Skip to content

Payment fraud prevention: types of fraud, card testing, 3-D Secure and the controls that work

Payment fraud prevention starts with knowing where fraud occurs. Learn card testing, 3-D Secure, velocity rules, account takeover and bank transfer fraud.

Payment fraud prevention means stopping fraudulent payments without turning away good customers. It works by layering controls where each type of fraud occurs: card testing is stopped by velocity limits, stolen-card fraud by 3-D Secure and address checks, account takeover by strong login controls, and bank transfer scams by payee verification and limits.

This guide organizes fraud by type, with the signals, controls and trade-offs, as of November 2026. It connects to chargeback ratios and Visa and Mastercard thresholds, PCI DSS compliance and KYC and KYB in payments.

Key takeaways

  • Fraud differs by channel. Match each control to the fraud it stops, and review the cost of each control to conversion.
  • Card testing is cheap for attackers and costly for you. Velocity rules and bot defenses stop it early.
  • 3-D Secure shifts the identity question to the issuing bank, at the price of occasional friction.
  • Bank transfers have no chargebacks, so prevention and verification before sending matter more.
  • Card networks monitor both fraud and enumeration. Visa counts fraud reports in its ratio.

What are the main types of payment fraud?

The main types are stolen-card fraud in card-not-present sales, card testing, account takeover, friendly fraud and authorized push payment fraud. Each has different signals and controls.

Fraud typeHow it worksSignalsControls
Card-not-present fraudA criminal uses stolen card details onlineMismatched billing and shipping, new device, odd order size3-D Secure, address and security code checks, risk scoring
Card testing (enumeration)Bots try many numbers to find valid cardsBursts of small attempts, many declines, same device or IPVelocity limits, CAPTCHA or bot defense, decline caps
Account takeoverA criminal logs in as a real customerNew device, password reset then purchase, changed emailMulti-factor authentication, login throttling, step-up on changes
First-party (friendly) fraudThe cardholder buys and then disputesRepeat disputers, delivery confirmed, usage logsClear descriptor, delivery proof, signed terms, fast refund
Authorized push payment fraudA victim is tricked into sending moneyNew payee, urgent request, changed bank detailsPayee verification, new-payee limits, delay for large first payments

How does card testing work?

Card testing works by running many cheap authorization attempts to find which card numbers are live. A bot submits small purchases or card verifications in bulk, and the cards that succeed are used or sold.

It hurts you in four ways: authorization fees on every attempt, damaged approval rates, reputation with issuers, and monitoring. Visa's Acquirer Monitoring Program also tracks an enumeration ratio, which is enumerated authorizations divided by total authorizations. Per Checkout.com, Solidgate and Shopify, a merchant is monitored when it has at least 300,000 enumerated transactions in a month and a ratio of 20% or more (2,000 basis points). Shopify describes this as monitoring only, with no fines. Visa's primary document was not located, so confirm the figure with your acquirer.

Do not wait for that threshold. Attackers can do damage at much lower volumes.

How do 3-D Secure and tokenization help?

3-D Secure has the issuing bank authenticate the cardholder before the charge, and tokenization replaces the card number with a token. Together they reduce both fraud and exposure.

  • 3-D Secure is often invisible to the shopper. The bank approves low-risk payments silently and challenges risky ones. A successful result is strong evidence in a dispute and can shift liability.
  • Tokenization means a stolen database holds tokens, not usable card numbers. It also helps you stay at a smaller PCI scope (see PCI DSS compliance).

The trade-off is conversion. A challenge adds friction and some shoppers abandon. Apply it by risk, not to every order.

Which velocity, device and behavior signals matter?

The strongest signals are velocity (how often), device and IP (from where) and behavior (how). One signal is weak, and several together are strong.

  • Velocity: attempts per device, IP, card, email or account in a time window.
  • Device and IP: new device, proxy or data center IP, many accounts on one device.
  • Address and card checks: billing address mismatch and security code mismatch.
  • Behavior: instant form fills, no browsing before purchase, unusual order patterns.
  • History: past undisputed orders, past disputes, account age.

Here is a worked velocity rule. The figures are illustrative.

ItemValue
RuleBlock a device or IP after 5 authorization attempts in 10 minutes
Normal shopper1 to 2 attempts, never trips the rule
Card-testing bot60 attempts in 10 minutes
Attempts blockedThe rule allows 5 and blocks 55 of 60, or 91.7%
Cost of false positives30 genuine shoppers a day behind shared office IPs trip the rule
Value at risk30 shoppers x $50 order = $1,500 a day
MitigationStep up to 3-D Secure or email verification instead of a hard block for the 30, recovering part of that value

The point of the example is the trade: the rule blocks nearly all of the attack, and the cost is small if you step up instead of blocking.

How does account takeover happen?

Account takeover happens when a criminal gets a customer's credentials and acts as them. They come from data breaches, phishing and credential stuffing, which tries leaked passwords across many sites.

