Portals eNewsletters Web Seminars dataWarehouse.com DM Review Magazine
DM Review | Covering Business Intelligence, Integration & Analytics
   Covering Business Intelligence, Integration & Analytics Advanced Search

View all Portals

Scheduled Events
Archived Events

White Paper Library

View Job Listings
Post a job


DM Review Home
Current Magazine Issue
Magazine Archives
Online Columnists
Ask the Experts
Industry News
Search DM Review

Buyer's Guide
Industry Events Calendar
Monthly Product Guides
Software Demo Lab
Vendor Listings

About Us
Press Releases
Advertising/Media Kit
Magazine Subscriptions
Editorial Calendar
Contact Us
Customer Service

Data Mining and Modeling:
Modeling to Reduce Attrition

online columnist David S. Coppock     Column published in DMReview.com
April 12, 2002
  By David S. Coppock

A critical issue for many businesses is customer turnover. The cost of customer acquisition is an investment that is paid back over time only if customers remain loyal. Much has been said about the strategic benefits of shifting the focus of marketing from customer acquisition to engendering customer loyalty. Since most people are familiar with these issues, I am not going to repeat them here. Instead, I will explain the role that data mining and modeling play in the effort to reduce customer attrition.

Attrition models (also called loyalty or retention models) are simply a statistical prediction of which customers are likely to remain loyal and which are likely to leave. They are estimated by using historical customer data that includes 1) observable predictive variables for each customer as of some specified point in time, and 2) whether or not each customer remained loyal (and for how long) after that point in time. By applying the model to current customers, a prediction of future loyalty is obtained. As with all models, this prediction is a matter of degree. It is not possible to guarantee loyalty or attrition for specific customers. But customers can be separated into high-risk and low-risk groups.

Retention Program Targeting

How are attrition models used to improve marketing and business results? The most common business application of attrition modeling is to target retention programs toward customers who are at high risk for attrition. (See Figure 1.)

Figure 1: Attrition Model

The cost of retention programs is usually incurred for everyone in the program, whether they would have remained loyal or not. Incremental revenue from the retention program comes only from converting a customer who would have left into a customer who stays loyal. Therefore, it will not be profitable to target retention programs to groups of customers who mostly would have been loyal anyway. (You can't "save" a customer who was loyal to being with.) The profitability will be increased if the program can be targeted toward people with a high likelihood of attrition.

This strategy can be particularly effective if customer "triggers" can be identified that identify a point in time when a customer becomes most vulnerable to attrition. For example, customers now expect an integrated experience between their sales, marketing and service touches with a firm. When these functions do not connect and customers have a bad service experience, they often respond by taking their business elsewhere. By monitoring for bad service experiences many of these customers can be saved by taking corrective action (such as an apology and a rebate).

Note however, that even if you can find customers who have a high probability of attrition, you still have to be able to influence this decision. Retention programs that don't substantially change attrition behavior will not pay off. In effect, retention programs have a double hurdle to clear before they attain profitability: 1) high risk customers must be identified and 2) for a reasonable cost, the program must have a significant influence on the loyalty decision.

Acquisition Targeting

An alternate means to attack the problem of high attrition is by targeting acquisition efforts. This strategy tries to find customers who will value your products and services as they are, rather than trying to change your offerings to please your current customers. Acquiring customers who are known to have a high average loyalty and eliminating acquisition efforts toward customers who are likely to quickly leave will result in a customer base that is loyal and profitable.

Incorporating acquisition targeting into the marketing mix in a profitable manner is much more straightforward than with retention marketing. Acquisition targeting is a decision to reduce marketing efforts (and therefore reduce cost) against unprofitable prospects. Retention marketing requires an increase in marketing expenses in an effort to turn unprofitable customers into profitable customers. Thus, retention marketing requires "cooperation" (that is, a change in behavior) on the part of customers. Acquisition targeting does not require customers to change - it is under the control of the marketer.

Acquisition targeting should actually be based on expected profitability, not just expected loyalty. But expected loyalty is an important determinant of profitability that is sometimes overlooked. When a prediction of loyalty is available, it can be combined with model scores for acquisition probability and profit per period to provide a targeting methodology based upon a complete calculation of expected profit.

Of course, this all assumes that there are a substantial number of prospects that value your current product and service offerings. Clearly the core of a business strategy must provide valued and differentiated products and services, consumer friendly channels, competitive prices, etc. The point is: don't expect acquisition targeting to solve all your problems. But if one of your problems is that you are acquiring the wrong customers, this can be fixed.


For more information on related topics visit the following related portals...
Data Analysis.

David Coppock has more than 20 years of experience in technical and strategic marketing positions. As senior vice president of Data Mining and Modeling at ANALYTICi he has implemented leading edge methodologies for targeting, segmentation and marketing strategy applications. He also led the Analytic Database Marketing Division at AT&T.; He holds a Ph.D. in economics from Yale University and can be contacted at dcoppock@patmedia.net.

Solutions Marketplace
Provided by IndustryBrains

Autotask: The IT Business Solution
Run your tech support, IT projects and more with our web-based business management. Optimizes resources and tracks billable project and service work. Get a demo via the web, then try it free with sample data. Click here for your FREE WHITE PAPER!

See Enterprise Business Intelligence in Action
See how business intelligence can be used to solve real business problems with this live demo from Information Builders

OutlookSoft Business Intelligence & BPM Software
OutlookSoft is real-time, Microsoft-based business intelligence and BPM software that unifies query, reporting, analysis & OLAP with planning, budgeting, forecasting, consolidation, reporting & scorecarding. Free demo & white paper

File Replication and Web Publishing - RepliWeb
Cross-platform peer-to-peer file replication, content synchronization and one-to-many file distribution solutions enabling content delivery. Replace site server publishing.

Design Databases with ER/Studio: Free Trial
ER/Studio delivers next-generation data modeling. Multiple, distinct physical models based on a single logical model give you the tools you need to manage complex database environments and critical metadata in an intuitive user interface.

Click here to advertise in this space

E-mail This Column E-Mail This Column
Printer Friendly Version Printer-Friendly Version
Related Content Related Content
Request Reprints Request Reprints
Site Map Terms of Use Privacy Policy
SourceMedia (c) 2005 DM Review and SourceMedia, Inc. All rights reserved.
Use, duplication, or sale of this service, or data contained herein, is strictly prohibited.