Using Real-Time Behavior Data.

• Author: Kish, John

Imagine a technology that lets a bank capture and quickly react to important changes in a customer's behavior--changes like an unexpectedly large withdrawal combined with a deviation in ATM usage and a lapse in automatic paycheck deposit. And, armed with such information--suggesting a potential defection--what if you could intervene in time to keep this customer from closing his account? Now you can. Technology exists today that makes truly personalized, timely and meaningful customer dialogues like this a reality. But before we explain how this technology works, let's review some bank marketing basics.

The right message at the right time

The elements of a sound marketing strategy are easily summarized: You must deliver the right message to the right customer at the right time and through the right channel. This means making sure you send relevant messages to your customers at precisely the time they are most receptive to them. But the devil, as always, lies in the details. Surprisingly, though the details of constructing a unique profile of each of your customers--one rich enough to support effective marketing programs and tailored to individual customer needs--are easier than you might expect. Let me explain.

The best marketing companies are those that have learned to listen effectively to their customers. Companies like Nordstrom and American Airlines have become legends in their respective industries as a result of a relentless focus on customer service. But listening to what customers have to say is, by itself, not enough. Your customers communicate rich and important information about themselves in how they act--or don't act--in their everyday dealings with you. In fact, customer actions are typically a much more reliable indicator of preferences and propensities than are their words. But the trick is to determine which actions are significant, which follow or deviate from a customer's known behavior pattern, and which merit an intervention of some kind. And the only way to do all this is by comparing an individual's most recent behaviors in his relationship with you to his previous behaviors.

By understanding significant changes in a customer's regular behavior, it's possible to detect early which customers are considering changing their...