Abstract

In this paper, we use a partial least square (PLS) optimization method as a prediction model to estimate the churn probabilities of customers and as a control model after configuring optimization objective and constraints with relative management costs of controllable variables. In our experiment, we observe that while the training and test data sets are dramatically different in terms of churner distributions (50% vs. 1.8%), four controllable variables in three marketing strategies played a key role in optimization process. We also observe that the most significant variable for prediction does not necessarily play an important role in optimization model because of the highest management cost. In addition, we show that marketing managers even further maximize financial outcomes of marketing campaigns by selecting customers based on churn probability or management cost.

Share

COinS
 

Churn Management Optimization via Partial Least Square (PLS) Model with Controllable Marketing Instruments and Associated Management Costs

In this paper, we use a partial least square (PLS) optimization method as a prediction model to estimate the churn probabilities of customers and as a control model after configuring optimization objective and constraints with relative management costs of controllable variables. In our experiment, we observe that while the training and test data sets are dramatically different in terms of churner distributions (50% vs. 1.8%), four controllable variables in three marketing strategies played a key role in optimization process. We also observe that the most significant variable for prediction does not necessarily play an important role in optimization model because of the highest management cost. In addition, we show that marketing managers even further maximize financial outcomes of marketing campaigns by selecting customers based on churn probability or management cost.