Skip to content

PSP operations

PSP Fraud Prevention: Rules, Scoring and Limits

Card-not-present fraud is the loss a PSP sees most often. UK Finance's Annual Fraud Report 2026 puts UK remote purchase fraud losses at £423.5 million in 2025, up 3%, with cases up 13%. Within that, internet and e-commerce card fraud came to £250.8 million, up 9%, and the report names data compromise, including hacks at retailers, as a major driver.

A PSP sits between those losses and its acquirer. It can't stop every stolen card, but it decides how many reach authorization, and its merchants' fraud counts end up in its acquirer's scheme ratios. This page covers PSP fraud prevention in layers, the controls for card testing, the European authentication rules, and how to compare fraud detection vendors.

The layers of PSP fraud prevention

No single control catches most fraud. Each layer stops a different kind, and together they keep good customers moving while bad payments stop early.

Layer What it does Example
Merchant risk Keeps fraudulent or careless merchants off the platform Underwriting and ongoing monitoring
Transaction rules Blocks or flags payments that match known patterns Five cards from one device in ten minutes
Risk scoring Scores each payment from many signals at once A third-party model's score above a set level
Authentication Asks the cardholder to prove it's them 3-D Secure before a high-value payment
Limits Caps the damage while risk is unclear A daily volume cap on a new merchant
Tokenization Makes stolen card data less useful Network tokens in place of card numbers

The order matters for cost. A rule that blocks a payment before authorization costs nothing in fees, while one that catches it after capture still leaves a refund or a dispute.

Rules that catch card testing

Card testing, where fraudsters run stolen card numbers through a checkout in small amounts to see which work, shows up in authorization data before it shows up in fraud reports. Visa now measures it directly: its VAMP program flags enumeration when enumerated authorizations reach 20% of all authorizations (2,000 basis points) and 300,000 transactions, according to its VAMP fact sheet.

A PSP's rules can stop it much earlier than that. These are the usual starting points.

  • cap the number of attempts per card, device or IP address in a short window
  • flag runs of low-value declines on one merchant
  • block card numbers that arrive in sequence
  • require authentication after a set number of failed attempts
  • slow down or challenge checkouts that see sudden spikes from new countries

Tune each rule per merchant. A donation page and a ticket seller see very different normal traffic, so one global setting either blocks good payments or misses bad ones.

Authentication and SCA in Europe

In the EU, Article 97 of PSD2 requires payment service providers to apply strong customer authentication when a payer initiates an electronic payment, with the authentication dynamically linked to the amount and the payee for remote payments. For card payments online, 3-D Secure is how that's done in practice.

Authentication adds friction, so the rules around it matter. Good setups send low-risk payments through exemptions where the rules allow them and reserve challenges for payments the risk score doesn't trust.

How to compare fraud detection vendors for PSPs

A list of the top fraud detection vendors for PSPs only helps if you judge it against what a PSP needs. PSP fraud detection is a different problem from a single shop's, because fraud detection for payment providers has to run for hundreds of merchants at once, each with its own normal traffic and dispute history. Ask these questions.

  1. Can each merchant, or merchant group, have its own rules and thresholds?
  2. Does the scoring API answer fast enough to sit inside authorization?
  3. Does it offer device fingerprinting and behavioral signals, or only card and transaction data?
  4. Does it learn from your chargeback and fraud report outcomes?
  5. Is it priced per transaction, per merchant or as a platform fee?
  6. Where is the data stored, and does that suit your regulator?

The first question rules out many tools. A vendor priced and designed for one merchant gets expensive and awkward when every sub-merchant needs a separate setup.

What's built into our platform

Our white label gateway comes with 150 built-in configurable fraud filters that can be set per merchant, and it connects to third-party risk scoring providers when you want a specialist model on top. Real-time fraud analytics for PSPs work best when rules and scoring see the same transaction data, so both run inside the payment flow.

3-D Secure runs in the same flow, including through an external MPI, while tokenization and network tokenization reduce what a breach can expose. For the merchant side of fraud, the platform's merchant monitoring tracks fraud and dispute ratios against scheme thresholds.

Questions

What is PSP fraud prevention?

The controls a payment service provider uses to stop fraudulent payments across all its merchants: merchant underwriting and monitoring, transaction rules, risk scoring, authentication, limits and tokenization.

How much does card-not-present fraud cost in the UK?

UK Finance reports remote purchase fraud losses of £423.5 million in 2025, of which £250.8 million came from internet and e-commerce card fraud.

What's the difference between fraud rules and fraud scoring?

Rules block or flag payments that match a set pattern, such as too many attempts on one card. Scoring weighs many signals at once and gives each payment a risk score that rules then act on.

Does PSD2 require 3-D Secure?

PSD2 requires strong customer authentication for electronic payments a payer initiates, with exemptions in defined cases. For online card payments, 3-D Secure is the standard way to meet it.

Review your fraud setup with us

Bring the fraud rules you run today, and we'll show how they map onto the platform's filters and scoring.