One of the early seeds for Wesabe was a conversation I had several years ago about credit card fraud detection. If you’ve ever gotten a call from your bank or credit card company asking whether some of your recent purchases were legitimate, you’ve seen the effects of fraud detection. These companies run increasingly sophisticated programs to find purchases that look “fishy” for one reason or another. Maybe a purchase took place in one city while you were withdrawing cash in a city across the country. Maybe your card was suddenly used to purchase ten cells phones at once. The fraud detection programs look for patterns of “normal” purchases, and try to find cases where something just doesn’t look right. This helps you, and it helps the card companies (since they bear most of the responsibility for fraud).
The conversation I had was with two people who write this kind of software. They were talking about how intricate the fraud detection programs had become over the years, and how much more effective it has gotten. “It’s really interesting stuff,” one of them said, “to see how patterns form in people’s spending. Patterns of purchases speak volumes.”
At the time I didn’t reply, but what I was thinking was, so, what else can you do with that software? If the patterns of “normal” purchases are that interesting, could you get a lot more out of them than just fraud detection?
Of course, banks and credit card companies do just that, and use that data to help set your credit limits and offer you new financial products. Where my thought led me, though, was to the idea of pattern-detection software that could help consumers. Some of the ideas you’ll see in Wesabe are simple beginnings of that process. Now that we’ve been running our software for a few months, I agree with the fraud detection developers: you can learn a lot from looking at patterns of people’s purchases, and I’m happy to say that there’s a huge amount in that data that can help consumers get more for their money.
After all, shouldn’t there be more tools that try to do just that? Help consumers — really help people with their money? It astonishes me sometimes that there aren’t.
Update: There’s a good story on fraud detection software in The Economist this week.