Fraud

Card cycling and cross-merchant card use

Many different cards added to or used on a single customer account, device, or email in a short time, or one card used across several merchants in a short time.

Many cards used on one account, device, or email in a short time, or one card used across several merchants in a short time One account, device, or email Many cards, short time or Merchant Merchant Merchant Several merchants, short time

What it catches

A fraudster who holds a batch of stolen card details does not stop at the first decline. They add card after card to the same account, or run them through the same device and email, until some go through. The approved cards are then spent, or used to load value that can be moved on: a wallet top-up, a prepaid or virtual card, an e-money balance. Other names: card cycling, card churn, too many cards per account, payment method cycling.

Visa’s guidance for merchants on enumeration and account testing names this directly. Under account creation it recommends to “Limit the number of cards that can be added per ‘account’ and session” and to “Monitor the frequency of payment method changes on accounts”; under anomaly detection it says that “multiple transactions with different payment accounts using the same email address and same device ID, may be a trigger for fraud classification or review”. For cards issued in the EU and EEA, the EBA and ECB report that in most card payment fraud the fraudster issues the payment order directly, “often using stolen credentials for remote card fraud”.

It also has an AML form, for issuers of prepaid and virtual cards and other e-money products. The FATF asks providers to have “systems in place to detect if a customer holds multiple cards or accounts”, because that can show a customer “circumventing the CDD procedures by structuring the funds into several ‘low risk’ products”, and to detect “cases where several products are funded by the same source” and where a customer “approaches a limit (on one product/transaction or cumulatively) beyond which full customer due diligence has to be applied”. The customer in question may not be one verified identity: the FATF’s 2010 passage is about products exempt from customer due diligence, and its 2013 guidance warns that buying prepaid cards online “could allow for multiple cards to be acquired by the same person using different names”. In one FATF case, 61 stolen credit card numbers were used to fund virtual prepaid cards, which then overpaid university tuition so the overpayment came back as a cheque from the university.

The second form, cross-merchant use, is the platform view: one card used at several merchants on the same platform in a short time. It suits payment service providers, aggregators and marketplaces that see many merchants. A stolen card can be spread across several merchants before it is blocked. No regulator or standard setter reviewed for this page describes cross-merchant use as an indicator, and the Visa guidance cited here does not name it either. It rests on the mechanism and on Visa’s general advice to set thresholds “on the number of transactions within a specified timeframe”; Visa’s nearest example is per card, not per merchant: “five authorizations from one IP address or PAN within a set timeframe”.

What it does not catch: a stolen card used once on an account that holds only that card; many cards spread across many fresh accounts, one card each, when the accounts share no device, email or IP address (Form B catches them when they do; otherwise see New account activity and Shared identifiers across customers on phone or address); and one card paying several customer accounts, which is the payment-instrument variant of Shared identifiers across customers.

How the rule works

Count distinct cards against the thing a fraudster keeps constant, and distinct merchants against the card.

  • Form A, cards per account. Distinct cards added to or used on a single customer account in a window.
  • Form B, cards per device or email. Distinct cards used from one device, email or IP address in a window, across accounts, with a much higher count for IP addresses. Catches the fraudster who opens a new account for each card.
  • Form C, cycling after declines. A declined attempt on a card followed within minutes by an attempt on a different card on the same account or device, repeated. Declines that point at the card details count; declines for lack of funds or a technical failure do not.
  • Form D, cards per person (issuers). Prepaid or virtual cards and other products held by one person, linked by the verified identity and also by device, IP address, funding source, or name and date of birth; products funded from one source; and a person whose products together approach the limit above which full due diligence applies.
  • Form E, one card across merchants. Distinct merchants at which one card was tried on the platform in a window, approved or declined.

Decide what “in the window” means before you set a number. Counting cards first linked to the account in the window catches new cards arriving; it misses a fraudster who rotates through cards already on file. Counting cards used in the window catches both.

Looks at

Every card payment attempt, approved or declined, and every card added to a customer’s account. For issuers, also the cards and products each customer holds, how each is funded, and the device, IP address, name and date of birth behind each.

