Hello December! If you’re reading this, you might be very short on time and very stressed out. Black Friday & Cyber Monday just passed, holidays orders are up (yay!) but so is your manual review queue. To welcome in the December month, here’s a quick read to help you wrangle your manual review queue as sales (hopefully) continue their upward trend.
Adjust your system
Whatever you use to put orders into a manual review, adjust your rules to account for seasonality and changing patterns. Normal fraud indicators like different shipping and billing, multiple shipping addresses, and/or multiple credit cards become more common in the holidays. If you keep your normal rules, this has likely increased your manual review queue unnecessarily. The goal is to prevent GOOD customers from going to manual review while still flagging the fraud.
To understand what rules to create or which to adjust, review historical data. With some analysis, see which factors triggered the most people going to manual review. Then see how many of those actually resulted in a rejected order/customer. Use this data to narrow the gap.
Be sure to also keep an eye on emerging fraud trends! In 2016, I observed a peak in fraud from Venezuela. Once these trends are observed, create a new rule to flag those orders to manual review.
Expect more chargebacks, maybe loosen up your prevention philosophy
Let the team know that fraud will slip through, it is inevitable. Rather than trying to catch all of the fraud, have your team focus on those that do definitely have a higher risk. For those that seem not very risky, let them slide (assuming you have a healthy monthly chargeback rate).
Another way to focus on spending manual review time where it matters is to set price caps. Customers with low purchase amounts – let them through during the holidays so that your team has the time to focus on the bigger fish.
Isolate key GOOD and BAD signals
Holidays equate to abnormal shopping patterns, which can make discerning fraud vs. gift shopping very hard. Since the holiday shopping behavior is abnormal, it can greatly increase the chance that your team will be rejecting good orders incorrectly.
To help mitigate this, find a core 2-3 indicators of good buyers, then make that fundamental to the manual review process. When your team sees these indicators, accept the order without further analysis. Likewise, create a few core indicators of real fraud. Teach your team to keep their eyes peeled for these signals and try to ignore the noise.
Give manual review team a timebox per order/customer
Only allow up to X minutes for review, depending on your company.
Further analysis will confuse the auditor and increase the chance of making the incorrect decision. Many people are risk averse and don’t want to process a potentially bad order. Create firm caps as to how long a team should review orders for – around 10 minutes has worked well for my teams in the past. Since shopping patterns are very irregular, the more time spent can actually lead to increased confusion for the analyst. Stick to those key GOOD and BAD signals!
With the overwhelming holiday sales boosting your manual review queue, there are two places to focus: reduce the number of GOOD orders going to manual review AND optimize time spent reviewing. Good luck!