Chargeback Reason Codes: Which Disputes to Fight
Chargeback Reason Codes: Which Disputes to Fight
Every chargeback carries a reason code, and the code predicts whether the merchant can win. In a stylized mix, fighting every dispute nets $571 per 100; fighting only the categories with positive expected value nets $764, from half the work. And with Visa's Compelling Evidence 3.0 data, card-absent fraud, usually the largest losing category, becomes the largest source of recovery, lifting the net to $1,494.
What you'll learn
What a reason code tells you
Every chargeback arrives with a reason code: the card network's label for why the cardholder's bank says the money should come back. The code decides which evidence the merchant must supply, how long it has to respond, and how often merchants win. Treating all chargebacks alike, either fighting every one or accepting every one, ignores the most useful piece of information in the dispute.
| Visa category | Example codes | What the cardholder's bank is saying | Mastercard counterpart |
|---|---|---|---|
| 10.x Fraud | 10.4 Other Fraud, Card-Absent Environment | The cardholder did not authorize the purchase | 4837 No Cardholder Authorization; 4863 Cardholder Does Not Recognize |
| 11.x Authorization | 11.2 Declined Authorization; 11.3 No Authorization | The transaction was not properly authorized at the time | 4808 Authorization-Related |
| 12.x Processing Errors | 12.6 Duplicate Processing or Paid by Other Means | The charge was processed wrongly: twice, late, or for the wrong amount | 4834 Point-of-Interaction Error |
| 13.x Consumer Disputes | 13.1 Not Received; 13.2 Cancelled Recurring; 13.3 Not as Described | The cardholder paid but did not get what was promised | 4853 Cardholder Dispute |
The categories differ in what wins. A fraud dispute is won by showing the genuine cardholder made the purchase. A not-received dispute is won with delivery proof. A not-as-described dispute is won with the product description and the customer's own communications. A processing error is usually lost because it is usually true. Per our dispute handling analysis, the response that wins one category is irrelevant in another, so the code should route the case before anyone writes a word.
The arithmetic of fighting
Take a merchant with an average ticket of $85 that spends $20 of staff time and tooling assembling each response. Fighting a dispute is worth it when the win rate times the ticket exceeds the cost: $20 ÷ $85 is a break-even win rate of 23.5%. Win rates vary sharply by category.
| Category | Share of disputes | Win rate | Expected value of fighting |
|---|---|---|---|
| Fraud, card absent | 40% | 20% | −$3.00 |
| Not received | 20% | 55% | +$26.75 |
| Not as described | 15% | 35% | +$9.75 |
| Processing error | 10% | 15% | −$7.25 |
| Cancelled recurring | 15% | 30% | +$5.50 |
Fighting everything spends $2,000 to recover $2,571. Fighting only the not-received, not-as-described and cancelled-recurring cases spends $1,000 to recover $1,764. The fraud and processing cases, half the volume, lose money on average at these win rates. That does not mean fraud disputes should always be conceded; it means they should be conceded unless the merchant has evidence that changes the win rate.
How better evidence changes the fraud column
Card-absent fraud disputes are the largest category and, for many merchants, the most frustrating, because a large share are not stolen-card fraud at all. Per our friendly fraud analysis, a cardholder who made a purchase and later disputes it as unauthorized files under the same fraud code as a genuine victim.
Visa's Compelling Evidence 3.0 rules, applicable to reason code 10.4, let a merchant show that the disputed transaction matches at least two earlier undisputed transactions on the same card, made 120 to 365 days before the dispute, on shared data elements such as IP address, device identifier, shipping address or account login, with at least one of the matches on IP address or device identifier. When the evidence qualifies, the liability shifts back to the issuer.
| Fraud disputes | Win rate | Expected value per case | Net recovery per 100 disputes (all categories) |
|---|---|---|---|
| Conceded | n/a | $0 | $764 |
| Fought with ordinary evidence | 20% | −$3.00 | $644 |
| Fought with qualifying Compelling Evidence 3.0 data | 45% | +$18.25 | $1,494 |
The policy did not change the merchant's customers; it changed what the merchant could prove. Lifting the fraud win rate from 20% to 45% turns the largest category from a loss into the biggest source of recovery, and nearly doubles the net. The precondition is data captured at checkout and retained: device identifiers, IP addresses and account logins stored long enough to match a dispute a year later. A merchant that does not log them cannot use the rule.
Codes that point to the cause
Reason codes are also a diagnostic. A rising share in one category names a specific operational failure, and the cheapest chargeback is the one the merchant prevents.
- Fraud codes with no prior relationship: stolen cards getting through. Per our 3-D Secure analysis, authentication on risky transactions shifts fraud liability to the issuer, and per our card testing analysis, bursts of small authorizations often precede a fraud wave.
- Fraud codes from repeat customers: customers not recognizing the charge. Per our billing descriptor analysis, an unclear statement descriptor is a common cause of disputes that are really confusion.
- Not received: fulfillment or tracking gaps. Signature on high-value shipments and proactive delivery notices are the fix.
- Cancelled recurring: cancellation that is hard to find or not honored. Per our subscription cancellation analysis, a difficult cancellation path converts cancellations into chargebacks.