Alerts when form A: cards per account

Counting this attempt or added card, 4 or more different cards were used on, or added to, the same customer’s account in the last 7 days.

Alerts when form B: cards per device or email

Checked separately for the device, the email address and the IP address of each attempt, across all accounts. Either of these, counting this attempt:

  • 4 or more different cards were used from the same device or email address in the last 7 days.
  • your number, much higher or more different cards were used from the same IP address in the same window.

Alerts when form C: cycling after declines

In your window before an attempt, your number or more other cards were declined on the same account, or from the same device, for a reason that points at the card details (such as a failed CVV check, an invalid card number, a lost or stolen card, or suspected fraud).

Alerts when form D: cards per person

For issuers of prepaid and virtual cards and other e-money products. Products count as one person’s when they share the verified identity, a device, an IP address, a funding source, or a name and date of birth. Any of these:

  • One person holds more cards or products than your product’s limit.
  • Several of the person’s products are funded from one source.
  • The person’s products together come within your margin of the limit above which full due diligence applies.

Alerts when form E: one card across merchants

For platforms with many merchants. Counting this attempt, one card was tried, approved or declined, at your number or more different merchants in your window. Your own and test merchants are not counted.

Counts as the same card

Cards are told apart by the Payment Account Reference or the processor’s card fingerprint. Not by the full card number, which brings PCI DSS into scope wherever it is stored, and never by a token that changes per merchant or per device.

The alert shows

  • Form A: the customer and the cards.
  • Form B: the device, email or IP address, and the cards.
  • Form C: the account, the declined attempts with their decline reasons, and the attempt that followed them.
  • Form D: the identities, the cards or products, what links them, and where they were funded from.
  • Form E: the card, the merchants, and which attempts were declined.

Highlighted values are settings you choose. Hover over one, tap it or tab to it for a suggested value.

Alert once per account, device or card and add later events to the same case, rather than re-alerting on every attempt.

Data it needs

ShowHide
FieldUsually comes fromIf it’s missing
A stable card key: Payment Account Reference (PAR) or the processor’s card fingerprintProcessor, card vault, token service providerWithout it, one card shows as several (a wallet token per phone, a new number after reissue, a different token per merchant) and Form A over-counts. A per-merchant token makes Form E impossible.
Whether the credential is a wallet or device tokenProcessor fieldsA phone wallet and the plastic card look like two cards.
Customer id, and what it actually identifiesYour account modelIf the “customer id” is itself the card, or a per-checkout id, every customer has one card by construction and Form A can never fire. Check this first.
Card-added events with a timestampYour app or vaultForm A sees only cards that were charged, not cards added and abandoned.
Every attempt with its outcome and decline reasonAuthorisation responsesWithout outcomes, Form C cannot run, and Form E cannot show which attempts failed, so a customer retrying at another merchant looks like a card spreading. Without reasons, Form C counts declines for lack of funds or a technical failure as if they pointed at stolen details.
Device id, email, IP addressCheckout session, device SDKForm B cannot run.
Stable merchant idYour merchant recordsKeyed on a name rather than an id, one merchant can count as two.
List of your own and test merchants, test BINs and staff cardsYour operations teamForm E fills with internal test traffic.
Cards or products issued per customer, their funding source, and the device, IP address, name and date of birth used for eachIssuer platformForm D cannot run, or sees only products under one verified identity and misses one person using several names.

Thresholds

Each row is a setting the rule needs. Each value is a starting point: test it on your own history, then adjust. Where there is no public value, set it from your own data.

