Chargeback Reason Codes Explained

When a customer disputes a transaction, their bank does not just send you a refund request. It sends a reason code — a short alphanumeric tag that classifies why the charge is being reversed. That code is the single most important piece of information you receive, because it dictates what evidence you need and whether the dispute is even worth fighting.

The problem is that each card network uses its own numbering scheme, and the codes read like airline fare classes. A "10.4" from Visa and a "4837" from Mastercard describe nearly the same event. Once you learn the underlying families, the noise clears up fast.

The four reason-code families

Nearly every chargeback falls into one of four buckets, regardless of the network:

• Fraud — the cardholder claims they did not authorize the purchase.
• Authorization — something went wrong with the approval (declined card charged anyway, expired authorization).
• Processing errors — duplicate charge, wrong amount, wrong currency, or a late presentment.
• Consumer disputes — "I never got it," "it arrived broken," "I cancelled and was still billed."

Roughly 60-70% of merchant chargebacks land in the fraud family, which is exactly where prevention pays off more than fighting after the fact.

Visa reason codes

Visa reorganized its codes under the Visa Claims Resolution framework into four dispute categories:

• 10.x — Fraud. The heavy hitters. 10.4 (card-absent fraud) is the most common e-commerce dispute: the cardholder says they did not make the online purchase. 10.1 covers EMV counterfeit fraud.
• 11.x — Authorization. For example, 11.3 means a transaction was completed after authorization was declined.
• 12.x — Processing errors. 12.5 is an incorrect amount, 12.6 is a duplicate processing.
• 13.x — Consumer disputes. 13.1 is "merchandise not received," 13.3 is "not as described or defective," 13.6 is "credit not processed."

Mastercard reason codes

Mastercard keeps a four-digit scheme:

• 4837 — No cardholder authorization (their fraud equivalent of Visa 10.4).
• 4863 — Cardholder does not recognize the transaction, often a soft fraud claim tied to an unclear billing descriptor.
• 4853 — Cardholder dispute (not received, defective, cancelled recurring, credit not processed).
• 4834 — Point-of-interaction error (duplicate or incorrect amount).

The lesson hiding inside 4863: a vague or unfamiliar statement descriptor generates disputes on its own. If your descriptor reads SP* LLC-8821 instead of your store name, customers panic and call the bank. Fixing your descriptor can cut "does not recognize" disputes measurably.

Amex and Discover

Amex and Discover act as both network and issuer, so their codes are simpler but follow the same logic:

• Amex F29 — card not present fraud. Amex C08 — goods not received. Amex C31 — goods not as described.
• Discover UA02 — no cardholder authorization. Discover RG — non-receipt of goods.

What the code tells you to do

Your response strategy flows directly from the family:

Fraud codes (10.4, 4837, F29, UA02). These are usually the hardest to win because the cardholder is claiming they were not involved at all. To have a chance you need proof the *legitimate* cardholder made the purchase — AVS match, CVV match, device fingerprint, IP location near the billing address, and ideally prior order history from the same account. If none of that lines up, the transaction was probably genuine fraud and fighting it wastes time. Prevention is the real answer here.

Consumer dispute codes (13.1, 4853, C08). These are the most winnable. Ship tracking with delivery confirmation, signed proof of delivery for high-value orders, and clear communication logs usually carry the day. "Merchandise not received" collapses the moment you produce a carrier-scanned "delivered" event at the customer's address.

Processing errors (12.x, 4834). Often your own back-office slip. Check for duplicate captures and refunds you forgot to issue before you spend energy disputing.

Authorization codes (11.x, 4859). Verify your gateway settings; charging after a decline is a policy violation you cannot win.

Where prevention beats representment

Here is the uncomfortable math. Even a strong representment win rate hovers around 30-40% for card-absent fraud disputes, and each dispute costs a non-refundable fee of roughly $15-$25 plus staff time. A store processing 4,000 orders a month with a 0.8% dispute rate is fielding ~32 disputes monthly — and winning maybe a dozen.

The cheaper lever is stopping the fraudulent order before it becomes a 10.4. That means screening the checkout for mismatched billing/shipping geography, VPN and proxy usage, disposable emails, and velocity spikes from a single device.

This is where a checkout-level tool earns its keep. Shieldy Fraud Filter scores orders in real time and can block high-risk IPs, countries, VPN, proxy, and Tor traffic before a fraudulent order is ever placed — so it never becomes a 10.4 or 4837 in the first place. Its AI fraud scoring flags the risky mismatches that quietly turn into "no cardholder authorization" disputes weeks later.

A simple triage workflow

When a chargeback lands, run this loop:

1. Read the reason code and map it to a family.
2. Pull the transaction record — AVS, CVV, IP, device, tracking.
3. Decide fight or fold. Fold on genuine fraud; fight winnable consumer disputes with evidence.
4. Log the pattern. If the same code keeps appearing, fix the root cause — descriptor, shipping process, or fraud screening.

Reason codes are not just paperwork; they are diagnostic signals. A spike in 10.4 means your fraud filters have a hole. A spike in 13.1 means your fulfillment or tracking needs work.

Want fewer fraud codes to fight in the first place? Start screening at checkout — the free plan is enough to see the difference.