The traditional credit/debit card has become passé—at least in terms of having a single device with access to only one funding account. Today, a single card can, at the point of sale, access different accounts to fund the payment transaction. Likewise, a mobile device has the capability to make payments through one or more funding accounts through the use of various apps.
Of course, this convenience requires changes to traditional fraud monitoring techniques. To detect fraud, we now must monitor the payment instrument and the customer’s activity across the selected funding accounts. Here's how:
1. Profile the overall use of the payment instrument in making payments similar to traditional payment card models. Since different funding accounts can be chosen, also look at outlier analysis across the funding accounts, and within the specific funding account chosen, to point to anomalous behaviors.
2. Maintain lists of favorites across funding accounts that indicate, say, how a customer uses his/her debit and credit accounts versus a points account. Such specific detailed monitoring of typical behavior on specific accounts provides a whole new dimension of detection.
3. Analyze the probability that given the transaction details, the customer would have chosen to fund the transaction with the selected account. You may see an already risky transaction as even more risky when it is determined that the legitimate customer is unlikely to have chosen a particular account for a type of transaction.
This analysis can help indicate abnormality in how the payment funding is occurring, and add additional and essential insight into fraud detection for these complex payment devices.