ParameterStarting pointWhere it comes from
Form A, cards per account4 or more distinct cards in 7 daysVisa recommends limiting cards added per account and session but gives no number. Two cards on a single account is ordinary (a second card after a decline, a partner’s card), so a threshold of two can flood review. Start here and lower it only after counting how many cards your customers actually hold.
Form B, cards per device or email4 or more distinct cards in 7 daysVisa names several payment accounts on the same email and device as a review trigger, without a number. Raise it for devices that are legitimately shared (an agent’s or shop’s terminal).
Form B, cards per IP addressNo public value; much higher than for device and emailMany unrelated customers can sit behind one mobile-network or public Wi-Fi address. Set it from the number of distinct cards your busiest shared addresses see, use a shorter window, or leave IP addresses out.
Form C, declined cards before a new cardNo public valueSet it from your own data: count how often a genuine customer tries a second or third card after a decline, and start above that. Needs decline outcomes and reasons.
Form C windowNo public value; minutesThe pattern plays out within minutes. Set it from your own data.
Form D, cards per person (issuers)No public valueThe FATF says providers should be able to detect it, without a number. Start from the product’s own limits, applied to all products linked to the person: a person who holds more cards than the product allows is the case to review. Several products funded from one source are reviewed whatever their number.
Form D, margin to the due diligence limitNo public valueThe FATF says systems “should detect where a customer approaches a limit” above which full due diligence applies, without a number. Set it from your product’s limits.
Form E, merchants per cardNo public value; set from your own dataCount distinct merchants per card over a short window on your platform, approved and declined attempts alike. Paying two merchants in a day can be ordinary, so start above the count that ordinary customers reach once your own and test merchants are excluded.

By country

  • Nigeria. The CBN’s Baseline Standards for Automated AML Solutions (10 March 2026) say that where an institution’s AML solution is also used for fraud monitoring, it shall “Monitor activities across relevant channels (e.g. cards, e-channels, deposits, lending) in real time or near real time to detect unusual patterns indicative of fraud”, and that fraud controls, “particularly for card and electronic channels, shall be calibrated to the faster intervention timelines appropriate to those channels” (section 5.6(a)(i)). The standards do not name card cycling. In the NFIU’s 2023 study of money laundering through fraud, payment service providers that answered its survey named “card theft, card testing and chargeback fraud” among the fraud patterns they observed linked to money laundering, and casino operators reported the control that “no betting wallet can be funded with a third-party account or debit card” (chapter 6). Survey answers, not an NFIU finding of its own.

    Reporting deadline

    Report to the NFIU immediately, with a written report within 24 hours, whether the transaction “is complete or not” (Money Laundering (Prevention and Prohibition) Act 2022, s.7(1)–(3)). The NFIU’s December 2024 STR guidelines start the 24 hours once the transaction is found suspicious, after an examination of no more than 72 hours, and require written reasons for every alert closed as not suspicious (STR-NFIU-2024-A0001, Part Six, para 2).

  • Côte d’Ivoire. In a case in GIABA’s 2025 cybercrime typologies, police found “several credit/debit cards associated with banks from France, and not belonging to” the suspect, who had made deposits from France to mobile accounts that were withdrawn as soon as they arrived. Suspicious transactions are reported to the CENTIF under the UMOA uniform law, which Côte d’Ivoire brought into national law by an ordinance adopted on 23 November 2023.

    Reporting deadline

    Report to the CENTIF “immédiatement” (Loi uniforme 2023, art. 60). The law defines “sans délai” as at most 24 hours (art. 2, point 58); “immédiatement” is not defined, and our reading is to treat 24 hours as the outer limit. Do not carry out an operation you suspect until the report is made, and then only once the CENTIF has not opposed it: an opposition, made before the execution deadline you state in the report, blocks the operation for up to four days unless a court extends it (arts. 61 and 65). If it has already gone through, inform the CENTIF “sans délai” (art. 61). The article numbers here are the uniform law’s.

  • West Africa (GIABA). GIABA’s 2025 cybercrime typologies open with electronic card fraud. A Ghana case shows the receiving side: a company owner took orders through a website and was paid through his bank’s online payment platform; within two months he “received GHS1,468,447.59 (USD341,500) into his account from several credit cards from different jurisdictions”. Visa and Mastercard chargebacks exposed the payments, and the cards were later found to be stolen. Many stolen cards paying one account is what a bank or payment provider sees from the merchant’s side.

  • All markets. The FATF guidance is written for providers of prepaid cards, mobile payments and internet payment services in any market; Visa’s merchant guidance is global.

