Valuing customer portfolios with endogenous mass-and-direct-marketing interventions using a stochastic dynamic programming decomposition

Abstract

Customer Relationship Management generally uses the value of customers to allocate marketing budget. But marketing interventions generally change the customer behavior, turning upside-down the customers ranking based on their initial valuations and making the budget allocation suboptimal. Rational Managers should allocate the marketing budget to maximize the expected net present value of future profits drawn from each customer, simultaneously planning mass marketing interventions and direct marketing effort on each individual. This is a large dimensional Stochastic Dynamic Program, which cannot be easily solved due to the curse of dimensionality. This paper propose a new decomposition algorithm to alleviate the curse of dimensionality in SDP problems, which allows forward-looking firms to allocate the marketing budget optimizing the CLV of their customer base, simultaneously using customized and mass marketing interventions