IMPLICATIONS OF CHAOS RESEARCH FOR NEW PRODUCT FORECASTING

The mathematics of chaotic dynamics are now familiar to Product foreca sters and marketing researchers. Do possible sightings of chaos in new product data sets have implications for the way new product studies a nd launches should be performed? Should these practices be affected by the knowledge that chaos is possible in principle? Although the mathe matics of new product diffusion models clearly allow for chaotic bifur cations and fluctuations, these phenomena have not been reliably obser ved for actual products. In this article we offer reasons why this has been so. The reasons include measurement and specification error, and aggregation and data collection interval effects. We conclude that ma rketers have not been looking in the right places to find chaos (or al least traditional market research reports do not lend themselves to a n effective search for chaos), and that brand managers behave in a way that minimizes chances of observing chaos. The exploration of chaos i n the context of new product management leads to an analysis of the im plications of chaos for the practice of new product forecasting. (C) 1 996 Elsevier Science Inc.