False positives and tuning

ShowHide

Common innocent explanations: a card expired, was lost or was reissued with a new number; a customer adds a second card after the first is declined, often for lack of funds, an issuer outage or a failed authentication step; a household where partners and children pay from a single account or device; a phone wallet and a watch wallet each carrying their own token for the same card; single-use or per-merchant virtual card numbers, which by design look like a new card each time; a business account with several staff cards; a customer who pays two different merchants on the same day.

Typical causes of misfires:

  • Tokens counted as cards. Phone and watch wallets, per-merchant tokens and reissued cards all change the number you see while the underlying card account stays the same. Key on the Payment Account Reference where you can get it: it is linked to the card number “and successor PANs” and “associated with all affiliated Payment Tokens” (PCI SSC). Otherwise key on the processor’s card fingerprint, and expect a reissued card with a new number to count as a new card.
  • Two is normal. A threshold of two cards per account or two merchants per card catches the second-card-after-a-decline and the household. Start higher and lower only after measuring.
  • Shared IP addresses. Mobile networks and public Wi-Fi put many unrelated customers behind one IP address, so Form B keyed on IP counts strangers’ cards together. Key it on device and email first; on IP, use a much higher count or a short window.
  • First-seen versus used. A window applied to when a card was first linked measures new cards arriving, not rotation through cards already on file. A rule built that way can never fire on a fraudster who loads cards in advance. Count cards used in the window if rotation is what you want to catch.
  • A customer id that cannot hold more than one card. Where the customer identifier is the card itself, or a new id per checkout, cards per customer is always one and the rule is structurally blind. Check what the id means before trusting a quiet rule.
  • Pointing the wrong way. Many cards on one account and one card paying many accounts are different signals. Where anyone can pay into an account, the second can be the stronger one. Run the reverse direction as well, through Shared identifiers across customers.
  • Every decline counted alike. A decline for lack of funds, an issuer outage or a failed authentication step says nothing about whether the card is stolen, and it sends genuine customers to their next card. Declines that point at the card details (a failed CVV check, an invalid card number, a lost or stolen card, suspected fraud) are the ones Form C is about. Visa advises analysing decline codes such as “CVV failed, Do Not Honor, Suspected Fraud” to spot enumeration. Count only those reasons, or weigh them more.
  • Your own test traffic. Scheme test cards, test BINs and staff cards used to check that payments work can spread one card across several merchants within minutes. Exclude them by card, BIN and merchant before counting.
  • Retries at another merchant. A customer whose payment fails at one merchant may try the same amount at another. With no outcomes, this looks like one card spreading across merchants. Get decline outcomes, and look at whether the attempts are the same amount repeated.
  • A quiet rule is not a proven rule. A rule that rarely fires may be blind rather than clean. Test it on a known or synthetic case before relying on it.

How to tune: get a stable card key; count cards used, not only cards first seen; start at four cards in seven days per account, per device and per email, and much higher per IP address; exclude your own merchants and test cards; add Form C once you have decline outcomes and reasons; review the reverse direction (one card, many accounts) on the Shared identifiers page.

When it fires

ShowHide

Analyst checklist:

  1. List the cards: issuer, issuing country, BIN, cardholder name where you have it, when each was added and first used, which were declined and why.
  2. Do the cards plausibly belong to the account holder? Different names, many issuing countries, sequential or same-BIN cards, and cards added minutes apart point to stolen details. Two cards from local banks used days apart point to a household or a replacement card.
  3. Look at what the approved cards bought or funded: goods shipped, a wallet loaded and drained, value moved to another account or to crypto. Stop fulfilment or withdrawal where you can.
  4. Check the device, email and IP address for other accounts (see Shared identifiers), and whether the account is new (see New account activity).
  5. For cross-merchant alerts, check whether the merchants are your own or test merchants, whether the attempts were declined, and whether the same amount was retried.
  6. Tell your acquirer or processor about the cards; block the cards on your platform; as an issuer, block and reissue.

