A return policy is a promise. A returns SOP is the machine that keeps it the same way every time.
Almost every DTC brand has a return policy. Far fewer have a returns SOP. The policy is the paragraph on your site: 30 days, unworn, tags attached. The SOP is everything that happens after a customer clicks return that the policy never mentions.
That gap is where money leaks. One agent approves a return the policy should have rejected. Another makes the customer wait four days for an RMA number. A third refunds before the item is inspected, and the box comes back empty. Same policy, three different outcomes, because there's no documented process behind it.
Returns run around 17 percent of online orders, and higher in apparel. At that volume, 'we'll figure each one out' is not a plan. This is the returns SOP we use with DTC brands: the RMA workflow, who owns each step, the inspection rules, and the exchange-first defaults that keep the sale instead of refunding it away.
A return policy states what you'll do. A returns SOP states how, and who. The policy says 'returns accepted within 30 days for a refund or exchange.' The SOP says: the customer submits through the portal, CX verifies the order is inside the window, the warehouse inspects against a checklist, and only then does finance release the refund.
Customers see the policy. Your team lives in the SOP. When the SOP doesn't exist, every agent writes their own, and your return outcomes drift apart within weeks.
A return policy is a promise to the customer. A returns SOP is the machine that keeps the promise the same way every time, no matter which agent or warehouse picks it up.
The policy your returns SOP operationalizes: windows, conditions, and what qualifies for a refund.
There is a third thing sitting between the policy and the SOP, and most teams never look at it. Shopify enforces part of your return policy for you, from Settings > Policies, in the Return and cancellation rules section. What is configured there is what customers actually experience, regardless of what the written paragraph says.
Four settings carry the weight:
| Setting | What Shopify offers |
|---|---|
| Return window | 14 days, 30 days, 90 days, unlimited, or a custom number. A separate Starting from menu decides whether the clock runs from each item's delivery or from delivery of the last item in the order |
| Return shipping cost | Free returns, a flat rate charged one time per return, or the customer buys their own label |
| Restocking fee | A percentage of the return, shown to the customer when they request one |
| Final sale items | Specific products or specific collections. You can add several of either, but you cannot mix products and collections in the same rule |
The Starting from choice is the one that quietly changes your window. On a split shipment, delivery of the last item can add a week to the clock on everything else in the order. If your policy paragraph says thirty days and that setting says last item, some customers genuinely have longer than thirty days.
Shopify's documentation is explicit: return fees aren't automatically deducted from refunds. When you create a refund, you need to manually deduct any applicable return fees. So a restocking fee can be configured correctly, displayed to the customer correctly, and still never collected, because the agent issuing the refund has to subtract it by hand every single time. That step belongs in your SOP as its own line, not as an assumption.
One more thing worth knowing before you go looking for it. Self-serve returns is a separate switch, under Settings > Customer accounts, and it only governs whether customers can submit requests themselves. Your return rules still apply when your team processes returns manually, even with self-serve turned off.
Which gives you a three-way check rather than a two-way one. The published paragraph, the settings, and the SOP all have to say the same thing. In practice the settings are the version that wins, because they are the version the customer meets.
Setting the fee, wording it on your policy page, and the three-way check that keeps it consistent.
There is a fourth copy of your return terms sitting in your helpdesk macros. This is how to check it still agrees with the other three.
The policy is one paragraph. The SOP is the full lifecycle, from the return request to the final disposition of the item. At minimum it covers:
Miss any one of these and you get the classic failures: refunds issued for items that never arrived, products restocked that should have been binned, return reasons nobody recorded.
Returns cross three functions, which is exactly why they fall apart without an owner. Name who does what:
At a small brand one person wears all three hats. Write them down separately anyway. The handoff between CX approving and the warehouse inspecting is where returns stall, and you can't fix a handoff you never named.
Configure the self-serve return portal so initiation and label generation run without a CX ticket.
Here's the workflow end to end. The goal is a return that moves without a human chasing it, tied together by an RMA number that links the request to the box to the refund.
The most expensive returns mistake is refunding on the label scan instead of the inspection. Refund the moment a box ships back and you teach customers to game it. Tie the refund to inspection, not to transit.
