Evolution toward fit - organizational development

Administrative Science Quarterly, March, 2002 by Nicolaj Siggelkow

Scholars in a variety of literatures have conceptualized firms as systems of highly interdependent elements (e.g., Miller, 1981; Milgrom and Roberts, 1990; Porter, 1996; Levinthal, 1997; Whittington et al., 1999). In these analyses, organizational elements such as firms' activities, policies, structural elements, and resources are seen to form complex systems, or configurations. Although scholars have established a long time ago that consistency, or internal fit, among an organization's elements is positively correlated with performance (e.g., Khandwalla, 1973; Drazin and Van de Ven, 1985), research about how organizations evolve toward such configurations of tightly reinforcing elements is much less developed (Miller, 1996). Moreover, no adequate vocabulary exists that would allow a consistent description of organizations' developmental paths toward configurations--a vocabulary that would be helpful in developing further theory that links organizational development and organizational performance.

In describing organizations, many scholars have argued that some organizational elements are more central or core to an organization than other, more peripheral elements (Hannan and Freeman, 1984; Singh, House, and Tucker, 1986). For instance, Hannan and Freeman (1984:156) identified as "core aspects of organization" the organization's stated goals, its forms of authority, its core technology, and its marketing strategy. While the distinction between core and noncore elements has become common in the organizational literature, little progress has been made to date in distinguishing them systematically. As a result, Hannan, Burton, and Baron (1996: 507), summarizing the research efforts concerning organizational core elements, concluded, "Although there seems to be a general agreement that some organizational features fall nearer the core than others, we see no consensus on exactly what constitutes the core."

To better understand the nature of core elements and the underlying developmental processes that lead to configurations, I engaged in an inductive study of the developmental path of The Vanguard Group, the second-largest mutual fund provider in the U.S. Through an iterative process, involving the data of the case, existing literature, and broader conceptual reasoning about organizational evolution, I identified four developmental processes that were intimately related to the creation and further elaboration of organizational elements that had core characteristics. To define these developmental processes more tightly, a characterization of organizational core elements was required that allowed a systematic identification of such elements. As a result, the inductive study of the developmental processes went hand-in-hand with creating a methodology that allowed the identification of an organization's core elements at various points in the organization's history. To identify core elements, the notion of interacti ons among the elements of an organizational system played a key role.

ORGANIZATIONAL CORE ELEMENTS AND ORGANIZATIONAL SYSTEMS

In previous work, core elements of organizations have been identified by declaring ex ante a number of critical activity domains and then selecting variables that represent these domains. For instance, Miller and Friesen (1982), in their study on quantum changes, measured structural variables falling into three categories: uncertainty reduction (e.g., use of formal rules), differentiation (e.g., decentralization), and integration (e.g., use of coordinative committees). In a similar vein, Romanelli and Tushman (1994:1147), in testing the punctuated equilibrium model, argued that organizational culture, strategy, structure, power distributions, and control systems are "important to organizational survival and central to organizational activities" and thus form core activity domains. Changes in these activity domains were measured by introductions of new product lines (strategy change), by general reorganizations of the firm (structural change), and by turnover of senior executives (power distribution change). D ue to data limitations, they dropped the culture and control systems domains from their analysis.

The advantage of an ex ante specification of core elements is that changes in these elements can be measured consistently across firms. The disadvantage of this approach is that it assumes that the same elements are equally central or core in all firms. Empirical evidence suggests, however, that the same elements are not equally central in all firms. For instance, Singh, House, and Tucker (1986) argued that chief executive change is a peripheral change in their

sample of voluntary social service organizations, while Romanelli and Tushman (1994) coded such a change as a core change for their sample of minicomputer producers. The ex ante specification approach also implies that the number of core elements is constant across different firms and constant over time for any given firm. Rather than specifying core elements (or domains from which core elements are selected) ex ante, it may be preferable to identify core elements more directly.