Friendly fraud: the chargeback no filter can stop

What friendly fraud is

Friendly fraud — also called first-party fraud or first-party misuse — is a chargeback filed by a real cardholder against a transaction they genuinely made. Nobody stole the card. The customer (or someone in their household) bought the thing, then disputed the charge with their bank instead of contacting you.

It lives on a spectrum:

• Honest confusion. The cardholder doesn't recognize an unfamiliar billing descriptor on their statement, or forgot they were on a recurring plan — "subscription amnesia" — and assumes the renewal is fraud.
• Convenience disputes. They wanted to cancel or get a refund, found it easier to call the bank than to deal with you, and the bank obliged.
• Deliberate abuse. They knew the charge was valid and disputed it anyway to keep the product or service for free.

For a subscription business this is the chargeback that hurts most, because — unlike a stolen card — there is no fraudulent signal at checkout to detect. The transaction was authorized by the legitimate cardholder. That single fact shapes everything that follows. It's also not a small slice: card networks attribute a large share of card fraud to chargeback misuse, and it's the fastest-growing category merchants report (see the sourced figures in chargeback statistics).

Three kinds of chargeback — and why the difference matters

Before you can act on a dispute you have to classify it, because the response is completely different for each. Every chargeback an issuer raises carries a reason code that falls into one of three broad buckets:

Category	What the cardholder claims	Your move
True (third-party) fraud	"I didn't make this — my card was used without permission."	Usually accept it; fix the hole at authorization (Radar rules, 3DS). Fighting a genuine stolen-card case rarely wins.
Processing error	Duplicate charge, wrong amount, currency, or a credit not processed.	Verify and refund if real; contest with the transaction record if not. Usually clear-cut.
Friendly fraud	"I don't recognize this / I cancelled / I never got it" — on a charge they actually made.	The judgment call. Contest with compelling evidence, or refund/deflect — covered below.

The trap: friendly fraud frequently arrives disguised as true fraud. A cardholder who forgot a subscription often tells their bank "I didn't authorize this," which lands as a fraud reason code even though they did authorize it. That mislabelling is exactly what the right evidence can unwind.

Why it's so hard to stop

Fraud prevention — Stripe Radar rules, velocity checks, 3-D Secure — all operates before authorization. It scores the transaction and decides whether to let it through. Friendly fraud happens weeks after a transaction the system correctly let through, when a real customer changes their mind or forgets. No model can predict a future dispute on a legitimately authorized charge, because at the moment of payment there is nothing wrong.

This is the single most important thing to internalise: friendly fraud is a post-purchase, behavioural problem, not a checkout problem. You don't solve it by tightening your fraud filters. You solve it downstream — by removing the reasons people dispute, by intercepting disputes before they post, and by selectively contesting the ones worth fighting.

Representment: what it is, and the uncomfortable truth

Representment is the formal process of contesting a chargeback — literally "re-presenting" the transaction to the issuing bank with evidence that it was legitimate. You compile a rebuttal, submit it through your processor, and the issuer either reverses the chargeback (you win) or upholds it. If it's upheld you can sometimes escalate to pre-arbitration and arbitration, but each step costs fees and time.

Why so low? Because most merchants fight blindly — submitting the same generic evidence packet regardless of the reason code, or fighting low-value disputes where the staff time exceeds the recovery. Win rates rise sharply when you fight selectively and match the evidence to the specific claim.

Fight or refund? The decision

Treat every disputable chargeback as an expected-value calculation, not a point of pride. Roughly:

• Fight when the disputed amount is meaningful, your authentication and usage evidence is strong, and the reason code is genuinely contestable (e.g. a "didn't authorize" claim on an account with a long, 3DS-authenticated, actively-used history).
• Refund or deflect the low-value, weak-evidence, "I forgot I subscribed" cases. At an average subscription chargeback of roughly $69 and an ~18% win rate, the expected recovery on a contested low-value dispute is often a few dollars — less than the labour to fight it, and it still counts against your ratio while it's open.

