Empirical Bayes Estimation of Customers’ Guarantee Time Length of Loyalty

Abstract

In this research we use soft computing methods and apply an empirical Bayes method to estimate the minimum customer alive duration. It is an important topic because the information of minimum length provides marketing decision maker to know the largest lower bound of which the customer will be alive. In this paper, we call this minimum duration the guarantee length of loyalty which means the value of each individual customer alive duration will be larger than or equal to this minimum length. This estimate can be used to help finding the best marketing timing for the extended of customer alive time. The model under consideration is based on a Bayes framework which is very flexible (general) so that many complicated factors that involve in marketing problem can be included in this model. In this research an asymptotic optimal empirical Bayes estimate will be derived. As the result, this model will be more practical in real situation.