If your country’s AML law makes you a reporting entity (in Nigeria, a financial institution or designated non-financial business and profession, MLPPA s.7) and your platform was used to turn stolen cards into value that was moved on, review it for a suspicious transaction report. An alert is a reason to review, not a suspicion in itself: the NFIU’s 2024 STR guidelines call monitoring alerts “unusual” and say they “should not be automatically regarded as suspicious” (Part Three, paras 1–2). Under the FATF standard, suspicious transactions are reported regardless of amount, attempted transactions included (Interpretive Note to Recommendation 20, para 3). Each country’s reporting deadline is under By country.

  • Card testing: checking which stolen cards work; card cycling is often what comes next.
  • Shared identifiers across customers: one card paying several customer accounts, the reverse direction of this rule.
  • Transaction velocity: too many transactions in a window, whatever the card.
  • New beneficiary, high-value payment: a large first payment to a beneficiary, wallet or card the customer has never paid before.
  • New account activity: large sums or many payments through an account in its first days or weeks; check it when the cards sit on a newly opened account.
  • Account takeover: someone other than the customer controls the account and pays out or drains it; check it when new cards follow a new device or a changed email or phone.

Sources

13 sourcesHide
  1. Visa, Enumeration Mitigation and Considerations for Merchants, V1.3, June 2026, section “Protecting the Ecosystem”: “Account Creation” (“Limit the number of cards that can be added per ‘account’ and session”; “Monitor the frequency of payment method changes on accounts”); “Anomaly detection” (“multiple transactions with different payment accounts using the same email address and same device ID, may be a trigger for fraud classification or review”); “Token” (“Implement intelligent rate limiting and velocity controls at the tokenization layer, specifically during the ‘Add Card’ or ‘Save Card’ steps”; “Analyze decline codes (CVV failed, Do Not Honor, Suspected Fraud, etc) aggregated across BIN ranges”, p.9); “Technical Tools” (“example five authorizations from one IP address or PAN within a set timeframe”, p.7); “Velocity thresholds” (“Thresholds should also be set on the number of transactions within a specified timeframe”, p.7). Supports the pattern. https://usa.visa.com/content/dam/VCOM/global/support-legal/documents/anti-enumeration-and-account-testing-best-practices-merchant.pdf
  2. FATF, Money Laundering Using New Payment Methods, October 2010: para 104 (p.33), “systems in place to detect if a customer holds multiple cards or accounts, which can be an indicator for a customer circumventing the CDD procedures by structuring the funds into several ‘low risk’ products”; para 109 (p.34), monitoring systems “must at a minimum allow the provider to identify … Cases where several products are funded by the same source”; para 110 (p.34), “Where products benefit from customer due diligence exemptions, systems should detect where a customer approaches a limit (on one product/transaction or cumulatively) beyond which full customer due diligence has to be applied”; Case 21 (p.42), 61 stolen credit card numbers used to fund virtual prepaid cards. Supports the pattern. https://www.fatf-gafi.org/en/publications/Methodsandtrends/Moneylaunderingusingnewpaymentmethods.html
  3. FATF, Guidance for a Risk-Based Approach: Prepaid Cards, Mobile Payments and Internet-Based Payment Services, June 2013, para 82 (p.25): “The transaction monitoring system could also be used to identify multiple accounts or products held by an individual or group, such as holding multiple prepaid cards”; para 101 (p.29): “Without sufficient CDD, however, acquiring prepaid cards via the Internet could allow for multiple cards to be acquired by the same person using different names.” Supports the pattern. https://www.fatf-gafi.org/content/dam/fatf-gafi/guidance/Guidance-RBA-NPPS.pdf
