Targeting Customers with Statistical and Data-Mining Techniques

Abstract

Operationalizing a relationship management program requires a retention strategy that is sensitive to an individual customer’s position in the service life cycle, while being financially sound for the provider. To this end, estimating a customer’s hazard function and remaining tenure with the company can lead to important insights into marketing tactics and constitute fundamental building blocks for methods of targeting important customers. The authors describe a way of estimating these quantities using a combination of statistical and data-mining techniques. The resulting customer hazard information leads to a generalization of lifetime value (GLTV) that explicitly accounts for company actions and their success in relationship management.