A Cluster Analytic Approach to Market Response Functions

The estimation of marketing policy effects with time series data is a fairly well-mined area of marketing discourse. Use of such data bases to examine the differences in effects on various market segments has received relatively less attention (except [3]). This is surprising given the premise of market segmentation, namely that certain groups of buyers behave differently than others. If a market consists of segments, ideally one would think that the effects of marketing policies should be estimated segment by segment. Then the market response function would be simply the aggregate of the functions estimated separately for each segment. That is, let: S, = f, (A, P, D)