Towards An Organic perspective on strateg. The Mechanistic perspective, страница 22

Conduct implications

Variable resolution

A better understanding of strategy-related phenomena is unlikely to be gained by attending to a single theoretical perspective, level of analysis, context, or time frame. Thus for example, the factors associated with the success of a single strategic decision, the tenure of a specific executive team, or firm survival across several generations of technological breakthroughs can vary widely (see Zaheer, Albert, and Zaheer, 1999, on the issue of time scale). Furthermore, what may be optimal at a collective level may not be optimal at the unit level. Progress is more likely to be made by using research with different degrees of resolution. By employing both finegrained and coarse-grained approaches alternately, a more holistic appreciation of strategy issues can emerge.

Much progress has been made in the study of highly specific phenomena such as acquisitions and multipoint competition. Attention at the level of individuals and to micro phenomena can also make new and important advances. At the other end of the spectrum, strategy research can benefit from using multiple time frames, comparative (historical) research, simultaneous exploration of different levels of analysis, and multiple theoretical lenses. Clearly, such a research agenda is more demanding and therefore it may be better approached in research programs, in large, booklength studies, and in periodical reviews rather than in the usual single-study format. However, it is likely to better place theoretical ideas and empirical observations in a broader and more temporal context.

History and process research

The organic perspective highlights the historical dimension of strategy-related phenomena. As illustrated in Chandler’s (1962) research, the nature of historical perspective makes it more likely to be eclectic, integrative, and sensitive to time, interaction, context, and multiple levels of analysis. Case histories of firms and industries that were instrumental to the field’s early development are sometimes labeled ‘prescientific’ (e.g., Rumelt et al., 1994). However, a renewed interest in historical and clinical research is not a sign of regression but of the field’s maturity. The benefits of such an approach are too great to be ignored by strategy researchers. New historical research is likely to be different from earlier work since it can now build on the cumulative progress made in the field. First, it can use both qualitative and dynamic statistical modeling. Second, it can use a better-developed theoretical base to frame the analyses. Third, it can be more sensitive to reading history forward as opposed to retrospectively, thus providing a better appreciation of how firms and managers cope with uncertainty, multiple trajectories, lags, and dead ends. Fourth, it can examine the development of firms, industries, and strategies before they become full-blown entities and thus add more knowledge on their early emergence, variation, and selection (see, for example, Aldrich, 1999). A revival of ‘neo-historical’ research in strategy may thus benefit from the path-dependent intellectual evolution of the field itself.

The content and spirit of the organic perspective require the use of longitudinal research and of less accepted methods such as sequence modeling, ethnography, and case histories. Cross-sectional studies can be useful but they cannot remain the predominant mode of analysis (Bowen and Wiersema, 1999). Process models look at different issues than variance models and therefore potentially produce different observations. Although there are different opinions with regard to the need to integrate variance and process approaches (e.g., Langley, 1999), we certainly see the use of process models as appealing in several respects. First, by disaggregating time, they introduce unique possibilities for path and sequence to affect final outcomes. Second, process models may be better suited to gain insights into duration variables in general and into sustained performance in particular. Third, by their greater sensitivity to multiple trajectories, process models and studies are more likely to reveal sources of both success and failure.[17]