All sales final: when a no-returns policy works (and when it just hides friction)

"All sales final" sounds like the simplest policy a store can publish: no windows to track, no return freight to eat, no judgment calls at the counter. It is actually the strictest policy, and strictness bills you in a currency that never appears on the policy page — the customers who don't buy, the ones who pre-buy and then preemptively email to ask, and the size-worriers who just shop elsewhere. So this article doesn't answer "should I write all sales final?" with a yes or no. It gives you the decision table for which parts of a catalog it genuinely fits, the disclosure rules that still bind you after you write it, and a checklist to run before publishing — because the stores hurt by this policy are almost never the ones that chose it carefully.

A policy built around accepting returns — windows, conditions, who pays freight — is designed in the Shopify return policy design guide. This piece handles the opposite decision: whether you accept returns at all, category by category, and what you still owe customers when the answer is no.

What "all sales final" can and can't override

Writing "all sales final" into a policy does not repeal consumer law. What it can and can't do:

• It can refuse change-of-mind returns — the customer who ordered the wrong color, the gift that missed. In the US there is no federal right to return for change of mind, so a disclosed no-returns policy generally holds (check your own state; some have specific disclosure duties, below).
• It can't erase statutory remedies for faulty goods. In the UK, the Consumer Rights Act 2015 gives statutory rights against faulty or misdescribed goods that store policy cannot contract away (accessed November 2026). The EU is similar for distance sales. Write the defect carve-out because it's fair and because it's the law in much of your market.
• It can't skip disclosure. Two US examples of how strict the rules are: California requires a store that doesn't accept returns to conspicuously disclose that fact — otherwise buyers may return goods within 30 days of purchase for a full refund (California Civil Code § 1723, accessed November 2026). New York's rule works the same direction: if no refund policy is disclosed at the point of sale, customers are entitled to a refund within 30 days with proof of purchase and the item in original condition (N.Y. General Business Law § 527-a, accessed November 2026). An undisclosed "all sales final" is, in those states, legally just a 30-day refund policy you forgot to write down.
• It doesn't apply to certain in-person sales at all. Under the FTC's Cooling-Off Rule, purchases above $130 made at the buyer's home or at temporary locations — craft fairs, pop-ups, trade booths — carry a 3-day cancellation right that no policy overrides (FTC, as of November 2026). If you sell at markets, that applies to you regardless of your policy page.

Where no-returns genuinely works

The policy isn't good or bad; it fits some inventory and chokes on others.

Where it just hides friction

The cost of a no-returns policy almost never shows up as returned units — those go to zero by definition. It shows up as friction you have to go looking for:

• Pre-sale anxiety. A store selling dog harnesses runs three sizes across two widths. If exchanges were running at roughly 8 per 100 orders (illustrative — set the number from your own exchange log), all-sales-final doesn't make those 8 fitting problems disappear. It converts them into 8 support negotiations, a handful of customers who keep the wrong size and never reorder, and — the invisible part — the shoppers who read "all sales final" on a three-size product and bought from a competitor with a friendlier page. The policy didn't cut your return cost. It cut your conversion, which is the same money lost a different way.
• Support tickets with worse outcomes. A return window is a valve: the awkward case gets handled once, by policy, in minutes. All-sales-final makes every awkward case a bespoke negotiation, which is slower, less consistent, and where exactly the resentful reviews come from.
• Chargebacks anyway. A customer refused a cash refund for defective or misdescribed goods can dispute the charge — and the dispute costs you the sale, a fee (estimated, and typically more than the refund you were withholding), and a processor relationship. The policy didn't save the money. It moved it to a worse channel.

This is why the useful framing isn't "returns are expensive" but "which returns am I actually getting?" If the answer is mostly size exchanges and gifts, the policy is taxing goodwill to save on freight you'd have paid anyway.

