As an e-commerce fraud analyst, you’re expected to decide whether a transaction is good or bad, often with ambiguous transaction and customer data. This can leave you feeling like Lucy, especially during the holiday season.
In the absence of a fraud detection system, here are five signals you can use to assess fraud risk. Remember, these are aggregate signals based on data from many companies. Your mileage may vary.
- Fraudsters have stacks on stacks of cards. If the customer has multiple credit cards on file from different banks, their order is 7x more likely fraudulent.
- fraudsters dislike capital letters. if a customer wrote their billing name in all lowercase letters, the order is 2.7x more suspicious.
- Fraudsters stay (virtually) on the move. A buyer with multiple billing zip codes within a week is 30x more likely to be fraudulent.
- Fraudsters favor disposable email addresses. An email address with two or more digits is twice as likely to be fraud than one with zero or one digit.
- Fraudsters are night owls. Transactions at 2AM are 50% more likely fraudulent, while 4AM transactions are 100% more likely fraudulent.
How can you further improve your fraud detection accuracy? Customization. Advanced fraud detection solutions like Sift Science can incorporate data unique to your business into our scoring. We call these custom events.
For example, an online shoe store like Zappos could send us the shoe size for each transaction as a custom event. It might turn out that size 10 shoes are more fraudulent than size 15 shoes. This makes sense intuitively: there are more people walking around with size 10 feet and fraudsters often focus on goods they can resell easily. Wondering how Sift Science can solve your e-commerce fraud challenges? Drop us a line, we’d love to help.