  4. EBA and ECB, 2025 Report on Payment Fraud (EBA/REP/2025/40), December 2025, Executive Summary, p.6, on “card payments (from an EU/EEA issuing perspective)”: “In most card payment fraud transactions, fraudsters directly issued the payment order, often using stolen credentials for remote card fraud”. Context. https://www.ecb.europa.eu/press/intro/publications/pdf/ecb.ebaecb202512.en.pdf
  5. PCI Security Standards Council, FAQ article 1374, “Is Payment Account Reference (PAR) as defined by EMVCo considered PCI Account Data?” (on EMVCo Specification Bulletin No. 167, January 2016): “PAR is a value that is intended to allow acquirers and merchants to link tokenized transactions to transactions that are based on the underlying PAN. PAR is generated and linked to a PAN (and successor PANs associated with the underlying issuer customer account) and will also be associated with all affiliated Payment Tokens”; “PAR data is not considered to be PCI Account Data”; “PCI DSS still applies anywhere PCI Account Data is stored, processed, or transmitted”. Supports the tuning advice on card keys. https://www.pcisecuritystandards.org/faq/articles/Frequently_Asked_Question/is-payment-account-reference-par-as-defined-by-emvco-considered-pci-account-data
  6. NFIU (Nigeria), Money Laundering Typologies Through Fraud in Nigeria, December 2023, chapter 6 “Analysis of questionnaire responses”: payment service providers, p.75; casino operators, p.82. Survey responses. https://www.nfiu.gov.ng/Trendstypo
  7. GIABA, Typologies of Money Laundering and Terrorist Financing Linked to Cybercrime in West Africa, May 2025, Typology 1 “Electronic Cards (credit/debit) Fraud”: Case 2 (Ghana), pp.14–15; Case Study 3 (Côte d’Ivoire), p.15. https://www.giaba.org/typologies/3495__typologies-of-money-laundering-and-terrorist-financing-linked-to-cybercrime-in-west-africa.html
  8. NFIU, Guidelines for the Identification, Verification and Reporting of Suspicious Transactions Related to ML/FT/PF for Financial Institutions, STR-NFIU-2024-A0001, 13 December 2024, Part Three paras 1–2 and Part Six para 2. Supports the reporting duty. https://www.nfiu.gov.ng/AdvisoryAndGuidance
  9. Money Laundering (Prevention and Prohibition) Act 2022 (Nigeria), s.7(1)–(3). Supports the reporting duty. https://www.nfiu.gov.ng/LawsAndRegulations
  10. UMOA, Loi uniforme relative à la lutte contre le blanchiment de capitaux, le financement du terrorisme et de la prolifération des armes de destruction massive dans les États membres de l’Union Monétaire Ouest Africaine (UMOA) (March 2023), arts. 2 (point 58), 60, 61 and 65. Supports the reporting duty. https://www.bceao.int/sites/default/files/2025-02/Loi-uniforme-LBCFTFP-Mars-2023-vf.pdf
  11. FATF, The FATF Recommendations, October 2025 edition (R.20 was not changed by the June 2026 update), Recommendation 20 and Interpretive Note para 3. https://www.fatf-gafi.org/en/publications/Fatfrecommendations/Fatf-recommendations.html
  12. Central Bank of Nigeria, Baseline Standards for Automated Anti-Money Laundering, Combating the Financing of Terrorism and Countering Proliferation Financing Solutions, BSD/DIR/PUB/LAB/019/002, 10 March 2026, section 5.6(a)(i). Supports the Nigeria entry. https://www.cbn.gov.ng/Out/2026/CCD/CBN%20issues%20Baseline%20Standards%20for%20Automated%20Anti-Money%20Laundering%20Solution.pdf
  13. Côte d’Ivoire, Communiqué du Conseil des Ministres du jeudi 23 novembre 2023: “Cette ordonnance introduit dans la législation nationale la Loi uniforme … adoptée par le Conseil des Ministres de l’UMOA en mars 2023”. Supports the Côte d’Ivoire entry and the reporting duty there. https://www.gouv.ci/uploads/publications/1700773380Communique-du-Conseil-des-Ministres-du-jeudi-23-novembre-2023.pdf

Ask us

A question about this rule?

A person at Archer reads every question and answers by email.

The Fraud & AML Library is a practical reference, not legal advice. What binds is the regulator’s own text, which every rule and regulation page links.

Get new rules and regulation changes by email

When we add a rule or a regulation changes, we’ll email you.

You can unsubscribe at any time. Privacy policy