Return vs exchange: the one that leaves an order open

Return vs exchange: the same request, two different obligations

A return sends the item back and closes the transaction: the money leaves the business and the order is finished. An exchange sends the item back and replaces it, which Shopify defines as sending the customer an alternative item as part of the return, so the transaction stays open and a second delivery is owed. Almost everything written on this argues about which outcome you should offer. The decision that costs money is which outcome you should take, order by order, and the exchange is the one that leaves an unfinished order behind. The break is specific: the replacement is promised at the moment the customer submits the request, and the stock is only committed days later when the original comes back.

NRF and Happy Returns put returns at 19.3% of online sales in 2025, so this isn't a rounding error. What follows is the comparison that matters operationally, the rule that decides it per order, the second order an exchange leaves behind, and where the law binds.

What actually differs between a return and an exchange

Both start as the same request, approved at the same RMA in shipping step, and both put the same parcel on the inbound lane. What separates them is what your operation still owes when it lands.

• What ends: a return ends once the refund clears. An exchange does not: the platform creates the return in an OPEN state with a reverse fulfillment order, the work required to process the return.
• Where the money goes: out on a return, and on a clock: a prompt refund is seven working days under the FTC's rule, or one billing cycle on a credit card sale, and Regulation Z then gives the issuer three business days to credit the account. Sideways on an exchange, and it can move toward you: the merchant action is release the exchange and collect a balance, invoiced at processing or taken from the card later.
• What you owe next: nothing on a return, once the refund lands. A second delivery on an exchange, with everything a first delivery can go wrong with.
• What has to be true for it to work: a return needs the item back and a payment rail. An exchange needs a specific unit of a specific variant to still exist when the original arrives, the condition nobody states.

To be plain about our edge: Keeyu isn't a returns portal. We don't host the return request, write the policy, print the label or issue the refund. We act on the break the choice leaves behind.

Exchange-first is a policy, and a policy cannot make this decision

The consensus advice is consistent, and it's not wrong. Free return shipping on exchanges and a fee on refunds. Bonus credit on top of the exchange value. An exchange the customer can take against the whole catalog rather than the same product in another size. Those are real levers, and they work on the choice the customer makes.

They do nothing about the choice you make. A lever that lifts the exchange rate lifts it on every return equally, including the returns where an exchange is the wrong answer and the expensive one. Widening the exchange to the entire catalog, a move this category recommends constantly, widens the exposure instead of narrowing it: more variants promised, from more locations, against stock nobody checked. One fee mechanic is worth knowing and no more: a restocking fee is a merchant-set percentage shown at the return request, and it's not deducted from the refund automatically. Someone applies it.

Policy sets the menu. It can't read an order.

The return reason decides which one is right

You already hold the deciding input. The reason is captured when the return is requested, before anything moves, and it's routinely treated as a reporting field somebody reads at quarter end. It's the decision, and it sorts into three classes that owe three different things.

• The customer chose wrong: size, fit, color, changed mind. This is the exchange case, and the only one where a second unit of the same product is genuinely the better outcome for both sides.
• You shipped wrong: a defect, a mis-pick, the wrong variant, damage found on opening. An exchange here ships a second unit of the thing that already failed, at your cost, and where the cause was a batch or a pick process it fails the same way. Fix the cause, then decide.
• The delivery went wrong: late, damaged in transit, left at the wrong address. The item was never the problem, so replacing it addresses nothing, and the reason code will say return while the operational break sits upstream of the return entirely.

The policy sets what is on the menu. The reason picks off it, one order at a time.

An exchange creates a second order with no stock reserved for it

Shopify's returns documentation puts it in one sentence: "Exchange items inventory isn't reserved until you process the return."

Read that as an operator. The customer is told the replacement is coming at the moment they submit the request. The stock is committed days later, when the original comes back and the return is processed. In between, the variant can sell out to somebody who paid for it today. Every order is a promise, and an exchange makes a second one before anyone has checked it can be kept.

The stock gap isn't the only place an exchange goes quiet. An exchange fulfillment with an outstanding balance is held until the payment clears or a human releases it. That's a queue, and a queue nobody owns is where an exchange quietly turns into a "where is my order" ticket three weeks later. One return, two broken promises.

The same documentation carries the constraints nobody mentions: exchange items can't be custom items, order-level discounts can't apply to them, and they can't go on an order carrying duties. Each is a case where the exchange your policy promised can't be built. I have yet to read a page on this subject that follows the exchange past the moment the customer picks it, which is where the work starts.

What the law settles, and what it leaves to your policy

The question underneath this keyword is whether you can offer exchanges only. In the US, mostly yes, conditionally. In the UK and the EU, mostly no.

California's Attorney General is explicit that there's no general legal right to a refund for a change of mind, and that the posted policy governs. California Civil Code 1723 contemplates a posted policy of "cash refund, store credit, or exchanges", requires it to be displayed conspicuously with its time period and conditions, and makes a retailer that fails to post liable for the purchase amount on goods returned within 30 days. New York law requires an online retailer to state whether refunds come "in cash, or as credit or store credit only", and to name any fees. In the UK a cancellation must be reimbursed within 14 days by the same means of payment, and the EU Consumer Rights Directive sets the same 14 days. An exchange-only policy satisfies neither.

One clock binds whatever your policy says, and it sits on the order you already took: ship within the time you stated or 30 days, and where you cannot, offer the buyer a choice of consenting to the delay or canceling for a prompt refund.

Detect the break, decide the outcome, act on it

Keeyu is proactive e-commerce operations, a different category from support tooling: a helpdesk replies about problems rather than resolving them, and one in five orders hits an operational break after checkout. We detect the break when it happens, not when the return request arrives. We decide from the reason and the real stock position, not a policy default. We act, which means checking the replacement against the stock on hand the moment the exchange is raised in the returns app you already run, then triggering the refund or the replacement order in the platform that issues it.

An exchange promised before the stock was checked is a second order nobody is watching, and it surfaces three weeks later as a customer asking where their replacement is. We check that second promise against real stock the moment it's made, usually before the customer knows anything went wrong. Every order is a promise, including the second one an exchange creates. If your exchanges are running on hope, book a demo.

• National Retail Federation and Happy Returns. 2025 Retail Returns Landscape, 15 October 2025. NRF returns research
• Shopify Help Center. Creating and processing returns and exchanges. Shopify returns and exchanges
• Shopify Help Center. Returns and exchanges. Shopify return creation
• Shopify developer documentation. Returns apps. Shopify returns apps
• Shopify developer documentation. returnCreate mutation, Admin GraphQL API. Shopify returnCreate mutation
• Shopify Help Center. Setting up return and cancellation rules. Shopify return rules
• 16 CFR 435.2, Mail, Internet, or Telephone Order Merchandise Rule, via Cornell Legal Information Institute. 16 CFR 435.2
• 16 CFR 435.1, definitions, including prompt refund, via Cornell Legal Information Institute. 16 CFR 435.1
• 12 CFR 1026.12(e), Regulation Z, prompt crediting of returns, via Cornell Legal Information Institute. Regulation Z 1026.12
• California Civil Code section 1723, display of refund policies. California Civil Code 1723
• New York General Business Law section 218-a, refund policies. New York GBL 218-a
• California Attorney General. Refund Policies. California Attorney General
• The Consumer Contracts (Information, Cancellation and Additional Charges) Regulations 2013, regulation 34. UK Consumer Contracts Regulations
• Directive 2011/83/EU on consumer rights, Articles 13 and 14. EUR-Lex, Directive 2011/83/EU