A datamining approach for retailing bank customer attrition analysis

Applied Intelligence 22, 47–60, 2005 c 2005 Springer Science + Business Media, Inc. Manufactured in The Netherlands.

A Data Mining Approach for Retailing Bank Customer Attrition Analysis
XIAOHUA HU College of Information Science, Drexel University, Philadelphia, PA, USA 19104
xiaohua hu@acm.org; thu@cis.drexel.edu

Abstract. Deregulation within the financial service industries and thewidespread acceptance of new technologies is increasing competition in the finance marketplace. Central to the business strategy of every financial service company is the ability to retain existing customers and reach new prospective customers. Data mining is adopted to play an important role in these efforts. In this paper, we present a data mining approach for analyzing retailing bank customerattrition. We discuss the challenging issues such as highly skewed data, time series data unrolling, leaker field detection etc, and the procedure of a data mining project for the attrition analysis for retailing bank customers. We use lift as a proper measure for attrition analysis and compare the lift of data mining models of decision tree, boosted na¨ve Bayesian network, selective Bayesian network,neural network and the ensemble of classifiers of ı the above methods. Some interesting findings are reported. Our research work demonstrates the effectiveness and efficiency of data mining in attrition analysis for retailing bank. Keywords: data mining, classification method, attrition analysis 1. Introduction

Deregulation within the financial service industries and the widespread acceptance of newtechnologies is increasing competition in the finance marketplace. Central to the business strategy of every financial service company is the ability to retain existing customer and reach new prospective customers. Data mining is adopted to play an important role in these efforts. Data mining is an iterative process that combines business knowledge, machine learning methods and tools and large amountsof accurate and relevant information to enable the discovery of non-intuitive insights hidden in the organization’s corporate data. This information can refine existing processes, uncover trends and help formulate policies regarding the company’s relation to its customers and employees [1]. In the financial area, data mining has been applied successfully in determining: • Who are the likelyattriters in the next two months? • Who are likely to be your profitable customers? • What is your profitable customers’ economic behavior?

• What products are different segments likely to buy? • What value propositions service different groups? • What attributes characterize your different segments and how does each play in the person’s profile? In this paper, our focus is on applying data miningtechniques to help retailing banks for the attrition analysis. The goal of attrition analysis is to identify a group of customers who have a high probability to attrite, and then the company can conduct marketing campaigns to change the behavior in the desired direction (change their behavior, reduce the attrition rate). In data mining based direct marketing campaign, it is well understood that targetingevery customer is unprofitable and ineffective. With limited marketing budget and staff, data mining models are used to rank the customers and only certain percentage of customers are contacted via mail, phone etc. If the data mining model is good enough and target criteria are well defined, the company can contact a much small group of people with a high concentration of potential attriters [1–5].The process of data mining for bank attrition analysis can be described in the following steps:

48

Hu losses. Revenue is constantly challenged by a high attrition rate: every month, the call centers receive over 4500 calls from customers wishing to close their accounts. This, in addition to approximately 1,200 writeins, “slow” attriters (no balance shown over 12 consecutive months) and...