Skip to main content

Fraud Types

TL;DR

Work out which of these you actually have before you read further.

The same customer keeps disputing? Match the payment before reviewing future orders.


Which Fraud Type Is YOUR Problem?​

If you see...Start here
Chargebacks on orders you know the customer madeCheck the claim first. Friendly Fraud if they knowingly disputed without a valid reason
Many small transactions, then fraudCard Testing
Good customer suddenly acting strangeAccount Takeover
New account, immediate high spendingThird-Party Fraud or Account Fraud
Coordinated attack across accountsFraud Rings
Return/refund abuseRefund Fraud
Promotion/coupon exploitationPromo Abuse
A partner's sales, leads or views look fakeAffiliate Fraud
ACH returns or unauthorized bank debitsACH Fraud
Customer sent a payment, then said they were scammedACH Fraud
A seller or customer account exists to receive and forward moneySecond-party fraud

How to Find Out What Fraud Type You Have​

Don't buy a tool yet. Find out what you've actually got:

  1. Pull your last 30 chargebacks (or however many you have from the last 6 months)
  2. Tag each one with the bucket that fits best. The claim on a dispute isn't proof, so note any overlaps:
    • Third-party fraud: Stolen card, customer says "I didn't do this" and the evidence supports them
    • Service issue: Customer had a real problem with the product or delivery
    • Billing confusion: Customer didn't recognize the charge on their statement
    • Friendly fraud: Customer made the purchase, got what they ordered, and knowingly disputes it anyway
    • Unknown: You can't tell yet. Keep this pile
  3. Count the cases and the money, on cases old enough to have settled. A small sample can point you somewhere, but its mix may not represent all your disputes.
Check Your Own Dispute Mix

If your own review shows mainly friendly fraud and billing confusion, start with the operational fixes. Better descriptors (how to fix them), easier refunds, clearer emails. Technical controls such as device fingerprinting can still help you collect evidence on repeat abuse, so choose them for a confirmed cause. Read Friendly Fraud first, then Descriptors and Comms, and buy nothing until you have.


Classification Framework​

Sorted by who commits it:

TypeActorKey Characteristic
First-PartyYour customerOwn identity, and they're lying
Second-partyYour customer, for somebody elseThey handed their card or account over. The money goes elsewhere
Victim-initiatedYour customerThey sent it themselves. Somebody deceived them into it
Third-PartyExternal fraudsterNever given access. It was taken
Fake IdentitySomeone behind a made-up personaNo such person or business. The details can be real, invented or combined, so real people can be victims

Second-party and victim-initiated look alike and aren't. One handed over an account. The other sent a payment. Both may have been deceived, and the deception doesn't change which one it is. Ask what they gave away: access, or money?

A deceived customer in those two middle rows can be telling the truth and still have lost the money. A knowing second-party participant isn't.

On cards, a criminal with a stolen number does the initiating. On a push payment or an ACH credit (how that fraud works), a deceived customer can tap send themselves. That's where this row lives. An ACH debit is different, because someone else initiates it against the account. Record who initiated and authorized each entry (credits and debits).

How the Federal Reserve Classifies This​

The Fed publishes two classification models, free and open. They're worth knowing for one reason. Nobody's selling you anything with them.

FraudClassifier opens with a different question than this page does. Not "who is the fraudster" but "who initiated the payment?" Everything else hangs off that.

Who initiated itHowThen
Authorized PartyActed fraudulentlyEmbezzlement, False Claim, Synthetic ID
Authorized PartyWas manipulatedProducts and Services Fraud, Relationship and Trust Fraud
Unauthorized PartyTook over the accountCompromised credentials, impersonation
Unauthorized PartyMisused account informationDigital payment, physical forgery
Unauthorized PartyModified the paymentCompromised credentials, impersonation, physical alteration

It works across payment types. The Fed's model classifies a payment by who initiated it, the method and the tactic, whatever the payment type or channel. At its launch the Boston Fed described the problem as inconsistent classifications "for fraud involving ACH, wire, or check payments". That's its history, not a limit.

That's why the two schemes look like they disagree. Take a customer's card details from a data breach, then use the card yourself. You initiated the payment. Third-party here, Unauthorized Party there, no argument. Now phish a customer into sending the money. They initiated it. The Fed calls that Authorized Party Was Manipulated. Here it's the victim-initiated row above. But if they hand over their credentials and the criminal initiates the payment, the Fed's branch is Unauthorized Party, even though this site calls the handover second party.

