experience of their competitors. Although the effect of the experience of competitors was considerably smaller than the effect of a firm’s own experience, the benefits of the experience of others accounted for a significant amount of the total variance in productivity. Thus, the Henderson and Cockburn results are similar to those of Irwin and Klenow as well as Zimmerman.

Firms generally learn from their own direct experience and from the experience of their competitors. Although they do not learn as much from their competitors as from their own experience, the transfer of knowledge from other firms contributes significantly to a focal firm’s productivity.

Levine and Prietula (2012) combined data from a consulting firm with an agent- based computational model to analyze the costs and benefits of knowledge transfer and found that the effect of knowledge transfer was contingent on various factors. Better organizational support for employee learning at the individual level as well as greater access to organizational memory weakened the benefits of knowledge transfer. Thus, knowledge transfer appeared to serve as a substitute for individual learning and organizational memory. Further, the benefits of knowledge transfer decreased in turbulent as compared to stable environments.

Thus, knowledge transfer has been found (to varying degrees) at all levels of analysis. An interesting issue that has received some attention and would benefit from more is the relationship between knowledge transfer at different levels. Does knowledge transfer at these different levels substitute for or complement each other? Related research is now discussed and the conditions under which knowledge is most likely to transfer across organizations are developed. The focus is primarily on knowledge transfer across organizations or units of organizations. Related work that helps illuminate these processes is included.

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 *