8.6 Transactional Data

One way of improving the accuracy of a fraud detection system is passing more data with each transaction, such as the Standard Industry Code (SIC) for the product or service being purchased. This data can be used to create a more detailed profile of a cardholder. For example, a $1,000 purchase at The Good Guys of a camcorder might score differently from the purchase of a TV for the same price. The difference between a $1,000 camcorder and $1,000 TV would result in a different risk score due to the fact a camcorder can be unloaded rather easily on an auction site while a large-screen TV cannot.

In addition, frequency of purchases for the same type of product would enable the fraud-detection system to score the transaction accordingly. Was another camcorder purchased this week? Another example where more data could be useful would be in airline tickets. A $500 charge from a travel site is not that suspicious, but a one-way ticket to Paris leaving tomorrow might affect the risk score. New data protocols, such as those based on XML, can enable the exchange of profile information at the time of a user visit and will make fraud-detection systems more accurate.