Identity Theft - decisioning in action?

I saw this article on Identity Theft and the entirely coincidental press release on Fair Isaac's new Identity Fraud prevention product (Falcon One). Identity Fraud prevention is a classic example of an EDM problem. Why? Well consider the core measures of EDM success - Precision, Consistency, Agility, Speed and Cost. Successful Identity Theft prevention requires:

• Precision - don't reject transactions that really are from the person they say they are from, do reject those that are fraudulent
• Consistency - identity thieves will use any and all channels to attack so you had better respond consistently across channels
• Agility - identity thieves are engaged in an "arms race" with you so you had better be able to keep changing your approach to keep ahead of them
• Speed - no-one is willing to wait while you check so you had better be able to check fast
• Cost - even though identity fraud is expensive there are bazillions of transactions taking place each day that are real so you can't afford to spend too much checking each one for fraud

Doing this well takes predictive analytics (like Neural Networks) to predict risk accurately, rules to enable agility and policy enforcement, and lots of good data. EDM.

Comments

[Metrics 2.0]

Do you have any metrics and stats on identity theft?