Retail promotional activities such as temporary price cuts, coupons and newspaper features produce multiple effects: increases in market share for the promoted brand, increased switching to the promoting store, and increases in category volume. Moreover, the promotional effects can be different for different segments of consumers. This research defines market segments as groups of consumers who are homogeneous in terms of their vector of choice probabilities for the brands in the product class. Segments are determined by an iterative Bayesian procedure. The variations in within-segment market shares within a store are first related to promotional variables using a Logit model estimated by nonlinear least squares. The effects on store choice are modeled in a Nested Logit framework. Finally, category volume is related to the overall promotional attractiveness of product category by incorporating both current and lagged effects. The research approach is applied to the HU ground coffee data. Results include: (i) the market can be characterized by brand loyal segments each of which mostly buys their favorite brand, and switching segments each of which switches mainly among different brands of the same type (e.g., Drip, Percolator), (ii) promotional variables have significant effects on within-segment market shares, with the effects being different across segments, (iii) store-share is significantly related to promotional attractiveness of a store, (iv) the overall promotional attractiveness of the product category has significant current and lagged effects on category volume, and (v) the lagged effects resulting from consumer purchase acceleration and stock-up last longer for brand loyal segments compared to switching segments.