One of the Biggest Mistakes in Fraud Prevention Is Assuming Every Decline Is a Win

One of the biggest mistakes in fraud prevention is treating every decline as a win. Teams measure fraud they stop. They spend much less time measuring legitimate revenue they stop with it. Aite-Novarica estimates false declines cost merchants $443 billion a year versus about $48 billion in actual card fraud losses, and some estimates say 60-65% of declined transactions may be legitimate customers. Fraud losses are visible. False declines are quiet - and they take the order, the acquisition cost, and often the customer.

One of the biggest mistakes in fraud prevention is assuming that every decline is a win.

Most fraud teams spend a lot of time measuring how much fraud they stop. Much less time is spent measuring how much legitimate revenue gets stopped alongside it.

Research by Aite-Novarica estimates that false declines cost merchants $443 billion globally every year, compared with around $48 billion in actual credit card fraud losses. Another estimate suggests 60-65% of declined transactions may come from legitimate customers.

Fraud losses are visible. False declines are quiet.

The reason this problem gets overlooked is fraud losses are visible. Chargebacks appear in reports, disputes are tracked, and fraud rates sit on dashboards. False declines are quieter.

You don't necessarily see the repeat customer whose $250 order was declined because they shipped it to their office. You don't see the first-time customer who placed an unusually large order and triggered an act. You don't see the customer whose first card failed, tried another, got flagged for velocity, and decided to shop somewhere else.

They just disappear, and the cost isn't limited to the transaction itself. You've potentially lost the revenue, wasted the customer acquisition cost, damaged the customer's lifetime value, and sent someone who was ready to buy directly to a competitor.

Lower fraud can be the wrong incentive

This is why optimising purely for lower fraud can create the wrong incentives. A fraud team can tighten thresholds, introduce more acts, and bring the fraud rate down. On paper, the system looks better. Commercially, it might be performing worse.

The difficult part is that improving authorization rates requires understanding:

• Which declines are preventing fraud?
• Which signals are predictive versus simply correlated with risk?
• Which customers deserve additional verification rather than an automatic decline?
• Where can thresholds differ by customer, transaction, or segment?
• How much additional fraud could you accept in exchange for more legitimate revenue?

Good fraud systems maximise legitimate approvals while keeping fraud within an acceptable level. Sometimes the transaction you decline is a good customer you just lost.

Related: the hidden cost of false positives, how to reduce false declines in Stripe, how it works, and the FAQ.

If you want to see which rules may be blocking good customers in your Stripe or Shopify data - book a demo.

Originally shared on LinkedIn.