Shared knowledge, “glitches,” and product development performance

Much recent thought in strategy has stressed the importance of organizational integration for competitive advantage. Empirical studies of product development have supported this emphasis by correlating integrating practices and superior performance. We propose, from a resource‐based or capability view, that this correlation results from integration leading to patterns of shared knowledge among firm members, with the shared knowledge constituting a resource underlying product development capability. To explore this connection, we examine the product development efforts of a scientific software company. We define the ‘glitch’ as a costly error possible only because knowledge was not shared, and measure the influence of glitches on firm performance. At this company, gaps in shared knowledge did cause the company to incur significant excess costs. We also identify a set of ‘syndromes’ that can lead to glitches, and measure the relative importance of these syndromes. The glitch concept may offer a general tool for practical measurement of the marginal benefits of shared knowledge. Copyright © 1999 John Wiley & Sons, Ltd.     

Response to Farjoun's ‘Strategy making,novelty, and analogical reasoning — commentary on Gavetti,Levinthal, and Rivkin (2005)’

In his thoughtful commentary on our 2005 paper (Gavetti, Levinthal, and Rivkin, 2005), Farjoun offers three critiques and extensions. First, he suggests our approach should have explicitly considered a constructionist logic. Second, Farjoun argues that we have neglected the full array of modes of cognition between rational choice and feedback‐based adaptive learning and have therefore overstated the role of our focal mode, reasoning by analogy. Third, he highlights some of the contingencies under which the various modes of cognition he identifies are effective. In response, we address each point. We first argue that a constructionist perspective is not alien either to the role of analogical reasoning or to the particular modeling apparatus we have developed. We then suggest that despite the richness of modes of cognition that lie between rational choice and adaptive learning, theorizing about them requires simplification and the identification of underlying categories that classify such modes, which is the approach our paper employs. Finally, we clarify how our paper adopts the contingent logic advocated by Farjoun. Copyright © 2008 John Wiley & Sons, Ltd.