Chapter Listing | Return to 1997 Topical Index


While the current results are consistent with those of other authors, thus validating the use of content analysis and the use of dichotomous variables, all authors need to heed the call for industry specific analyses Further, while the results may be similar to other research studies, they can be only similar in that there is a finite set of possible strategies. Data are collected over time and thus researchers should expect results to change as environmental conditions change. Even the current results will change when future years and additional firms are added to the data base.

There are inconsistencies in the literature in terms of what has been learned about firms with different growth rates (for example, Siegel et. al, 1993). The differences are likely the result of analyzing similar firms at different times, meaning that both the competitive environment and the firms have changed and thus have the firms' strategies. While strategies are a complex set of decisions which should be consistent and appropriate, the differences in results likely point to strategies being not as consistent or as appropriate as believed.

Formal planning, has been thought to be related to firm performance. Indeed a number of studies have observed this relationship (Olson and Bokor, 1995). However, the result of this study indicates a negative relationship between formal planning and performance. Perhaps formal planning is a reactive strategy to prior poor performance and while gain may be related negatively in the short term, in the long term it should become positive.

Finally, there is no reason to assume that strategies stay constant over a period of years or that strategies for similar firms will be constant even over a period of one year. What is surprising is that survey studies produce relatively consistent results. Some researchers have observed firm behavior which was counter intuitive (Dean and Snell, 1996) and have difficulty explaining it. Perhaps the reason is that the results are correct for the firms and times studied. Recent work supporting the inadequacy of Porter's (1980) and Miles and Snow's (1978) work regarding strategic typologies for new venture firms (Carter et. al, 1994) becomes problematic if it is argued that typologies can be general in nature, but the specifics of each depends on a particular point in time when surveys are completed.


The results of the contingency table analysis of the three sets of firms demonstrated that there were different strategies for firms in different industries, thus supporting the first hypothesis. The results of the cluster analysis indicated that most firms' strategies changed dramatically from year to year, thus not supporting the second hypothesis. The discriminant analyses correctly classified the vast majority of high and low gain observations in each of the three SIC Codes: 36, 38, and 73, 91.5%, 78.2%, and 79.7% respectively. In SIC code 36, the result of a multiple regression analysis for all observations was that cost leadership, high price, innovation, joint venture/licensing, market development and product development accounted for 18.6% of the variance in percent gain. For SIC code 73, 24.2% of the variance in gain was accounted for by cost leadership, differentiation, formal planning, joint venture/licensing, quality and uniqueness. Results of the regression for SIC 38 firms explained less of the variance, 14.0%. Only four strategy variables accounted for the variance in gain: differentiation, formal planning, innovation, and market focus. Thus for each SIC code faster growth firms evidenced distinctly different strategies than firms with slower growth, thus supporting the third hypothesis..

The identified strategy variables typically are considered to define strategic decisions of firms. Thus it was expected that firms would be following those decisions over relatively long periods of time. The results of the analysis indicated something different. Strategies were being used in a much more tactical sense. They were being changed from year to year. For these firms, what was normally masquerading as a strategy was really a tactic. The analyses further indicated that industry specificity was lost even when data from firms which were technical in nature within the same industry, were agglomerated and analyzed. Industry specific studies provide more insight into strategy and performance then do multi-industry studies. It also appeared that the choice of only a small number of "tactics" provided a firm with higher gains in revenue. The wrong choice of "tactics" produced lower gains. Thus there are industry specific "tactics" which lead to differences in performance.. Some strategies had little influence on growth.


With 317 observations and 62 firms in the current data base, other longitudinal studies are being performed. In particular, different algorithms are being developed to relate strategy to performance over time. One ongoing study involves market value and the relationship to the strategy variables. A second study covers "marketing at the entrepreneurship interface."

Top of page | Chapter Listing | Return to 1997 Topical Index

1997 Babson College All Rights Reserved
Last Updated 03/03/98