Fraud Detection
Fraud detection screens orders for stolen-card and bad-intent signals before they ship.
What fraud detection is
Fraud detection is screening orders for signs they were placed with stolen cards or bad intent: risk-scoring each transaction on signals like mismatched addresses, odd velocity, and known-bad patterns, before it ships.
Why fraud detection matters
Card-not-present fraud lands on the merchant: the cardholder gets refunded, the store loses the goods, the revenue, and pays the chargeback fee. Screening trades a small review effort for not shipping to thieves, without strangling good orders.
Signals fraud screening weighs
- Address and card verification mismatches
- Velocity: many orders, cards, or accounts from one place fast
- Geography: IP, shipping, and billing that don’t add up
- History: device, email, and address reputation across networks
Frequently asked questions
Should risky orders be blocked or reviewed?
Tiered: block the clear frauds, hold the gray zone for quick human review, pass the rest untouched. Hard-blocking everything suspicious cancels real customers with unusual setups.
What do false positives cost?
The declined sale, and often the customer permanently: someone whose legitimate order was refused rarely retries. Fraud tuning is a two-sided error problem, not a race to zero fraud.