Chapter 1 Introduction
In a world in which the access to information is almost free or insignificant and where there is a real plurality of offers, churn analysis has become one of the key points a firm needs to focus on. Whoever says plurality of offers needs to introduce the term competition. Thereby, the latter is more and more fierce and cut-throat. Furthermore, switching costs have decreased significantly thanks to market regulation laws. For instance in France when you switch TSP, the new provider pays you off cancellation fees. All of this being said, it is essential for firms to implement efficient strategies to enhance customer relationships. To that end, the development of both survival models and machine learning algorithms have enabled companies to really push-up their strategies in terms of customer portfolio management, monitoring of attrition and estimation of customer value.
After careful consideration of the issues at stake, the following key steps are focused on:
- Segmentation of customer portfolio as firms generally tend to partition their portfolio into multiple segments.
- Estimation of customer lifetime and prediction of attrition.
- Measurement of customer value.
In the following sections, the concepts of portfolio, attrition and customer value are defined. Then, some pieces of literature review are provided. Before embarking on data analysis and modelling, we present the theoretical basis of the models used in the study. We finally introduce the dataset and implement the methodology with the aim of estimating the overall value related to a fictional customer portfolio of a telecommunications service provider.