What the ratio is, where the limits sit, and when fighting a dispute is worth the cost.
BNPL marketing tells the merchant the provider absorbs chargebacks. For consumer-dispute reason codes, the chargeback comes through the card network behind the BNPL and lands the same as any other.
Read →The headline 2026 statistic — first-party fraud rose from 15% to 36% of all reported fraud — reflects better detection by acquirers, not worse behaviour by cardholders.
Read →While the industry panics about AI on the cardholder side, the bigger 2026 shift is happening inside the issuing banks themselves.
Read →Amex is structurally tilted toward the cardmember, and the merchant's instinct to refund earlier on small Amex disputes is the correct one.
Read →Two identical 13.1 disputes can produce opposite outcomes because Chase, Bank of America, and Capital One weigh the same evidence differently.
Read →The line on the cardholder's bill names you to the buyer; mismatch it and you create unrecognised-charge disputes for free.
Read →3DS is a partial shield with specific shapes. The carve-outs cluster around what counts as a real authentication and around reason codes that 3DS never covered at all.
Read →Volume of AI-generated content in disputes has roughly doubled in a year. Issuers still review the cases by hand, and they have learned to dismiss AI letters faster than the templates they replaced.
Read →The dispute is a he-said/she-said. Whichever side wins, the underlying problem is the same, and it sits with the merchant.
Read →At a fully-loaded $50 an hour and 30 minutes per careful representment, the in-house cross-over sits around 60 disputes a month. Most growing merchants pass it without realising.
Read →Card-absent fraud carries the highest representment cost and the lowest naive win rate. CE 3.0 is the only framework that helps, and it helps a narrow slice of merchants.
Read →Auto-submit platforms, AI-rebuttal vendors, and processor dashboards all send a document the merchant never reads. A merchant who cannot see the letter cannot improve the letter.
Read →A Mastercard 4853 is not a Visa 13.1 with a different number. The rules differ on timing, on reason-code structure, and on workflow, and merchants who do not notice lose cases they should have won.
Read →One tired analyst, a structured scoring sheet, a queue of similar cases behind. The letter that wins is written for that workflow.
Read →Visa 13.3 turns on whether what the merchant promised matched what the customer ordered. Photo evidence of the delivered item is rarely as useful as the product description on the page where the customer clicked buy.
Read →The issuer's question is not whether the digital goods were delivered. It is whether the cardholder used them after paying.
Read →Marketplaces take the customer relationship and the dispute interface. They leave the chargeback liability with the seller. The structure is deliberate.
Read →There is no tracking number on a haircut. The records that win services chargebacks are operational, not logistical, and most merchants do not capture them.
Read →Subscription chargebacks cluster in two reason codes that are nearly always refund-path failures, and neither pure automation nor pure letters wins them alone.
Read →Physical-goods cases are won in the order log, before the chargeback arrives.
Read →Fighting every chargeback loses money on cases that should have been refunded.
Read →The records that win a subscription dispute were captured at sign-up, or never.
Read →Pre-arbitration is the issuer's second move, with fresh evidence and a network fee for the loser.
Read →Two of Visa's April changes will actually reach a small merchant's desk.
Read →The dashboard hides roughly half the cost of a lost chargeback.
Read →The acquirer's response window is a third of the network's.
Read →Optimising for chargeback ratio rewards refunding winners and fighting losers.
Read →A tracking number proves a parcel reached an address, not the right person.
Read →The auto-submit default loses cases the merchant could have won.
Read →The case is decided by what the merchant captures on day one.
Read →Every reason code wins on the same letter arc.
Read →Most cardholders disputing their own purchases are picking the faster refund.
Read →Five fields inside the notification decide the case.
Read →Paying $19 to deflect a chargeback is the cost of missing order records.
Read →Order-time records decide the chargeback that arrives six weeks later.
Read →Visa's monitoring threshold reads as uniform but behaves as a step function.
Read →One year in, VAMP enforcement looks graduated rather than absolute.
Read →