There's a particular kind of chargeback that stings more than the others. Not the fraud claim, not the package-never-arrived dispute — the one where a customer asks for a refund your policy clearly doesn't allow, gets told no, and then goes to their bank anyway. You pull up your terms page, point to the paragraph that says all sales final after 30 days, and assume the matter is settled. Weeks later, the bank sides with the cardholder.
The mistake isn't the policy. It's the assumption that having a policy and disclosing one are the same thing. In the eyes of the card networks, they are not — and the gap between them decides a surprising number of disputes.
A policy nobody saw is not a policy
Both Visa and Mastercard build their dispute rules around a simple principle: a cardholder can only be held to terms they had a reasonable opportunity to see before they paid. Your refund policy might be thorough, fair, and legally sound. But if it lived on a page three clicks from checkout, behind a footer link the customer never touched, the network rules treat it roughly the way contract law treats fine print nobody could read: as if it barely existed.
This is why disputes with reason codes like Visa's 13.6 (Credit Not Processed) and 13.7 (Cancelled Merchandise/Services) so often turn not on what your policy says but on where it was when the customer clicked pay. Visa's guidance for card-absent transactions is unusually concrete on this point: cancellation and refund terms should be presented during the checkout sequence itself, near the final confirmation, ideally with some affirmative act — a checkbox, a 'click to accept' — tying the cardholder to them. A policy disclosed that way is evidence. A policy discovered later by the merchant's lawyer is an afterthought.
When a bank analyst reviews your response to a refund dispute, the questions run in a fixed order. Was there a policy? Was it disclosed before the transaction completed? Did the cardholder have to acknowledge it, or at least pass directly by it? And does the policy actually cover the situation the cardholder is describing? Merchants tend to prepare obsessively for the fourth question and lose on the second.
What disclosure looks like when it works
Picture the checkout page the way an issuing bank's reviewer will reconstruct it. The strongest position is boring and specific: a short statement of the refund terms — not a link, the terms themselves or a one-line summary of them — sitting within the visual field of the button that completes the purchase, plus a record that this is what the page looked like on the date of the transaction.
That last part is the piece most merchants can't produce. Checkout pages change. If your evidence is a screenshot of today's checkout and the disputed purchase happened in February, a careful reviewer can discount it, and a cardholder who says 'that wasn't there when I bought' has an open lane. Merchants who win these disputes consistently tend to keep dated captures of their checkout flow — every time it changes, a screenshot goes in a folder. It feels like paranoia until the first time a dispute arrives from four months ago and you can show exactly what the buyer saw.
The acknowledgment itself matters too. A checkbox reading 'I agree to the refund policy,' logged with a timestamp and the customer's IP address, converts your policy from something you published into something the customer did. Disputes are decided on documents, and a logged action is a document. A paragraph of prose is scenery.
The psychology underneath the paperwork
There's a reason disclosure works beyond the rulebook, and it comes from a well-studied corner of social psychology: procedural justice. Research going back to John Thibaut and Laurens Walker's work in the 1970s found something counterintuitive — people's willingness to accept an unfavorable outcome depends heavily on whether the process that produced it felt fair, sometimes more than on the outcome itself. Tom Tyler's later work extended this: when people feel they had voice, notice, and a transparent set of rules, they comply with decisions that go against them at strikingly higher rates.
A refund denial lands very differently depending on the process around it. A customer who checked a box acknowledging '30-day return window' and asks on day 45 knows, somewhere, that they're outside the rules — the denial may frustrate them, but it doesn't feel arbitrary. A customer who never saw the policy experiences the same denial as an ambush. And ambushed people escalate. The chargeback, from the cardholder's chair, is not fraud; it's an appeal to a higher authority against a decision that felt rigged. Good disclosure doesn't just win the disputes you get. It quietly prevents a share of them from ever being filed, because the 'no' you eventually deliver arrives inside a process the customer already agreed to.
This also reframes what a refund policy is for. It isn't armor for the merchant. It's advance notice for the buyer — and it only performs either job if it's delivered in advance.
Writing a policy that survives a dispute
A few properties separate policies that hold up from policies that collapse under a reviewer's eye.
Specificity beats severity. 'All sales final' feels protective but reads as boilerplate, and some issuers treat blanket no-refund language skeptically, especially for services. 'Refunds available within 14 days of purchase; after 14 days, purchases are non-refundable' is narrower and far more enforceable, because it demonstrates the customer was told exactly where the line was.
The policy must match the dispute. If a cardholder claims they cancelled a subscription and were billed anyway, your return policy for merchandise is irrelevant — the reviewer needs your cancellation terms and your record of when (or whether) the cancellation actually happened. Policies fail in disputes most often by answering a question nobody asked.
Consistency is evidence too. If you've made goodwill exceptions — refunded other customers outside the window — that's fine, but a cardholder who can show you bent the rule for others has a fairness argument. Decide what your real policy is, then follow it, and let exceptions be rare and documented.
Keep the receipt of the disclosure, not just the disclosure. The order confirmation email that restates the refund terms is quietly one of the best pieces of evidence you can generate, because it puts the policy in the customer's own inbox with a timestamp neither of you can edit.
None of this requires legal sophistication. It requires deciding, once, that the policy lives at the moment of payment rather than in the site's basement — and keeping the records that prove it was there.
When the dispute comes anyway
Even perfect disclosure won't stop every refund dispute, and when one arrives, the burden flips to you: gather the dated checkout capture, the acceptance log, the confirmation email, the refund-request correspondence, and assemble them into a response the bank can follow — inside Stripe's unforgiving submission window. This is exactly the moment Argeback was built for. It ingests the dispute the day it lands, matches it to the right reason code, drafts an evidence-backed response around your policy and your records, and files it before the 7-day deadline — all from your phone, so a dispute that arrives during your busiest week doesn't lose by default. The policy protects you at checkout; Argeback makes sure it shows up at the hearing.