DSpace@MIT

This dissertation investigates how rapid learning occurs in high technology industries, many of which operate in what Bourgeois and Eisenhardt (1988) term "high velocity environments." The dissertation consists of three empirical studies, which follow the method of extended case study research (e.g. Yin, 1981; Eisenhardt, 1989a). Semiconductor manufacturing and process development are chosen as settings for this dissertation because they exemplify some of the attributes of high velocity environments. The first study looks into how firms organize for learning. Existing definitions of modularity (McClelland & Rumelhart, 1995; Ulrich, 1995b; Baldwin & Clark, 1997) are expanded to provide a theoretical framework for differentiation and integration of organizations (Lawrence & Lorsch, 1967, 1967a, 1969), technology (Iansiti & West, 1997; Iansiti, 1998), learning activity, accumulated knowledge and performance metrics in high velocity environments. The results of the study imply that differentiation in high velocity environments occurs with a high degree of modularity (the individual modules exhibit little interdependence), which fosters learning efficiency. The results also indicate that integrating technological subsystems in is significantly more complicated than integrating organizational subsystems. The second study explores the inner mechanisms of rapid learning by building a model of the lifecycle of a semiconductor manufacturing process. The output of the model suggests that the performance of a system is proportional to the performance of its most limiting subsystem. Learning in an organization occurs by paying attention to the weakest mechanism within that organization until that mechanism is no longer the weakest one, and shifting attention to the mechanism that replaces it as the weakest one.(cont.) Learning in high velocity environments is likely to be highly punctuated and to include substantial engineering efforts, which may occur prior to the release of a product. The third study investigates the effects of rapid learning on a firm's profitability, suggesting that the ability to conduct and accelerate punctuated learning serves as the primary source of competitive advantage in high velocity environments. In conjunction, these studies lay the foundation for a normative, metrics-driven, pragmatic theory of learning. The theory recommends that firms should 1) Define the global objective of a venture and select the global metrics that best measure whether the venture is making progress towards its stated objective; 2) Build a learning architecture that supports the global objective of the venture, i.e. define a hierarchy the metrics that allow the firm to achieve the objective in the most effective manner; 3) Organize for learning by defining a hierarchy of activities that are aligned with the metrics of the learning architecture; 4) Manage bottlenecks; and 5) Prepare for change before change occurs, by spreading knowledge as rapidly as possible throughout the organization.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Sloan School of Management, 2003.Includes bibliographical references (p. 524-569).