Refunds and Chargebacks in Crypto Payments

What exactly does crypto remove?

The forced reversal. On cards a chargeback lets a cardholder’s bank take money back from a merchant, and the merchant carries the cost of disputing it whether or not the claim is honest.

Crypto has no equivalent primitive. There is no issuer, no scheme, and no mechanism by which a third party can move a settled payment. For merchants in categories where friendly fraud is routine, this single property is worth more than every other difference combined.

It also removes the per dispute fees, and the scheme monitoring programmes that reprice or terminate an account whose ratio crosses a threshold.

What does it cost you?

Your customer’s route to redress. Someone with a genuine grievance cannot escalate to a bank, so the only available process is yours. If yours is slow or unreachable, the complaint goes to a public channel instead, and that is a worse outcome than a chargeback for most businesses.

The practical conclusion is that refund handling should be better on crypto than on cards, not merely equivalent. You have removed the safety net that was catching your mistakes.

How does a refund actually work?

You send a new payment. That means three decisions that card refunds never require.

Which address. The sending address is often not one the customer can receive at, particularly if they paid from an exchange. Ask for a destination rather than assuming.

Which amount. If the payment was in a volatile asset, the value has moved. Returning the same quantity of crypto and returning the same fiat value are different numbers, and you need a stated policy rather than a case by case decision.

Who pays the network fee. Somebody does. Deducting it is defensible and needs saying in advance rather than discovering.

Refunding in a stablecoin removes the second problem entirely, which is one more reason to prefer them.

What about partial payments and overpayments?

Both are refund shaped and neither is a dispute. Underpayment is common enough to need an automatic tolerance rather than a manual decision. Overpayment is rarer and more awkward, since the customer has sent you money you did not ask for and the return trip costs a network fee.

Decide the thresholds before launch and put them in your terms. Accepting crypto payments covers where these sit in the wider flow.

Do reserves still exist?

At some providers, yes. A reserve holds back a percentage of your balance against future refund obligations and releases it after a period. It is a card processing mechanism that carried over, it is rarely mentioned on a pricing page, and it decides where your working capital sits.

Ask for the reserve percentage and the hold period as numbers. Merchant services covers the rest of the contractual terms that matter more than the headline rate.

Writing a refund policy that works on this rail

State the window, the method, and what the customer needs to provide. Because a refund is an outbound payment to an address, you need that address, and asking for it late in an already unhappy conversation goes badly.

State which amount you return: the original fiat value or the original quantity of the asset. For stablecoins these are the same and the question does not arise. For anything else they diverge, sometimes sharply, and the policy has to say which before a case appears.

The operational cost nobody budgets

Refunds cost a network fee every time. On a cheap network that is negligible; on an expensive one, at small order values, it can approach the refund itself. Deciding whether you absorb it is a policy question, and stating it in advance is the difference between a term and an argument.

Partial and over payments

Set an underpayment tolerance and apply it automatically. Overpayment is rarer and needs a decision too: returning the excess costs a network fee, and keeping it without saying so is not defensible.

Both belong in your published terms rather than in an operator’s judgement, because both will happen and neither is worth a case-by-case conversation. The underpayment entry covers why the first is so common.

Reserves, and how to price them

If a provider holds a reserve, it is funding your refund exposure with your money. Ask for the percentage, the hold period, and what changes them.

Then price it: a ten percent reserve on a rolling ninety day basis is roughly a month of revenue permanently unavailable. That is a bigger number than any fee difference in this category, and it appears on no pricing page.

Where to go next

The chargeback entry covers the mechanism this rail removes, and the reserve entry covers how providers fund the exposure that remains. Merchant services covers the contractual terms that govern both.

What this changes about fraud

It removes the payer-initiated variety entirely, which is the dominant form in card-accepting businesses and the reason several verticals struggle to keep processing at all.

What remains is ordinary commercial risk: goods not delivered, disputes about quality, and mistakes on both sides. Those were always yours to handle, and the difference is that now there is no external process to fall back on when your own handling is poor.

Businesses that treated chargebacks as their quality feedback loop should replace it deliberately rather than discovering they no longer have one.

One thing to publish

Your refund window and method, on a page a customer can find before buying rather than after complaining. In a category without disputes, visible terms are what stands between a disagreement and an escalation you cannot control.

The short version: you have traded a dispute process you could not control for one you own entirely, and owning it means running it well rather than not running it.