A customer orders a medium. It arrives, it doesn't fit, and now there's a return. Nothing about the product was wrong - the size was. That single mistake triggers a return label, a trip back to a warehouse, an inspection, a restock, and a refund, when the actual fix - swapping a medium for a large - takes about five seconds to type.
The reason it turns into a return instead of a five-second fix is timing. By the time the customer realizes the size is wrong, the order has already shipped. There's no form for "actually, make that a large" on a package that's already in a truck - so the only tool left is the one built for products that are actually wrong: return it, refund it, hope they reorder.
The mistake isn't the wrong size. It's that the fix only becomes available after the correction is already the expensive kind.
Why a return costs more than the refund
A refund is the visible cost. It isn't the whole cost. The item travels back at your expense or the customer's, someone inspects it, and if it's apparel, there's a real chance it comes back creased, worn once, or out of season by the time it's back on a shelf - which makes it a write-off, not a restock. None of that touches the Shopify Payments processing fee, either: refunding an order doesn't return the 1.5-2.9% you already paid to accept the original payment, so a full refund is a guaranteed loss on top of the product.
And a return doesn't reliably become a reorder. Plenty of customers who send a wrong size back simply don't come back to buy the right one - the moment of intent has passed, and now it's a new decision instead of a correction to one they'd already made.
Catch it before it ships, not after it's returned
The fix is the same principle that works for address typos: move the correction earlier, onto the thank-you page and order status page the customer is already looking at, before the warehouse has touched the order. Instead of a size or color swap requiring a full return cycle, it becomes an edit to the order that's already sitting there - a variant swap, settled in place, on the same order number.
What a good pre-shipment swap flow includes
- Variant swaps for size and color that check current inventory before confirming, so a customer never swaps into a variant you can't fulfill
- An edit window tied to your actual fulfillment cutoff, so the option disappears the moment the order is picked, not on a fixed timer
- Automatic settlement of any price difference in either direction, so a swap to a pricier variant doesn't require a separate charge
- Auto-approval for straightforward swaps, with a queue only for the ones you've decided need a human look
- A record on the order of what changed and when, so support isn't reconstructing it later from an email thread
The cheapest return is the one that never ships.
This works because it's the same edit either way - a customer changing their mind about a size doesn't need a different system than a customer fixing a typo. It needs the same self-service window, applied to a different field on the order.
Set the swap window by fulfillment status, not a guess
A round-number cutoff - "edits open for 24 hours" - is a reasonable starting point, but the safer version ties eligibility to what's actually happened to the order: swaps stay open until the order is picked, and close automatically the moment that's no longer true. A store that ships same-day needs a tighter window than one that batches overnight, and tying the cutoff to fulfillment status instead of the clock means it adjusts itself as your operation speeds up.
It's also why hiding ineligible swaps matters as much as offering eligible ones. Once an order has been picked, showing a "change size" button that can't actually be honored just produces a request someone has to decline. Hide it instead, and there's nothing to decline - the customer sees only what you can still deliver on.
What this saves beyond the refund
Compare the two paths on the same wrong-size order. The return path costs a shipping label, a restock or a write-off, a refund, and the processing fee you don't get back - plus a real chance the customer never reorders at all. The pre-shipment swap costs nothing beyond the customer picking a different size on a page they were already on, and it keeps the sale, the order number, and the fees you already collected.
It also removes a support step most stores don't count separately from returns: the back-and-forth confirming the return was received, approving the refund, and answering the follow-up about when it'll post. A swap that happens before the order ships generates none of that thread, because there was never anything to send back.
- Add variant swaps to wherever customers can currently request a size or color change.
- Check live inventory at the moment of the swap, not just at checkout, so customers can't swap into a variant you're out of.
- Tie the swap window to your fulfillment cutoff, not a flat number of hours.
- Auto-apply straightforward swaps and route only the sensitive ones to a queue.
None of this requires a looser return policy or a bigger returns team. It requires giving customers the fix five minutes after checkout, instead of the one that only shows up two weeks later in a box.