Mine data to model and 
predict customer behavior

Share this article:
John Keenan
John Keenan

Many companies use their customer databases, including transaction history, simply to predict future actions based on previous behaviors of the same type, such as when, what and how much customers will buy in the future. Yet, existing customer data can also be used to model attitudes, which allows a brand to tailor messaging and offers to individual needs. 


The type of attitudes that a business decides to model depends on the organization. Let's consider a recent program Anthem helped run for a provider of home services. The goal was to decrease attrition rates by supporting a call center's "save-desk" initiative. They wanted to understand not just the reason a customer might want to cancel their service, but the deeper motivation and how to arm their customer service representatives.


This provider generally has two types of customers. The "outsourcers" are those that prefer an outside provider to handle the services in their entirety. The "do-it-yourselfers" are those that prefer to do things themselves, but occasionally desire help from a provider, either to deal with a specific problem or when they are short of time. By predicting the likely type for each customer, we could respond appropriately if they called in to cancel. 


One of the things that we noticed through our exploratory data analysis was a sizable number of customers who, over a several year period, repeatedly enrolled in the service, cancelled after some time, and then re-enrolled at some future point. There were others that enrolled and subsequently cancelled, and who we never heard from again.


This suggested we could predict the likelihood that a customer who cancels will re-enroll at a future date. By inference, a customer of that type is likely a do-it-yourselfer. Now we had a way to categorize the customer base by attitudinal difference, through a creative use of existing transaction data. When a do-it-yourselfer called to cancel, the customer service representative could graciously thank them, and put them in the stream for a future win-back program. When an outsourcer called to cancel, the representative could probe deeper and try to save the customer, using appropriate price or service adjustments and retention incentives.


Depending on the business, it might be more important to identify other attitude factors, such as convenience or purchase triggers. If you think creatively, you might be able to find a way to use your existing customer data to do so.


John Keenan is managing partner at Anthem Marketing Solutions. Reach him at jkeenan@anthemedge.com.
Share this article:
You must be a registered member of Direct Marketing News to post a comment.

Sign up to our newsletters

Follow us on Twitter @dmnews

Latest Jobs:

Featured Listings

More in Data/Analytics

12 Big Data Facts for Marketers in 2014

12 Big Data Facts for Marketers in 2014

The idea of Big Data is nothing new, but its potential to solve today's problems and spark innovation is unprecedented.

Harvard Prof: Marketers Need to Step Up Their Predictive Abilities

Harvard Prof: Marketers Need to Step Up Their ...

Statistics expert Edo Airoldi says data must be paired with predictive analytics before marketers can truly forecast customer behavior.

The (Marketer's) TV Guide

The (Marketer's) TV Guide

Public broadcasting station WGBH in Boston cleans up its dirty data and boosts donations.