What is fraud screening? Definition and example

Fraud screening

Fraud screening is checking each payment against rules, lists and risk models before or during authorization, to approve it, block it, send it to manual review or ask for 3-D Secure, based on how likely it is to be fraudulent.

Also called: fraud check, risk scoring, risk score, fraud scoring, fraud detection, transaction screening

Updated 3 November 2026

How does fraud screening work?

A screening step collects signals about the payment: card and BIN data, amount, customer account age and history, device and IP address, email, shipping and billing addresses, and the results of security code and address checks returned by the issuer. Rules and models turn these into a decision. Stripe’s documentation describes four possible actions on a payment: request 3-D Secure, allow, block or send to review. Screening can run before authorization, to avoid sending risky payments at all, or after, before the order ships.

Why does it matter?

It decides two costs at once: fraud losses and chargebacks on one side, false declines on the other. As an illustrative example: a store with 50,000 card payments a month tightens its rules and cuts fraud disputes from 100 to 40, but blocks 400 more payments, half of them genuine. Whether the change pays depends on the average order value and the cost of each dispute, so both numbers need to be tracked.

What are good practices?

Use several signals rather than one, send borderline payments to authentication rather than blocking them, feed confirmed fraud and chargebacks back into the rules, and review blocked payments regularly. New customers and returning customers with a clean history deserve different thresholds.

Where does Omno use it?

Omno scores every transaction for fraud before it is processed, alongside any checks run by PSPs or the merchant’s own fraud tools. Its risk controls add customer rules by tag and volume, customer blocking, fully or per method, limits per provider and method, and card blocking with blacklists shared across partner merchants, so cards linked to fraud or chargebacks are stopped before any provider sees them. Customer groups and customer verification levels decide which providers a customer can reach.

Frequently asked questions

What is a risk score?

A number from a fraud model that estimates how likely a payment is to be fraudulent. Stripe's Radar, for example, gives scores from 0 to 99, treats 65 or above as elevated risk and 75 or above as high risk by default, and blocks high-risk payments unless the merchant changes the rules. Where to set thresholds is a business decision.

Should risky payments be blocked or challenged?

Often challenged. A 3-D Secure challenge lets a genuine customer prove who they are, and successfully authenticated payments typically shift fraud liability to the issuer. Blocking suits clear fraud, such as known stolen cards. A middle band of scores is where challenges usually work better than blocks.