Hey, you! Take out your wallet for me. Inside that wallet, how many credit cards do you have? Two? Three? Did you know that, on average, Americans have 2.6 credit cards per eligible person. In this final installment of the “10 Warning Signs of eCommerce Fraud” series, we examine the patterns related to credit card and billing data, signals 6, 9, and 10. With that in mind, when merchants encounter IP addresses with more than two credit cards associated, a red flag might be raised. Considering the number of cards that the average American carries, too many credit cards can signal fraud.
Perhaps your shop is the victim of a single fraudster with a stash of stolen credit card numbers. Or is a reshipping service at work? Perhaps it’s friendly fraud, with a family member using Mom, Dad, Uncle Roger, and Aunt Linda’s cards all at once. It could even be “liar-buyer fraud” (oooh).
There are, however, plenty of explanations for a high credit-card-to-IP-address ratio. Perhaps that IP address is linked to a business in a busy metropolitan area where employees often order personal items. Maybe the IP address belongs to an especially savvy online shopper that only attaches one card to each merchant site to quickly shut down fraudulent activity on his accounts. Regardless, merchants risk off-putting good customers if they hold or cancel an order without seeing the user’s big picture.
Similarly, email age and history contain telltale signs. Quick question: how old is your email address? How about the email address that you use for your Amazon or Netflix accounts? Like most users, chances are good that the email addresses that you use for your purchases and account creations have a history. Not so much for fraudsters.
Sometimes, criminals create email addresses purely for the purpose of creating accounts on merchant sites. Orders from these newly-hatched emails that don’t have attached billing relationships are suspicious.
Big data from a global network offers a solution to this conundrum. Yes, there are basic IP address matching services, but machine learning can also streamline your fraud detection process. In order to quickly sort those fraudulent orders from good ones, allow global data to quickly and accurately pinpoint those IP addresses with a history of fraud.
All in all, fraudsters leave signs of their dirty work. We’ve walked through some of their most common tells, but using rules to fight criminals is too rigid and outdated to work. Cybercrime takes advantage of all the anonymity and internationality that the worldwide web has to offer — fight back with global data to stay ahead of fraud. Like training a laser-focused guard lion, machine learning can fit your unique needs.
As always, our free Guides and Resources library can provide extra insights into fraud, ecommerce, merchant challenges and solutions. Feel free to check our webinars, and tweet us questions @siftscience.