detect patterns of suspicious behavior, and all of the participants in the network would in turn have the benefit of that collective visibility in terms of detecting and preventing fraud.”
The key advantage, Hansen says, is that a big bank like JP Morgan Chase will have lots of data from processing its own credit applications. But it would not know about suspicious activity involving credit card accounts at Citibank, or wireless accounts at AT&T or Sprint. “We did it so the data comes in real time,” Hansen says. “But all that comes out of the network, back to the participants, are analytics, statistics, variables—things that can help them make decisions. But the data stays in the network. So AT&T didn’t have to worry that Sprint would have their data.”
Their unspoken goal, of course, was to replicate the success of HNC Software.
“The very first product we called the ID Score and it’s still one of our most successful products today,” Hansen says. “Virtually every one of our clients buys it, and it’s a three-digit number just like a FICO score, and in our case the higher the number the more risk is associated with your identity.” A big bank processing a million credit card applications a month could set the level of risk it was willing to tolerate—like anything over 700—and refer those accounts for additional investigation.
“These enterprises would get a lot of value out of it,” Hansen says. “A large enterprise would take just the 1 or 2 percent of riskiest identities, based on our score, and they could sift out 40 or 50 percent of all the fraud.” Customers write off the remainder as a cost of doing business, because of the costs associated with addressing them (and in confronting consumers over their potentially fraudulent credit applications).
The analytics applied to the data uses pattern recognition technology to identify fraud.
“The best example starts with a new account—applying for a new credit card with Bank of America, or going into get a cell phone at the Sprint store,” Hansen says. “These enterprises only have a couple of seconds to figure out if these people really are who they say they are. So they send a ping to our network, we would do the analysis in half a second on all the stuff we’ve seen on that identity. If that identity had applied for 12 new credit cards in the past 24 hours, then the likelihood of fraud is a lot higher. Or if the address being used with the application for that credit card has been involved with a dozen prior fraud attempts, then it’s a higher likelihood.”
The system can look at all kinds of variables. It might be that
