This is why at Enova Decisions, we focus on helping clients model normal behavior, leveraging both 1st & 3rd-party data. Since majority of transactions are not going to be fraudulent, it’s much easier to establish a baseline of what is normal and therefore makes it much easier to identify when behavior is abnormal. An example of how to model normal behavior is looking at velocity. For example, if on average, 30% of new customers apply with a Gmail email address and one morning, that percentage jumps up to 80%, that may be an indication that your business was victim of a fraud attack, and the applications with a Gmail email address can be isolated for further review.
While abnormal behavior may not necessarily be fraudulent, this information can be used to isolate cases where additional verification and review is necessary. In other words, by scoring transactions on abnormality, businesses have a roadmap toward auto-approving clear cases of normal behavior and auto-flagging abnormal behavior for additional review. So, rather than applying the same level of friction across the board, friction is added only when necessary. Therefore, an efficient and effective fraud detection and prevention program is not necessarily one that is completely automated. Rather, it is one that enables you to automatically identify clear cases of non-fraudulent and fraudulent transactions so that your investigations team can focus their time reviewing the unclear cases.
Want to learn how you can better segment your customers and more efficiently detect and prevent identity fraud?
Contact us today to learn how you can take advantage of our application fraud detection solution accelerator.