Although highly relevant for marketing practice, few studies provide conceptual and empirical insights into customer portfolio management. Furthermore, most approaches to analyzing customer portfolios are static. This article discusses three neglected key issues relevant for a dynamic customer portfolio analysis: (1) Does a static versus a dynamic valuation lead to a different prioritization of customer segments in a portfolio? (2) How does offensive or defensive management of segment dynamics affect portfolio value? and (3) Do reliable predictors for dynamics of a customer's position in the portfolio exist? As a tool for customer portfolio analysis, the authors develop a segmentbased customer-lifetime-value model. They capture customer dynamics by analyzing how customers switch between segments of different values across time. The authors apply their tool with longitudinal data from four firms with up to 300,000 customers. The results from the empirical analysis and a simulation study provide answers to the three key issues raised. First, compared with a dynamic analysis, a static approach overestimates the value of some customer segments but underestimates others. Second, a defensive versus offensive management of value dynamics is relatively more appropriate for middle-tier segments, whereas the opposite holds true for bottom-tier segments. Third, general customer characteristics and aggregated transaction characteristics indicate future segment dynamics, whereas specific product usage data differentiate customers according to current value.

No abstract