After login the attacker changes the email, adds a payee or card, or makes purchases. Defenses are multi-factor authentication, throttling login attempts, alerts on changes to contact details, and a step-up check before sensitive actions such as adding a bank account. These controls add friction, so scope them to the moments that matter.

What fraud affects bank transfers and payouts?

Bank transfers face authorized push payment fraud, account takeover and fraud by the sender or the recipient. They have no card network dispute right, so controls come before the payment.

Instant rails raise the stakes because the money is gone quickly. Pix has a Special Return Mechanism (MED) that lets a participant's payment provider debit an account identified as fraudulent, per the Banco Central do Brasil guide, so some recourse exists. In the US, Nacha's fraud monitoring rules took effect in two phases in 2026, on 20 March and 19 June (22 June in practice, because 19 June is a federal holiday), as described on its Phase 2 page. Controls that work: verify the payee and bank details, hold new-payee payments for a short review, set limits by customer tier and monitor for changes in payout details. See ACH payments explained and how to evaluate automated risk monitoring vendors.

How do you balance fraud and approval rates?

You balance them by applying friction by risk. Low-risk orders go straight through, medium-risk orders get a step-up, and high-risk orders are declined or reviewed.

  1. Measure your approval rate, fraud rate and dispute rate together, not separately.
  2. Set a risk score band for approve, step up and decline.
  3. Review declines. Sample them to see good customers you lost.
  4. Test each rule on historical data before you turn it on.
  5. Track the cost of each control against the loss it prevents.

Over-blocking is a cost too. If your decline rate rises, your chargeback count may fall while your revenue falls faster. Also watch the networks: Visa's merchant ratio counts both disputes and fraud reports, so catching fraud early helps your ratio. For thresholds, see chargeback ratios and Visa and Mastercard thresholds.

A fraud controls checklist

Use this list as a starting baseline and adjust it to your risk.

  1. Add velocity limits per device, IP, card, email and account.
  2. Put bot defense or a CAPTCHA on card-entry and login pages.
  3. Cap declined attempts per session and per payment.
  4. Run address and security code checks and score every payment.
  5. Use 3-D Secure on medium and high-risk payments.
  6. Tokenize card data and keep it out of logs.
  7. Require multi-factor authentication for logins and sensitive changes.
  8. Verify payees and hold first payments to new payees for large amounts.
  9. Monitor fraud reports, disputes and enumeration ratios monthly.
  10. Review rules quarterly with approval-rate data.

Where BlindPay fits

BlindPay handles fraud controls on the payout side through KYC, KYB, sanctions screening and transaction monitoring inside the API, so each payment runs through a verified customer. Monitoring flags suspicious payins and payouts for compliance review, as described in the on-hold transactions docs. See KYC and KYB in payments. BlindPay is not a card acquirer, but it offers card payins up to $500 per payin with 3-D Secure required, a billing address check, and up to 3 declined attempts per payin, per the card payins docs.

Methodology and sources

Checked in November 2026. Enumeration figures come from processor documentation, and the velocity example is illustrative.

FAQ

One API.
Multiple payment rails.

Blind Pay, Inc. (“BlindPay”) is a United States-based financial technology and payments infrastructure company, incorporated in the State of Delaware on May 2, 2024. BlindPay is duly registered as a Money Services Business (“MSB”) with the Financial Crimes Enforcement Network (“FinCEN”) under registration number 31000277566153. BlindPay is not a state-licensed money transmitter and does not, in its own capacity, engage in regulated money transmission services. All such regulated activities are conducted exclusively through, and under the licenses of, duly authorized financial institution partners. BlindPay's role is limited to providing financial technology, virtual account management, payment processing, and related payments infrastructure services that facilitate payment flows and account functionalities. These services are provided in coordination with its licensed partners and in compliance with applicable federal MSB requirements and relevant state regulatory frameworks.

Blind Pay, LLC (“BlindPay”) is a Delaware limited liability company formed on June 24, 2025. BlindPay is duly registered as a Money Services Business (“MSB”) with the Financial Crimes Enforcement Network (“FinCEN”) under registration number 31000318334247 and is listed in the Nationwide Multistate Licensing System (“NMLS”) under NMLS ID #2745309.

BLIND PAY SOCIEDADE PRESTADORA DE SERVIÇOS DE ATIVOS VIRTUAIS LTDA. (“BlindPay”) is a Virtual Asset Service Provider (“VASP”) operating in Brazil. The company is presently completing the regulatory adaptation process required of Virtual Asset Service Provider Companies (SPSAV) under Central Bank of Brazil Resolution No. 520/2025, and remains subject to the transitional regime set forth in Article 88 of that Resolution. Registered address: Avenida Princesa Isabel, nº 15, Sala 1405, Edifício Martinho de Freitas, Centro, Vitória/ES, 29010-361.

BlindPay is a stablecoin API that enables individuals and businesses to move money worldwide.