Special Report: 4 Best Practices for Combating Online Fraud
The good news is that online retailers understand fraud is a growing problem and have taken steps to combat it. Their efforts are making an impact. Studies point to a decrease in the percentage of orders lost to fraud. Our own Annual Trends Survey revealed that for 75 percent of respondents, less than 1 percent of their online orders are fraudulent. While this is a step in the right direction, you shouldn't be satisfied till that figure is 0 percent. Here are tips to help you get there:
1. Gather as much data as possible on every transaction, no matter how trivial it may seem. The more data you have at your disposal, the more likely you are to detect fraud. Using this broad data set as the foundation, correlate each individual transaction element with historical instances of fraud to determine its predictive value. Also, correlate the various combinations of these elements with fraud to uncover the links between data elements that are indicative of fraud.
Behavior analysis is a newer fraud tool for online merchants. It assesses whether a customer's visit and website activity is consistent with that of a typical user. For example, the time spent on the checkout page or the number of pages viewed varies greatly between a typical user and a fraudster using a bot. The use of proprietary data analysis by retailers is likely to increase as fraudsters learn how to circumvent traditional mitigation tools.
2. Conduct manual reviews of orders. Online orders that don't pass through an initial automated screening stage — e.g., card verification number, address verification service, postal address validation services — generally are segmented for a manual review process, where additional information is collected to determine if orders should be accepted or rejected due to excessive fraud risk. That review process often includes evaluating the customer's order history, contacting the customer directly, and using Google Maps and IP geolocation to investigate delivery addresses.