Modelling customer retention with Rough Data Models

Banks are very interested in identifying some mechanisms behind decisions and determining clients that are about to leave the given bank, and one way of getting such an insight is to analyse historical data that describe customer behaviour in the past.

Abstract

Banks, as many other companies, try to develop a long-term relationship with their clients. When a client decides to move to another bank it usually implies some financial loses. Therefore, banks are very interested in identifying some mechanisms behind such decisions and determining clients that are about to leave the given bank. One way of getting such an insight is to analyse historical data that describe customer behaviour in the past. In this paper we present a methodology and some results of an analysis of a large data set provided by a big mutual fund investment company. Our approach, based on the concept of Rough Data Model, [7], resulted in the identification of key factors that influence customer retention. Moreover, a number of rules that characterise various groups of clients have been generated. Our results have been highly appreciated by the company and led to specific actions aimed at increasing customer retention.