Customer Profiling

Blog Administrator | Analyzing Data, Analyzing Data Quality, Data Quality, Identity Resolution | , , ,

By David Loshin

In my last post I suggested that reviewing the exchanges of value at different customer touch points exposed opportunities for improvement, depending on the context. One of those contexts is framed by who the customer is, what that customer likes and dislikes, and how that customer’s decisions are influenced.

This is more simply referred to as a customer profile. We have already looked at some aspects of customer profiling in previous blog entries, mostly concentrating on the relationship between location and generalizations that can be made about transactions that take place at a specific location.

However, a specific customer’s profile must encompass more than just location data. The most comprehensive models for customer profile take into account all aspects of the decision-making processes presented to the customer, ranging from small decisions such as which links to click though on a web site to large decisions such as committing to buying a product.

The result of each decision presented to the customer adds to the profile, and can be used in analytical models to look for probabilities that certain kinds of decisions will be made at each touch point.

Managing customer profiles implies two things: that at any touch point, the identity of the entity involved can be discerned, and that that entity’s profile can be located and accessed as part of the process. Both of these assertions require identity resolution, and that will be the topic of the next post.