What The Sift?

Gut check: When you hear about “online fraud”, what comes to mind? Chargebacks? Data breaches? Purchases made with stolen credit card numbers? Well, did you know that Sift Science can detect many different types of fraud, not just transactional fraud? It’s true! Our very varied customers use us for a wide array of detection and prevention services. Whether you sell socks online or offer a dating service, run a crowdfunding site or offer a sales platform to others, Sift Science can help you catch fraudsters.

Our fraud detection use-cases fall into two camps: Transactional Fraud and Other. I know, illuminating, right? Well let me run through the categories:

Transaction Fraud

Of course, the majority of transactional fraud takes the form of chargebacks. But chargebacks aren’t the only type of fraud that Sift can help merchants — or anyone that works with the exchange of goods or services — detect online. Some of our users leverage Sift Science to detect coupon abuse, promotion/referral code fraud, and funneling. Never heard of funneling? Funneling fraud is when a fraudster creates various accounts for the purchase of garnering promotional credits. For shoppers, some of these types of fraud feel harmless, but their consequences add up for shop owners.

Other Types of Fraud

Yes, this second category is more nebulous. It speaks to the versatility of Sift’s machine learning-based system that fraud analysts use us to detect fake or spam… <blank>. These include:

• fake or spam accounts
• fake classifieds
• fake job listings
• fake resumes
• fake or spam account creation
• spam site messages (you know, like: “HAPPY THANK YOU FOR APPRECIATION CHECK OUT THIS COOL VIDEO”)
• fake content

And those are just a handful of the types of spam that Sift Science helps our users catch. That Sift Science allows you to send any data that you want gives us the flexibility to be truly customizable. For example, messages that appear in ALL CAPS and are poorly punctuated are more likely to be fraudulent than other messages. Yes, perhaps your message sender was eating a sub sandwich and typing with his feet while sleepwalking and the message is legitimate; and that’s where the other thousands of data fields come into play, such as associated email address, behavioral signals, and device ID indicators.

So, do you have a fraud problem? We can help you with it. Our network of analysts is big and growing, teaching our system every time a user labels an account, email address, transaction, or device. Let’s fight fraud together.

Questions? Feel free to tweet us @siftscience or click through our website. We’d love to talk to you.