Before you publish: the disclosure checklist

Run this before the policy goes live. Each line is a check you can do in an afternoon:

1. Sort every category in the table above. All-sales-final is a per-category decision, not a storewide banner. Mixed catalogs usually want final sale on consumables and standard windows on everything else.
2. Write the defect carve-out, in one sentence, even in categories that are otherwise final sale. If you ship to the UK or EU this isn't optional in practice; if you don't, it's still the clause that keeps chargebacks down.
3. Disclose before purchase, conspicuously. Banner on the product page, line at checkout, not a footer link only. In California and New York, conspicuous disclosure is the difference between a valid policy and an automatic 30-day refund obligation (CA Civ. Code § 1723; N.Y. GBL § 527-a, accessed November 2026).
4. Write the support script. What exactly do you offer the customer who asks? Exchange, store credit, partial gesture — decide once. Store credit is usually the cheapest fair offer, and the reason why is priced out in the store credit versus refund comparison.
5. Check the market-and-pop-up rule. If you sell in person at temporary locations, the FTC Cooling-Off Rule gives buyers a 3-day cancellation right on qualifying sales over $130 (as of November 2026) — your policy can't touch that.
6. Instrument the friction. Add a "couldn't buy because of returns policy" tag to support email and pre-sale chat. After 60 days, count it. That count is the real price of the policy, and it's the one your sales dashboard won't show you.
7. Revisit in one quarter. If the friction tag count is growing faster than the freight you saved, the experiment failed — publish a standard window instead.

The middle path most stores end up on

Very few small stores that try pure all-sales-final keep it unmodified. The stable versions:

• Final sale + defect replacement. The consumables standard. Tight defect window (7 days is common — illustrative), replace-or-refund your choice, statutory-rights line included.
• Exchange-only. No cash refunds, but swaps and store credit welcome. Keeps the valve open at nearly zero cash exposure — most of what customers want is the right item, not the money back.
• Credit-only for change of mind, cash for your fault. The split that survives chargeback scrutiny best.

Any of these needs written clauses; the copy-paste versions with annotations are in the annotated return policy templates. What you should avoid is the accidental policy — all sales final as a default inherited from a theme's sample text, never chosen on purpose. That's the version that hides friction, because nobody is measuring it.

Common questions

Is an all-sales-final policy legal? For change-of-mind returns in the US, generally yes if disclosed before purchase — and some states punish undisclosed policies by granting an automatic 30-day refund window (California Civil Code § 1723, accessed November 2026). It can't override statutory rights for faulty goods in the UK/EU, or the FTC Cooling-Off Rule's 3-day right on qualifying in-home and temporary-location sales. Disclose conspicuously and keep the defect path open.

Does a no-returns policy actually reduce returns? It reduces change-of-mind returns, yes. It doesn't reduce defect claims, which move to chargebacks instead if you refuse them. And it can reduce sales: the policy's real cost is conversion friction on size-dependent and gift purchases, which appears as fewer orders rather than fewer returns. Measure both sides for a quarter before judging.

Do I still have to accept a defective return under all sales final? In the UK and EU, yes — statutory fault remedies survive whatever your policy says. In the US it varies, but the practical answer is yes too: the alternative is a chargeback plus a fee (estimated to exceed most refunds withheld) plus a public dispute. Write the defect carve-out; it's cheaper than enforcing the banner.

What should I offer instead of a refund when I say no? Decide once and script it: an exchange, store credit, or both. Credit keeps the cash in the store and gives the customer a reason to return, which is usually the cheapest fair outcome — the payout-path cost math is worked through in the store credit versus refund guide. The worst answer is improvising per case, which is how reviews happen.

Can I make just one category final sale? You can and usually should — consumables, personalized goods, and clearance are the classic fits, because the goods can't be restocked or the margin is gone. Size-dependent and gift categories keep a standard window. Category-level policies need the same conspicuous disclosure as storewide ones.

Next step

Take 30 minutes today and sort your catalog into the table above — every category, one verdict. Then write the defect carve-out and the disclosure line before touching the policy page, and if the verdict is "final sale with exceptions," build it from the annotated template set linked above rather than from a blank page. If any category that lands on final sale is personalized or made-to-order, put the cancellation terms in writing before the next such order — the stage table and deposit math for that case are in the cancelled orders guide.