A Fast Fraud Field Guide
By Emily Chin /
10 Aug 2015
It’s a jungle out there. Whether you run a retail site, deal in Bitcoins, offer travel bookings, or engineered a social media app, there’s probably some form of fraud plaguing your business. Fraudsters are on the prowl, trying to capitalize on incentives, easy sign-ups, and anonymous buying.
The first step in fighting fraud is recognizing who you’re dealing with. Think you’re all alone in the struggle against spammers? We’ve got you covered with this quick online fraud field guide.
Hey, big spender!
“I’ll take 22 Rolex watches and 16 Louis Vuitton bags, please!” Can you say stolen credit card? When a user purchases an unusually high value of similar goods, you should (obviously) take a closer look. Likely such a user is trying to maximize a stolen credit card number, cashing out with luxury or high-volume orders before the card is cancelled. Definitely take the time to review these users and orders, since chargebacks hurt a lot.
Your customers and users trust your site. Having your users’ information stolen makes for an awful experience for them and loss of business for you. Sometimes fraudsters will create accounts and post fake profiles or job listings to phish for consumer data. If you find that certain users are capitalizing on your business to lure others into providing personal information, you might have a fraud ring on your site. How can you tell? Pulling user activity, device information, and account creation statistics can help to separate the naughty users from the nice.
They say a picture is worth a thousand words – but what if you notice a certain user posting thousands of nonsense words or diluting your content? Especially in marketplaces where reviews and comments lend validity to offerings (think vacation rentals or social media), spammy messages can hurt good users who just want to get the word out or better understand what they’re looking to buy. It takes too much time to moderate every word that gets posted on your site or app, but using automated tools to detect bots or spam messages keeps out the clutter.
The Million-Account Man
Do a handful of users on your site share a single profile picture? Or maybe you notice some repeating or obviously disposable email addresses (email@example.com; firstname.lastname@example.org; email@example.com)? Smells like fake accounts. With this kind of account abuse, fraudsters are up to no good and often trying to game the system. Whether they hope to spam other users or abuse incentives like coupons, referral codes, and other promotions, these accounts should be shut down.
Testing, Testing, 1,2,3…
Notice a user making lots of small purchase attempts with a bunch of different credit card numbers? Chances are good that you’ve got a credit card tester on your hands. What’s this tester doing? Trying a stack of stolen credit card numbers on low-priced goods or donation amounts in order to ascertain which numbers are still usable and which have been canceled.
Account takeovers? Is someone using your users’ information for evil? Scammers are the bad apples that sour the online experience for your good users. From fake crowdfunding pages to the cybercrimes that grab headlines, scammers can encompass almost any type of fraud.
Fraud comes in many forms and from countless directions. The best thing that you can do is proactively predict and prevent against these cyber criminals. How? By leveraging the power of big data and deep learning. With the global network that Sift Science employs, you can learn from the fraud findings of analysts and businesses across the world. Use the network effect to catch bad behavior and reward your legitimate users; nothing’s worse than a sub-par customer experience for good users.
Organizations small and large, non-profit and for-profit are arming themselves against the fraudsters that are moving increasingly toward online targets. Don’t get left in the dust! Learn how other companies have successfully faced off against fraud and put Sift Science’s machine learning system to the test with a free 30-day trial.