One place the Fed splits from this page. It files Synthetic ID under Authorized Party Acted Fraudulently. That treats the fabricated persona as an account holder with rights. This page gives fake identity its own actor row instead. The persona isn't real, though somebody real runs it, and real people's details can be in it.

ScamClassifier is the companion. It sorts the scam itself into nine types. The Fed says it can run before or after FraudClassifier. Reach for it when you're describing what happened to a person, not to a payment.

A rough crosswalk. Run each actual payment through the Fed's own questions.

On this siteFraudClassifier
First-party, friendly fraudAuthorized Party Acted Fraudulently → False Claim
Refund fraud run by your own staffAuthorized Party Acted Fraudulently → Embezzlement, if they were entrusted with the funds
Fake identityAuthorized Party Acted Fraudulently → Synthetic ID
Victim-initiatedAuthorized Party Was Manipulated
Second-party, knowingAuthorized Party Acted Fraudulently
Second-party, deceived into itDepends who initiated the payment. Authorized Party Was Manipulated if the victim did, Unauthorized Party if a criminal used what they handed over
Third-party, card testingUnauthorized Party → Misused Account Information
Account takeoverUnauthorized Party → Took Over Account (compromised credentials or impersonation)

Quick Reference​

By Method​

Fraud TypeDescriptionWhen You See It
Account FraudFake account signupsBot attacks, promo farming
Account TakeoverHijacked customer accountsPassword breaches, phishing
ACH FraudUnauthorized bank debits, BECACH returns, payment redirects
Card TestingValidating stolen or guessed cardsBursts of small charges or card setups
Fraud RingsOrganized multi-account attacksCoordinated patterns
TriangulationThree-party resale schemeMarketplace fraud
BEC & PhishingAttacks on your operationsFake invoices, credential theft
Affiliate FraudPaid partners faking sales, leads or viewsOne partner's refunds or disputes well above the rest

First-Party Fraud Subtypes​

These are first-party fraud when the customer is lying. A customer who forgot a purchase, or used an offer the way its terms allow, isn't a fraudster:

Fraud TypeDescriptionYour Defense
Friendly FraudKnowingly dispute a real purchase without a valid reasonEvidence collection. CE 3.0 where the history qualifies
Refund FraudExploit return policiesPolicy enforcement
Promo AbuseGame promotions/discountsDevice linking, limits

Comparison at a Glance​

These are starting points. Who carries a disputed payment depends on its condition and your contract (Visa rules, 24 October 2026, Tables 11-29 and 11-32, pp. 706-716).

TypeWho Can LoseCan You Answer a Chargeback?A Place to Start
Third-PartyYou, unless an eligible 3DS sale or another defense moves itCheck the condition. Eligible 3DS is one defense3D Secure
First-PartyYouYes, with evidence for the claimPolicy enforcement, evidence collection
↳ Friendly FraudYouYes, with evidence. CE 3.0 where the Visa history qualifiesDescriptors, evidence, easy refunds
↳ Refund FraudYouIf a dispute follows, answer its claimPolicy enforcement, pattern tracking
↳ Promo AbuseYouN/ADevice linking, limits
Fake IdentityYou, and anyone whose real details were usedSometimesIdentity verification
ATOCustomer, you, or neither if no payment was madeCheck the claim and any credit already given. Respond if an authenticated-payment or other permitted defense applies (Mastercard 4837, Visa 10.4)A second login check (MFA), behavioral analytics
Card TestingYouA test charge that succeeded can be disputed. Answer it under its conditionVelocity rules, CAPTCHA
Fraud RingsYouSometimesDevice fingerprinting
Affiliate FraudYouN/ACommission terms, refunds and disputes tracked per partner

Prevention Priority​

Rank your own confirmed losses and control gaps from the buckets above. This order is only a place to start.

1. Worth Checking First​

TypeAction
Third-Party FraudEnable 3D Secure for liability shift on eligible sales
Card TestingAdd velocity rules and CAPTCHA
Account FraudRequire email/phone verification

2. Worth Checking Next​

TypeAction
Friendly FraudCollect evidence. Check CE 3.0 eligibility
Refund FraudTighten policies, track patterns
Account TakeoverRequire a second login check (MFA), monitor logins

3. Specialized Threats​

TypeAction
Fraud RingsDevice fingerprinting, fraud data shared across businesses (consortium data)
Promo AbuseDevice linking, redemption limits
TriangulationShipping address analysis

Something wrong on this page? Email matt at paymentsandrisk.com and say what needs fixing.