The terms you're proud of might not exist

Most merchants have a terms-of-service page. It took an afternoon to write, or a lawyer an hour to bless, and it sits in the footer of every page linked in tidy gray text. It says the customer agreed to a no-refund window, a recurring charge, a cancellation process. When a dispute lands, that page feels like the trump card. They agreed to this. It's right here.

Here is the uncomfortable part. In the eyes of a card network reviewer — and, in the rare case it escalates, a court — your terms may not legally exist for that customer. Not because the words are wrong, but because there is no proof the person ever agreed to them. A policy nobody assented to is a policy nobody is bound by. And the difference between terms that bind and terms that evaporate comes down to one design decision most merchants never think about: how the customer met the agreement in the first place.

Two ways to present an agreement, two very different outcomes

Contract law has a plain-English name for the mechanism that matters here: manifest assent. For an agreement to bind someone, they have to have done something that signals yes, I agree — not merely had the opportunity to find out what they were agreeing to.

That single principle splits online terms into two families.

Clickwrap is the version where the customer takes an affirmative action. A checkbox next to "I agree to the Terms of Service." A button that reads "By subscribing, you accept our terms." The act is deliberate, recorded, and unambiguous. Courts have upheld clickwrap agreements consistently because the click is the assent. There is a moment, logged with a timestamp, where the person did the equivalent of signing.

Browsewrap is the version most merchants actually have. The terms live behind a link in the footer or at the bottom of a signup form, and the theory is that by continuing to use the site, the customer has silently accepted them. No box. No click. Just a link that was, in principle, available.

The legal record on browsewrap is far shakier. In Nguyen v. Barnes & Noble (2014), a federal appeals court declined to enforce a browsewrap agreement because a reasonable user wasn't put on notice that using the site meant accepting the terms. More than a decade earlier, in Specht v. Netscape Communications (2002), another court refused to bind users to terms they would have had to scroll down to even see. The through-line across these cases is consistent: availability is not agreement. If the customer could have missed it, they aren't bound by it.

Why this decides chargebacks, not just lawsuits

A chargeback almost never reaches a courtroom. So why does a body of contract law matter to a dispute settled by a card-network reviewer working through a queue?

Because the reviewer is asking the same question the court asks, just faster and with less patience. When you dispute a subscription chargeback or a "services not as described" claim, the card networks want to see that the cardholder agreed to the specific term you're leaning on — the renewal, the cancellation policy, the delivery timeline. Visa's compelling-evidence standards for subscription and digital-goods disputes explicitly favor proof that the customer accepted terms at the point of purchase, not that terms existed somewhere on your domain.

Paste a link to your terms page into a dispute response and you've shown the reviewer a document. Show them a record that this cardholder, on this date, checked this box or clicked this button before the charge — and you've shown them assent. One is a wall of text. The other is a signature. Reviewers reward the signature, because it's the thing that would survive if anyone ever pushed further.

The evidence that actually carries weight

The useful reframe is to stop thinking of your terms as a page and start thinking of them as an event. The page is static; the event has a witness and a clock. What you want in a dispute file is the record of the event.

Concretely, the strongest version looks like this:

  • The exact language the customer agreed to, as it read on the day they agreed — not today's revised version.
  • A timestamp of the acceptance, tied to the customer's account or email.
  • The mechanism of acceptance: "Customer checked the box confirming the recurring monthly charge and 14-day cancellation window at 2:14 PM on the signup date."
  • Ideally, a screenshot or description of the checkout screen showing that the terms were adjacent to the action, not buried three clicks away.

Notice that none of this is about having better legalese. A short, clear sentence the customer actively accepted beats an exhaustive twelve-page agreement they were merely near. The card networks, like the courts, care about proximity and action — was the term right there at the moment of commitment, and did the person do something to accept it.

What to change before your next dispute

If you take one operational thing from all this, make it this: audit the moment of agreement, not the agreement.

Walk through your own checkout as a stranger would. Is there a point where the customer must actively confirm they accept the terms — a checkbox, a button with acceptance language baked into it? Or do your terms rely on a footer link and a hope? If it's the latter, you have browsewrap, and you're carrying a policy that may dissolve the instant it's tested.

Then check whether you're actually storing the acceptance. Plenty of merchants have a perfectly good clickwrap checkbox and no record that any given customer ever clicked it. The box protects you only if the click is captured, dated, and retrievable months later when the dispute arrives. An unrecorded click is, for evidentiary purposes, no click at all.

And keep versioned copies of your terms. When a customer disputes a charge from four months ago, the term that binds them is the one that was on screen four months ago — not the one you rewrote last week. If you can't reproduce what they saw, you can't prove what they accepted.

The quieter lesson

There's a broader idea underneath the legal vocabulary, and it's genuinely useful whether or not chargebacks are on your mind. Agreement is a behavior, not a document. People are bound by what they did, and "did" means a deliberate, observable act — a click, a check, a signature. Everything a business wants to hold a customer to has to trace back to a moment the customer acted. Design for that moment, capture it, and most of your disputes are half-won before they begin. Ignore it, and the finest terms of service ever written are just words on a page nobody agreed to.

That gap — between the terms you have and the assent you can prove — is exactly where disputes are lost, and it's easy to overlook when a chargeback lands and the clock is already running. Argeback works this angle for you: it ingests the dispute, pulls the acceptance record and the terms the customer actually agreed to into an evidence-backed response, and files it before the seven-day deadline, all from your phone. When you want your terms to count when it matters, you can start at argeback.lumenlabs.works.