- Not as described: product pages that oversell, or a refund process customers cannot reach. Per our refund analysis, a refund costs less than a chargeback, its fee and its ratio impact.
- Processing errors: duplicate captures, late presentment or amount mismatches. These are almost always lost and almost always fixable in the payment integration.
Read the code before you fight the case
HL Hunt AI Payment Processing classifies each dispute by reason code, scores whether the available evidence is likely to win, and assembles the response for the cases worth fighting, so staff time goes where the expected value is positive.
Why the dispute count matters beyond the money
Card networks monitor each merchant's dispute and fraud levels, and merchants above program thresholds face fines, mandatory remediation and, ultimately, loss of the ability to accept cards. Winning a chargeback recovers the money but, under most network rules, the dispute still counted when it was filed. That changes the calculation: prevention lowers the count, representment does not.
Pre-dispute tools sit between the two. Per our dispute alerts analysis, issuer alert and resolution services let a merchant refund a disputed transaction before it becomes a chargeback, at a per-alert fee. For categories with negative expected value, such as processing errors, refunding through an alert is usually better than either fighting or absorbing a formal chargeback, because it costs the ticket without the chargeback fee or the ratio impact. Per our chargeback prevention analysis, the merchants with the lowest ratios combine all three: prevention for the causes they control, alerts for the cases they would lose, and representment for the cases they can win.
The consumer rights underneath
Chargebacks are a network mechanism, but they rest on legal rights. The Fair Credit Billing Act, implemented in Regulation Z, gives credit cardholders the right to dispute billing errors, including goods not delivered as agreed, and limits their liability for unauthorized use to $50. Regulation E gives debit cardholders error-resolution rights for unauthorized transfers. Card networks typically extend zero-liability policies further than the statute requires.
That is why the burden falls on the merchant. The cardholder's bank must investigate a timely dispute, and the network rules give it a simple path to recover the funds from the merchant's acquirer unless the merchant supplies evidence. A merchant that understands this stops treating representment as an argument and starts treating it as a proof exercise matched to the reason code.
The case against selective fighting
- Conceding a category invites more of it. If customers learn that disputes are never contested, friendly fraud rises. Fighting some negative-expected-value cases can deter future ones.
- Win rates are not fixed. A merchant that never fights fraud cases never learns how to win them; some disputes should be fought to build the evidence process.
- The averages hide variation. A $600 fraud dispute has a break-even win rate of 3.3%, not 23.5%. Ticket size belongs in the rule, not just category.
Each objection refines the rule rather than defeating it. The decision should use the expected value of the individual case, combining reason code, ticket size and available evidence, with a deliberate budget for deterrence and learning. What it should not do is fight or concede everything by habit.
Fight the disputes you can win, prevent the rest
HL Hunt AI Payment Processing tracks dispute rates by reason code, flags the operational cause behind each rising category, and helps keep your ratio well below network monitoring thresholds.
Frequently asked questions
The network's label for why a transaction is reversed. Visa uses 10.x fraud, 11.x authorization, 12.x processing and 13.x consumer disputes.
Those where win rate times ticket beats response cost. Here, fighting only positive-value categories nets $764 per 100 versus $571 for fighting all.
A Visa rule letting merchants beat code 10.4 fraud disputes by matching the transaction to two prior undisputed ones on device, IP or account data.
Generally not. Disputes usually count when filed. Prevention and pre-dispute refunds lower the ratio; winning only recovers the money.
Key takeaways
- The reason code determines what evidence wins and how often merchants win.
- Fighting is worth it when win rate times ticket exceeds response cost: $20 on $85 needs 23.5%.
- In the example, fighting only positive-value categories nets $764 per 100 versus $571 for fighting all.
- Visa Compelling Evidence 3.0 lifts card-absent fraud win rates for merchants that retain device, IP and login data.
- Rising codes in one category point to a specific cause: descriptors, fulfillment, cancellation or integration errors.
- Winning a chargeback recovers money but generally does not remove it from the dispute ratio.
- Use per-case expected value, including ticket size, with a deliberate budget for deterrence.
References
- Visa. Visa Core Rules and Visa Product and Service Rules, Dispute Resolution chapter. Visa Inc.
- Visa. Visa Merchant Dispute Management Guidelines. Visa Inc.
- Visa. Compelling Evidence 3.0 for Dispute Condition 10.4, Visa Business News. Visa Inc.
- Mastercard. Chargeback Guide: Merchant Edition. Mastercard International.
- Fair Credit Billing Act, 15 U.S.C. §§ 1666–1666j; liability for unauthorized use, 15 U.S.C. § 1643.
- Regulation Z, Billing Error Resolution, 12 C.F.R. § 1026.13.
- Regulation E, Procedures for Resolving Errors, 12 C.F.R. § 1005.11.
This guide is educational and does not constitute legal or payments advice. Reason code names, evidence requirements, timeframes and monitoring thresholds are set by the card networks and change periodically; consult current network rules and your acquirer. Dispute shares, win rates, ticket sizes and response costs are stylized illustrations and do not describe any particular merchant.