The fraud-to-sales ratio is the value of fraudulent card transactions in a period divided by the total value of card sales in the same period. It is usually expressed in basis points (hundredths of a percent) because the numbers are small: a ratio of 0.08% is written as 8 basis points. Card networks use it to monitor merchants and acquirers, issuers use it to monitor their portfolios, and European payment providers use a version of it to qualify for exemptions from strong authentication.
The ratio measures losses by value, which separates it from the chargeback ratio, which counts disputes of every kind. A merchant can have a low chargeback ratio but a high fraud-to-sales ratio if fraud is concentrated in a few large orders, or the reverse if most disputes are about service rather than stolen cards. It is a core metric in fraud risk management for anyone handling payment fraud.
Fraud-to-sales ratio in plain terms
Think of a shop counting how much of its takings turned out to be paid with stolen cards. If the till took in a large sum this month and only a sliver of that came from crooks, the shop is in good shape. If a noticeable slice of the money was fraudulent, the shop has a problem, even if only a handful of orders were involved.
The fraud-to-sales ratio is that slice, measured by money rather than by number of orders. Banks and card companies like it because it shows how much is actually being lost. A small number of very expensive fraudulent orders can matter more than many cheap ones, and this measure captures that.
Formula and example
The standard calculation, in basis points, is:
- is the total amount of sales later reported as fraud by issuers, or confirmed as fraud chargebacks, depending on the programme.
- is total settled sales value on the same card brand and merchant account. Network programmes often use the prior month's sales, since fraud is reported with a delay.
- Multiplying by 10,000 converts the fraction into basis points; multiplying by 100 instead gives a percentage.
Worked example: an electronics retailer settles $4,000,000 in card sales in June. Issuers later report $36,000 of those sales as fraud. The ratio is 36,000 ÷ 4,000,000 × 10,000 = 90 basis points, or 0.9%. For an e-commerce merchant that is high, and would draw attention from most acquirers.
How networks and regulators use it
- Network monitoring. Mastercard's Excessive Fraud Merchant programme, as described by acquirers in 2026, combines a fraud level of 50 basis points or more with a minimum fraud value and transaction count, and a test of how much volume used 3-D Secure. Visa's acquirer monitoring programme now blends reported fraud with disputes into a single count-based ratio rather than tracking value separately.
- PSD2 exemptions. Under the regulatory technical standards for strong customer authentication, a payment provider may skip authentication for a remote card payment through transaction risk analysis only if its own fraud rate stays under set reference levels. For card payments these are 0.13% for transactions up to €100, 0.06% up to €250 and 0.01% up to €500, measured over a rolling quarter.
- Issuer reporting. Issuers report fraudulent transactions to the networks (for example through Visa's TC40 records), which is the data that feeds merchant-level ratios. Central banks and industry bodies publish national card fraud rates in the same form.
What drives the ratio
The ratio rises with exposure to card-not-present sales, high-value resaleable goods such as electronics, gift cards and tickets, and fast delivery that leaves little time for review. Attacks such as card testing can also raise it, although many tested cards never complete a large purchase. It falls with stronger fraud detection, device and velocity checks, address and card security code verification, and authentication that shifts liability to the issuer.
Limitations
The figure depends on what counts as fraud. Issuer fraud reports can include first-party fraud, where the real cardholder made the purchase and later claimed otherwise, so a merchant may be charged with fraud it could not have prevented. Reports also arrive weeks or months after the sale, so the ratio for recent months keeps rising as late reports come in. Finally, a single average hides where the loss sits: risk teams usually break the ratio down by product, country, channel and payment method before acting. Acquirers that see the ratio climbing may tighten underwriting terms or hold back funds.