Valuing Customer Portfolios with Endogenous Mass and Direct Marketing Interventions Using a Stochastic Dynamic Programming Decomposition

The customer relationship management allocation in marketing budgets is potentially misleading when it uses individual customer lifetime value estimations from historical data. Planned marketing interventions would change the purchasing behavior of different customers, and history-based decisions would thus be suboptimal. To cope with this inherent endogeneity, we model the optimal allocation of the marketing mix by accounting simultaneously for mass interventions and direct marketing interventions for each customer. This is a large stochastic dynamic problem that, in general, is computationally rather intractable as a result of the "curse of dimensionality." We present an algorithm to derive the optimal marketing policies how the firm should allocate its marketing resources and the expected present value of those decisions, which maximize the long-term profitability of firms. This allows the firm to value customers/segments and helps the firm to target those that maximize long-term profitability given the optimal marketing resources allocation. We apply the proposed approach in the context of a kitchen appliance manufacturer. The results identify the most effective marketing policies and the endogenous customer values. It is in this context that we also dynamically identify the most profitable customer and the short-and long-term effects of marketing activities on each customer.