Innovators and imitators: Organizational reference groups and adoption of organizational routines

Abstract

Firms vary greatly in their rates of creating and adopting technological and organizational innovations, in part because of their choice of reference group. We argue that the selection of a reference group is a crucial and neglected source of firm heterogeneity. Comparisons to the average of other firms in a population cause most firms to adopt innovations once they are widely accepted. A distinctive feature of the minority of innovating firms that create innovations or adopt them early is that they compare themselves with, compete with, and try to differ from other innovating firms, whereas the majority of firms compare themselves with, and conform to, a broader group of firms. We elaborate this argument based on the behavioral theory of the firm and institutional theory, and test it on a two-period survey on adoptions of innovative organizational forms in Europe and the US. The analysis shows the predicted differences in the adoption patterns of innovating and imitating firms.

Introduction

Firms differ in their strategies of adoption and adaptation of technological and organizational innovations, with some firms being pioneers (innovators) and others being late adopters (imitators). The difference between innovators and imitators has been central to theories of creation and adoption of innovations. Economic theory maintains that greater preference for risk or greater ability to innovate generates innovations and higher performance (Schumpeter, 1942), which in turn compels other firms to imitate (Nelson, 1991, Nelson, 1995). Diffusion theory suggests that differences in risk propensity, and capability to generate and adopt innovations determine the order of adoption in a population of otherwise similar actors (e.g., Rogers, 1983, Karshenas and Stoneman, 1995, Lissoni and Metcalfe, 1994). Institutional theory argues that imitation of organizational structures and procedures is driven by norms of rationality (Meyer and Rowan, 1977) and uncertainty (DiMaggio and Powell, 1983), and that differences in social proximity to the innovators explain the order of adoption (Strang and Soule, 1998).

Mimetic behaviour – organizations imitating other organizations or comparison groups – is an established phenomenon supported by an extensive theoretical and empirical literature (Rogers, 1983; Strang and Soule, 1988). Mimetic behaviour has been shown in connection with firm strategies (Fligstein, 1991, Haveman, 1993, Haunschild and Miner, 1997), organizational structure (Fligstein, 1985, Burns and Wholey, 1993), and organizational processes (Sutton and Dobbin, 1996, Massini et al., 2002). Imitation has the effect of decreasing heterogeneity of organizational strategies and practices in an organizational field (DiMaggio and Powell, 1983) and offering legitimacy to organizations that adopt the modal behaviours (Scott, 1987). Imitation of prevalent behaviours as a way of resolving uncertainty and gaining legitimacy raises the question of how organizations come to adopt unusual and innovative practices. Two proposed solutions are that the early adopters have a greater need for the practice to solve political or economic problems (Tolbert and Zucker, 1983, Kraatz and Zajac, 1996) or are embedded in local networks with norms that diverge from those of the organizational field as a whole (Davis and Greve, 1997). Similarly, the technological innovation literature has investigated the characteristics of innovating firms (Pavitt, 1984) and their motivations to innovate (Tether, 2003) without settling on an explanation of how they differ from later adopters.

We propose a general explanation of why actors observing the same technological or competitive environment and having the same basic preferences for risk differ in their organizational routines for initiating and implementing change (Cyert and March, 1963, Nelson and Winter, 1982). The behavioral theory of the firm (Cyert and March, 1963) and its extension to evolutionary economics (Nelson and Winter, 1982) explain change as consequence of feedback on performance or fitness of heterogeneous firms which are characterized by specific routines and meta-routines (Nelson and Winter, 1982, Aldrich, 1999). In this view, managers make technological, strategic, or organizational changes when organizational performance falls short of an aspiration level determined by past experience and/or comparison with a reference group (Cyert and March, 1963, Bromiley, 1991, Greve, 1998). For example, comparison with the average profitability in the industry has been shown to drive strategic change and risk taking (Lant et al., 1992, Fiegenbaum and Thomas, 1995, Greve, 2003). Extending the theory to cover comparison on institutional, not economic, criteria, we propose that the relevant performance measure and determinant of the aspiration level is the prevalence of a new routine in a reference group of firms. In other words, adoption is influenced by whether the organization has the structures and procedures thought to be appropriate and legitimate in its chosen reference group of other firms (Meyer and Rowan, 1977, DiMaggio and Powell, 1983). This proposal implies a closer integration of institutional theory and the behavioral theory of the firm than in previous work. Whereas it is conventional to assume that adoption of new institutions is presented to boundedly rational managers as a solution to various organizational problems, we propose that managers also have an implied goal of incorporating institutions viewed as modern, and thus that the problem solved by adopting a new institution is that the firms in the organizational reference group have adopted the institution, but the focal organization has not.

Once adoption is seen as resulting from the comparison with a reference group of other firms that possibly varies across firms it is important to consider how the reference group is determined. Traditional formulations assume that aspiration levels are determined by comparison to the average of a population (Greve, 2003). This would suggest that most firms are more likely to adopt a technological innovation or new organizational practices once they are sufficiently widespread to become institutionalized. This formulation has the same problem as seen in other theories of imitation of failing to explain the behaviors of early adopters that go against the actions of the majority. The solution is to propose that innovating and imitating firms have different reference groups (Lewin and Massini, 2003). Just as social comparison theory proposes heterogeneity in the reference groups of individuals (Festinger, 1954, Wood, 1989), we propose that innovating firms are more likely to select other innovating firms as their reference group and imitating firms are more likely to select the average firm in the population, or in the sub-population of non-innovators, as their reference group.

These different choices of reference groups occur when innovativeness becomes an important component of firm identities (Dutton and Dukerich, 1991). A history of innovating or adopting new routines or emerging innovations early will cause managers to see their firms as innovators and compare themselves to other firms with a history of early adoption or sustained innovations. Innovating firms, therefore, are more likely to react to actions of other innovating firms than to changes in the population average, and thus are likely to stay innovative (Webb and Pettigrew, 1999). The innovator identity also implies a different reaction to the reference group than the one seen in the majority of firms. Imitating firms primarily change by moving toward the average of their reference group, thereby decreasing heterogeneity. For innovating firms, the average of their reference group of other innovating firms serves as an anchor for choosing distinctive actions, thereby generating greater variety of change relative to their reference group as well as the population average. This explanation advances a more nuanced theory of action than the institutionally anchored social mimetic processes or theories of herd behavior. By distinguishing the reference groups and behavioral strategies of firms with and without an innovator identity it accepts the institutional argument that practices become institutionalized, but adds a mechanism for generating innovations that prime the subsequent process of mimetic adoption.

Our empirical investigation examines whether groups of innovator and imitator firms, identified from their current level of adopting innovations, demonstrate differences in adoption behavior consistent with different selection of reference groups. We test the model using data on the adoption of new organizational routines thought to increase the structural and procedural flexibility of organizations. The data come from a large-scale survey of changes in organizational routines in large European and US firms in the 1990s (Whittington et al., 1999, Massini et al., 2002, Pettigrew and Massini, 2003). We expect the adoption pattern of the majority of firms in the population (imitating firms) to be explained by comparisons to the average of the population. We expect the adoption pattern of a small number of innovators and early adopters (innovating firms) to be explained by comparisons to the average of the top quartile of population. Moreover, we expect the variance around the adoption pattern of innovating firms to be greater than for imitating firms because innovating firms are more likely to differentiate themselves through unique patterns of adoption.