Skip to main content
Real-Time Learning About Changing Customer Behavior

I’m just back from the Credit Scoring and Credit Control XIII conference in Edinburgh, where last week several FICO colleagues and I presented on various topics around analytic innovation. The conference is widely recognized as Europe's top conference for credit scoring and related topics, with the noble goal of bridging the “academic practitioner divide.”

Indeed, that's a key goal for my fraud analytics team here at FICO—finding new ways to apply advanced analytics in real-world business scenarios. It was the central theme of my conference session “Big Data Developments in Transaction Analytics,” where I highlighted two new streaming analytics that we are using to improve real-time fraud detection:

  • Analytics that learn from individual behavior. We're all creatures of habit. We may have, for instance, favorite ATMs close to work or home, favorite gas stations along a daily commute, and preferred online stores for shopping. FICO has developed an advanced algorithm that identifies and updates these favorites in real time. The more frequent entries appear with greater recurrence and are ranked at the top of the favorites list, while less frequent entries fall away and are replaced with new entries. These analytics provide a more granular view into an individual's typical behavior based on their stream of transactions, improving a bank’s ability to distinguish between normal and suspicious transaction activity.
  • Analytics that learn from global behavior. While individual behavior patterns can tell us a lot, so too can the behavior patterns of other customers with similar characteristics. That's why we've developed patent-pending Collaborative Profile analytics that examine patterns of transactional behavior across millions of customers to continuously update “behavior archetypes” for individual customers. Think of these archetypes as the “atoms” in the periodic table of customer behaviors. Real customers are “compounds,” composed of a unique mixture of various archetypes. This archetypal breakdown is updated in real time, with each transaction. By looking at individuals through this global lens, it improves our understanding of customers in the context of similar customers, enabling banks to anticipate new behaviors.

I had a very engaging dialogue about these customer behavior transaction analytics with conference attendees. For many, it was the first time they considered such challenges as 10 millisecond decisions.

If you missed us in Edinburgh and are interested in learning more about the analytics I describe above, I invite you to download the Insights white paper that explores them in greater detail (requires registration).

related posts