How to actually win a representment

When you do fight, win rate is a function of evidence quality and fit. The evidence that moves issuers proves the real cardholder both made and benefited from the purchase:

• Authentication: AVS and CVV match, and especially 3-D Secure — a 3DS-authenticated transaction shifts fraud liability to the issuer, which often ends a "didn't authorize" claim outright.
• Identity continuity: device fingerprint and IP matching the customer's prior legitimate sessions; the same login used before and after the disputed charge.
• Proof of benefit: usage and access logs showing they logged in and used the product after paying; download/streaming/feature-access records.
• Consent: a timestamped record of accepted terms of service or the recurring-billing agreement (clickwrap), and the cancellation policy they agreed to.
• Transaction history: prior undisputed charges on the same card. Visa's Compelling Evidence 3.0 framework was built precisely for this — it lets merchants use a pattern of prior, undisputed transactions (same card, device, IP, account) to rebut a first-party fraud claim.

The discipline that lifts win rates: assemble evidence to the exact reason code. A "fraud — card not authorized" code needs authentication and identity continuity. A "services not provided" or "not as described" code needs proof of access, delivery, or the terms they agreed to. A generic packet stapled to every case is why the average win rate sits at 18%.

The better play: prevent and intercept

Because friendly fraud can't be blocked at checkout and is expensive to fight after, the leverage is upstream and midstream — removing the reasons people dispute, and catching disputes before they harden into chargebacks.

Prevent (remove the reasons)

• A billing descriptor people recognise. "SP* YOURBRAND APP" on a statement is the #1 driver of "I don't recognise this." Make it your brand name plus a recognisable tag.
• Pre-renewal reminders. An email before each renewal kills "subscription amnesia" before it becomes a dispute.
• An easy self-serve refund/cancel path. A one-click "refund my last payment" or cancel flow deflects a large share of would-be chargebacks — customers reach for the bank only when you make leaving hard.
• Clear, representment-ready terms. Plain renewal language, agreed at sign-up and logged, that also doubles as evidence later.

Intercept (catch it before it posts)

Ethoca (Mastercard) and Verifi / CDRN (Visa) are dispute-interception networks. When a cardholder opens a dispute, they alert you before it becomes an official chargeback, giving you a window to issue a proactive refund — which keeps the dispute off your chargeback ratio entirely. For a subscription app near a monitoring threshold, this single layer typically does more for the ratio than any representment program. The trade-offs between the two networks are in Ethoca vs Verifi.

Frequently asked questions

What is friendly fraud?

A chargeback filed by a genuine cardholder against a transaction they actually made — from honest confusion (an unrecognised descriptor, a forgotten renewal) to deliberate abuse (disputing a valid charge to get it free). Because the purchase was authorized, no checkout fraud filter can block it, which makes it the dominant chargeback category for subscription apps. More in the glossary.

What is chargeback representment?

Contesting a chargeback by re-presenting the transaction to the issuer with evidence it was legitimate. The issuer then reverses or upholds it. Merchants win only ~18% of the chargebacks they fight and recover ~20% of the value (statistics), so it's a selective tool.

Should I fight a chargeback or just refund it?

An economics decision. Fight when the amount is high, the authentication/usage evidence is strong, and the code is contestable. Refund or deflect low-value, weak-evidence "I forgot" cases — at ~$69 average and ~18% win odds, fighting often costs more than it recovers, and an open dispute still hits your ratio.

What evidence wins a representment?

Proof the real cardholder made and used the purchase: AVS/CVV match, 3-D Secure authentication (shifts liability to the issuer), device/IP continuity with prior sessions, usage/login logs, accepted terms, and prior undisputed transactions on the same card (the basis of Visa's Compelling Evidence 3.0). Assemble it to the exact reason code cited.

How do you prevent friendly fraud?

You can't block it at authorization, so prevent it structurally (recognisable billing descriptor, pre-renewal reminders, easy self-serve refunds, clear terms) and intercept it with Ethoca and Verifi before disputes become chargebacks. Near a VAMP/ECM threshold, that layer beats chasing representment wins.