Reflections on 'A Meta-analysis of Applications of Diffusion Models'
Coauthor(s): Fareena Sultan, John Farley.
Diffusion of really new products is of increasing concern to marketers because of the evident impact of innovation on firm profitability (Capon, Farley, and Hoenig 1995). A meta-analysis of a disparate set of related, but not identical, marketing studies can provide the kind of empirical generalizations that are both valued by practitioners (e.g., Marketing Science Institute, Research Priorities 1994-1996) and an appropriate goal of basic research. We were fortunate to see a link between the two and recommend both areas for further research. We hope that our reflections and the much- appreciated O'Dell award given to our article will help encourage the use of meta-analysis in marketing. At a minimum, the now-discredited null-hypothesis of zero values for diffusion-model parameters should be abandoned in favor of a null hypothesis that is based on existing results. There is a tendency to treat anything that looks new as having no precedent. Although each new product or service has certain unique characteristics and consumer reactions to them need to be tracked (e.g., Sultan and Winer 1995), we have much to learn from the past. Formal meta-analysis of previous diffusions and in-depth historical studies of previous analogies. (e.g., the telegraph) can provide useful insights into the adoption of revolutionary innovations, such as the Internet.
Source: Journal of Marketing Research
Sultan, Fareena, Johh Farley, and Donald Lehmann. "Reflections on 'A Meta-analysis of Applications of Diffusion Models'." Journal of Marketing Research 33, no. 2 (May 1996): 247-49.