Lewis Platt, the former CEO of Hewlett-Packard, once reportedly said that if Hewlett-Packard only knew what it knows, the company could dramatically improve its performance. That problem, as many executives recognize, is not unique to Hewlett-Packard. More and more companies are beginning to realize that the existence of knowledge—about, say, the best way to make widgets—in one area of a company need not mean that the company as a whole knows it. In other words, as Platt says, organizations do not necessarily know what they know.

Why does knowledge move slowly within organizations—a phenomenon that researchers describe as "stickiness?" A common explanation makes politics the culprit. If knowledge is power, those who hoard it may remain powerful. Another related factor may be competition. Those who know how to do a task exceptionally well may be unwilling to share their know-how for fear of losing their competitive advantage. Companies that accept these explanations usually go one step further. To "correct" the tendency toward knowledge hoarding, they offer incentives to employees to encourage them to share knowledge.

Such efforts, however well meaning, may be off the mark, according to research by Gabriel Szulanski, assistant professor of management at Wharton. In a study titled "The Process of Knowledge Transfer: A Diachronic Analysis of Stickiness," he examines the process of knowledge transfer—and it is a process, rather than an instant, technology-enabled, one-time act. Szulanski also develops a framework to explain the factors underlying stickiness, illustrated by in-depth studies of eight companies—AMP, AT&T Paradyne, British Petroleum (now BP Amoco), Burmah Castrol, Chevron, EDS, Kaiser Permanente and Rank Xerox.

Szulanski argues that stickiness is not an anomaly but fairly typical within organizations. Finding out and dealing with the factors that impede knowledge transfer can help free up the organization’s structure. For example, firms where friendly ties exist between a source of knowledge and its recipient are likely to promote exchanges. Such fertile organizations are different than so-called barren organizations, which only impede such activity.

Szulanski identifies four phases of the knowledge-transfer process, each of which causes stickiness in a different way.

    1. Initiation stickiness: This refers to the difficulty that people face in recognizing opportunities to transfer knowledge and act upon them. The search for opportunities and the decision to transfer knowledge occurs under a degree of uncertainty. This may happen because the source as well as recipient of knowledge may have incomplete understanding about the environment where new knowledge is to be used. Also, if the source of knowledge is seen as being untrustworthy, that can create a problem.
    2. Implementation stickiness: After the decision to transfer knowledge has been made, the attention shifts to the exchange of information between the source and the recipient. In this context, the problem of bridging the communication gap and improving coordination between the two is important. These difficulties can be reduced through planning and mutual understanding. Szulanski says that "The effectiveness of planning, coordination, and mutual adjustments is likely to depend on the quality of the relationship between the source and the recipient."
    3. Ramp-Up Stickiness: Once a recipient starts using newly acquired knowledge—for example, by changing a manufacturing process—problems might arise because the transferred knowledge gets an unexpected response. Difficulties could also occur because of an ineffective use of new knowledge.
    4. Integration Stickiness: The use of new knowledge and practices becomes gradually routinized and accepted in the organization. Difficulties, however, can lead to the abandonment of these practices and a reversal to the previous status quo.

By distinguishing stickiness at various levels, Szulanski presents an approach to the evolution of problems in knowledge transfer. He illustrates this model with an example—the transfer of best practices within a firm. Such a transfer refers to "the replication of a practice that is performed in a superior way somewhere within the organization." The study of the transfer of best practices is important since it lets organizations discover which structures support knowledge transfers. The resulting insight can help managers make decisions.

Szulanski’s research examines instances of stickiness and what causes it. One of his conclusions is that in the first three phases of knowledge transfer, motivation and the perceived reliability of the source are significant barriers. "Lack of motivation may result in foot dragging, passivity, feigned acceptance, hidden sabotage or outright rejection in the implementation and use of new knowledge," he says.

Another off-beat finding is that a motivated recipient can intensify rather than mitigate problems during the ramp-up stage by complicating the implementation of new practices. Barriers to the implementation of new practices can also arise as a result of the voluntary dismissal of outside help, the pride associated with ownership, or the making of unnecessary modifications.

Szulanski points out that the institutionalization of knowledge already available in the firm can sometimes pose a problem. The more institutionalized such knowledge is, the greater the effort required to dismantle it. Such complex responses lead Szulanski to conclude that there may be a so-called natural pace for knowledge transfers. A fast-paced transfer may prevent the full implementation of changes. On the other hand, a pace that is too slow may only retard change by institutionalizing practices.

The most important predictor of stickiness that Szulanski found in the implementation phase, however, is the ability of the recipient to absorb new knowledge. This suggests that the policy measures needed to overcome stickiness may be different than those suggested by conventional wisdom. If, as Szulanski shows, knowledge transfers are impeded by the inability of recipients to absorb knowledge, incentives may do little to promote knowledge flows. Companies may need instead to implement educational measures to increase the ability of knowledge recipients to learn. More and more companies should act on such prescriptions.

As more executives recognize that knowledge transfer is a process, they will be able to identify its different stages and use this model to identify and catalog transfer problems. According to Szulanski, this can help them "prevent stickiness as well as unstick sticky transfers." That, in turn, should enable companies to get closer to knowing what they know.