What is CVV fraud detection?

CVV fraud detection is the process of checking the card verification value against the issuer's records during an authorization, then combining that result with other signals to decide whether a card-not-present order is genuine. The CVV is a 3-digit code on the back of Visa and Mastercard cards and a 4-digit code on the front of American Express cards. A mismatch means the person typing the number does not have the physical card in hand.

No single check proves a buyer owns a card. Detection works when several weak signals are read together: CVV result, address match, device history, order velocity, and authentication status.

How does a merchant check a CVV?

The merchant passes the code in the authorization request, and the issuer compares it to the value on file. The reply comes back as a result code, not a simple yes or no.

  • Match: the code is correct. Risk drops, but the order can still be fraud if the full card record was stolen.
  • No match: the code is wrong. Decline the order or route it to manual review.
  • Not processed or not supported: the check never ran. Treat the order as unverified, not approved.

That last case causes real losses. A gateway that returns "not processed" gets misread by some order systems as a pass, which lets a card with no CVV data slip through.

Why do CVV checks fail to stop fraud?

Card data sold in bulk often includes the CVV, so a match proves less than it seems. Fullz records bundle name, address, card number, expiry, and CVV into one file, and a fraudster holding that file passes a CVV check without ever seeing the plastic.

Skimmers and page-jacking scripts on checkout pages capture the code at the moment a real customer types it. Both cases produce a valid CVV attached to a stolen identity.

Recurring billing adds another gap. PCI DSS forbids storing the CVV after authorization, so subscription renewals and stored-card checkouts run with no CVV check at all.

Which signals work best alongside CVV results?

Address Verification System (AVS)

AVS compares the billing street number and ZIP code to the issuer's file. A CVV match with an AVS mismatch on a first-time order is a classic carding pattern: the thief has the card data but guesses the address.

Velocity and BIN rules

Count attempts per card, per IP, per device, and per email over rolling windows. Card testing shows up as many small orders or declines across a range of BINs in a short span.

BIN rules add context. A high-risk BIN paired with a billing address that does not match the BIN's issuing country deserves a second look.

Device and session data

Device fingerprint, IP geolocation, browser language, and typing behavior separate a returning customer from a script. Watch for a device seen across many cards, or a checkout session that jumps countries between page load and payment.

3-D Secure

3-D Secure pushes authentication to the issuer, which challenges the buyer with a password or a one-time code in the banking app. A completed challenge shifts chargeback liability to the issuer for the covered transactions.

How do you build a detection stack that holds up?

  1. Send CVV and AVS on every authorization. Confirm the gateway actually transmits both fields.
  2. Map each response code to a rule, including the "not processed" cases. Never default to approve.
  3. Set velocity limits on cards, IPs, emails, and devices, with tighter limits on new accounts.
  4. Require 3-D Secure above a value threshold or when CVV, AVS, or device checks disagree.
  5. Route disagreements to manual review with a call or email to the cardholder before shipping.
  6. Log every decision and review chargebacks monthly to find rules that miss.

What should a merchant do when a CVV check fails?

Decline, then look at why. A single mismatch from a long-standing customer who mistyped the code is different from five mismatches from one IP in ten minutes.

Do not ask the customer to resend the CVV by email or chat. That puts card data into systems that were never meant to hold it and creates a compliance problem.

For high-value orders, contact the cardholder through a channel already on file. A short callback confirms the order and catches fraud before fulfillment.

Frequently asked questions

Is a CVV match proof the buyer owns the card?

No. It proves the buyer has the code. When the full card record was copied from a breach, the code comes with it.

Can merchants store CVV data after a sale?

No. PCI DSS prohibits storing the CVV, and the issuer does not send it back in any response file. Systems that keep it fall out of compliance.

Does 3-D Secure replace CVV checks?

No. It adds a layer. Issuer authentication catches more fraud than CVV alone, but a stolen card with a passed challenge still reaches the merchant.

Why do some legitimate cards return CVV "not processed"?

Some issuers, regions, and card types do not support the check, and some gateways drop the field. Treat those orders as unverified and lean on other signals.

How fast should a fraud rule react?

Fast. Card testing burns through dozens of numbers in minutes, so blocks based on hourly velocity catch it while daily reviews do not.