Representment is the moment your billing system gets cross-examined. A cardholder filed a dispute, the issuer pulled the funds, and now you have one chance to send the acquirer a file that proves the charge was valid. For a subscription business this is not an occasional event. It is a recurring operational load, because every billing cycle creates a fresh population of charges that some fraction of customers will dispute rather than cancel.
The general dispute system (who routes what to whom, and the arbitration ladder) is covered in the sibling course. Here we stay in the subscription seat: what evidence wins a recurring-charge representment, how Compelling Evidence 3.0 changes the math on fraud claims, and why the win rate your vendor reports is probably not the number that matters.
What a representment file actually is
Representment is your re-presentation of the transaction to the issuer with evidence that the original charge should stand. You are not arguing in front of a judge. You are submitting a structured packet that the issuer reviews against the dispute reason code, often with limited human attention and a tight clock.
The clock is the first constraint. Under Visa, the acquirer or merchant has 30 days to respond to common card-absent dispute conditions, including fraud (10.4) and consumer disputes such as 13.2. American Express compresses this further: the response window is around 20 days, and once the acquirer's notification lag and your internal review are subtracted, teams often have five days or fewer to assemble and submit. Build for the worst case, not the nominal one.
Match the evidence to the reason code
A subscription dispute almost always lands in one of two buckets, and they want different files.
The fraud claim (Visa 10.4, "Other Fraud: Card-Absent Environment") says the cardholder did not authorize the charge. Your file needs to prove the account holder is the person who has been using the service: AVS and CVV match results, IP address, device fingerprint, login records, and proof of cardholder communication.
The cancellation or "I did not get what I paid for" claim (Visa 13.2, Amex C28, Discover AP) says the customer canceled or the service stopped. Your file needs the terms and cancellation policy in effect at signup, the customer's acceptance of those terms, and access logs showing the service was used during the disputed period. The distinction that wins these cases is documentary proof that the customer canceled their service and not their card, or never canceled at all.
Compelling Evidence 3.0: the historical footprint
For fraud claims specifically, Visa's Compelling Evidence 3.0 (CE3.0) gives subscription merchants a stronger remedy than a generic evidence dump. It applies only to dispute condition 10.4, the card-absent fraud code, and it lets you establish a "historical footprint" of legitimate prior use.
The mechanism: you submit two prior undisputed transactions from the same cardholder that are between 120 and 365 days older than the disputed charge, neither previously reported as fraud. Each must share matching data elements with the disputed transaction (customer login ID, device ID or fingerprint, IP address, or shipping address), and under Visa's spec at least one of the two matching elements must be the IP address or the device ID (some processors frame this as a main/secondary element split, so confirm your acquirer's exact requirement). The billing descriptors also need to align, so the first six characters should be identical across the transactions.
When the packet qualifies, the issuer is generally required to accept liability and cannot simply push the case forward, which closes off the pre-arbitration counter-move that otherwise drags fraud disputes into a 30-day second round.
Why this favors subscriptions
CE3.0 was practically built for recurring billing. A monthly plan that has run for a year produces exactly the kind of repeat, same-device, same-IP history that satisfies the two-transaction rule. Instrument your system to capture login ID, device fingerprint, and IP on every successful charge and to retain them for at least 13 months, and a large share of your "fraud" disputes become near-automatic wins. Skip that instrumentation and you are stuck arguing the same cases on weaker generic evidence.
The gross-versus-net win-rate trap
Here is where most subscription dispute programs lie to themselves. Two different numbers both get called the "win rate," and they are far apart.
The gross win rate is successful representments divided by disputes you chose to fight. The net recovery rate is successful representments divided by total disputes filed against you, including the ones you never contested.
As of 2025, US merchants win roughly 54 percent of the chargebacks they fight, which is the gross number that vendors love to quote. But net recovery, measured against all disputes, averages around 12 percent, and some 2024 estimates put it closer to 8 percent. The gap is everything you let through without a fight.
Worked example
Say you take 1,000 subscription disputes in a quarter, each worth $40, so $40,000 is in play.
Your team contests 300 of them (the ones with clean evidence) and wins 162. That is a 54 percent gross win rate, and the dashboard turns green.
But the other 700 disputes were conceded automatically, often because the billing descriptor was unrecognizable, the evidence was not captured, or nobody triaged them in time. Of the full $40,000, you recovered 162 charges, or $6,480. Your net recovery rate in this scenario is 16 percent, already better than the industry's roughly 12 percent average but still less than a third of the 54 percent your dashboard shows. The 54 percent number described one slice of your behavior; the 16 percent described your business.
The trap is that you can raise the gross rate by fighting fewer, stronger cases, which makes the program look better while recovering less money. Manage to net recovery, and the levers change: capture CE3.0 data so more fraud cases qualify, fix descriptors so fewer disputes are conceivable, and triage every dispute rather than cherry-picking. In accounting terms the conceded disputes are not a line you can ignore. A chargeback is a reduction of revenue, usually booked to a contra-revenue account, so the 700 you waved through are revenue you reported and then gave back.
The takeaway
Representment for subscriptions is two jobs, not one. The first is building files that match the reason code, with CE3.0's two-transaction historical footprint as your strongest play on fraud claims, which means instrumenting login ID, device, and IP capture today and retaining it past 365 days. The second is measuring honestly: report net recovery against all disputes, not gross win rate on the cases you chose to fight, or you will optimize a number that goes up while your recovered dollars go down.