We packaged this as a returns SOP pack: the RMA workflow, the condition checklist your warehouse inspects against, and the role split for CX, warehouse, and finance. Grab it from the Loop Returns and refund policy SOP pages linked here.
Try it on one of your own procedures.
Record a process once, AI writes the structured SOP. 3 free SOPs, no credit card.
Inspection is where a return stops being a refund and becomes a recovery decision. The warehouse needs a checklist, not judgment. For each return:
Disposition follows from condition. Restock the resellable, refurbish what's worth the labor, liquidate or donate the rest, write off what never came back. The point is that someone decided on purpose, instead of a pile of returns sitting in a corner losing value.
The adjacent procedure for items that arrive broken, which often re-enter your warehouse as returns.
A refund gives the money back. An exchange keeps it. Most return portals default to refund, which trains customers to expect their cash and quietly drains revenue on every return.
Flip the default. Offer the exchange or store credit first, with the refund available one step further down. Brands that lead with exchanges retain a real share of returns as revenue instead of reversed sales. The customer still leaves satisfied; you just don't hand back the order to get them there.
A denied or botched return is a common path to a chargeback. Here's the process for when one lands.
Every return carries a reason, and the reasons are the most underused data in DTC ops. Roughly 86 percent of returns are effectively decided before checkout, at the product page. The return reason tells you where.
If 30 percent of a product's returns say 'not as described,' that's a product page problem: the photos or copy are overselling. If 40 percent say 'wrong size,' that's a sizing problem: the guide is wrong or missing. The returns desk is a feedback loop into merchandising, but only if the reasons are captured at initiation and someone actually reads them.
Review return reasons monthly by SKU. The worst offenders earn a product page fix, a sizing chart update, or a packaging change. That's how a returns SOP pays for itself twice: once in cleaner operations, once in fewer returns next quarter.
Returns SOPs rot like every other operational doc. You move from native Shopify returns to Loop, you add a 3PL, you launch a final-sale category, and the SOP written before those changes now tells people to do things that no longer exist. One that references a manual label process you automated months ago is worse than none, because your team still trusts it.
Review it quarterly, and immediately after any change to your returns app, fulfillment partner, or policy. This is the same drift that quietly breaks every SOP, and it's why we treat them as living records instead of one-time documents.
Why every operational doc, including your returns SOP, degrades within 90 days, and how to catch it.
Don't document the whole lifecycle at once. Start with the one rule that saves the most money: refunds release on inspection, not on the label scan. Write that down, tell CX and your warehouse, and you've closed the empty-box hole today.
Then map the RMA workflow on a single page and name the CX, warehouse, and finance owners. A returns process that runs the same whether you handle 20 returns a week or 200 is worth more than any returns app bolted on top of chaos.
ReccordSOP turns a workflow like this into a documented SOP with timestamped screenshots, and flags drift when your tools change underneath it. Generate your first SOP free at reccordsop.com.
30 days is the DTC standard and what most customers expect. Shorter than 14 feels stingy; longer than 60 invites wardrobing. Match it to your category and state it clearly at purchase, with any final-sale exceptions called out.
Exchanges first. A refund reverses the sale; an exchange or store credit keeps it. Default the portal to exchange and credit with the refund one step further down, and reserve instant refunds for defects and your own errors.
Let the policy auto-approve the clear cases (inside the window, standard reason) and route only edge cases (outside window, final sale, high-value, reason mismatch) to a human. Auto-approving everything invites fraud; manually approving everything buries CX.
About 17 percent of online orders on average, higher in apparel where 20 to 30 percent is common. Track yours by SKU, not just overall, since the average hides the few products driving most of the returns.
Native tooling is fine under roughly 50 returns a month. Above that, a dedicated app like Loop Returns pays for itself through self-serve portals, automated labels, exchange-first flows, and reason analytics that native returns don't surface.
3 free SOPs to start. No credit card required. See if drift detection keeps your docs honest.
Start for freeI built ReccordSOP after watching too many DTC ops teams lose months to undocumented workflows. These SOPs are battle-tested with Shopify operators running $1M to $50M brands.
Last reviewed October 11, 2026
Most SOPs are wrong within 90 days of publishing. Here's how to detect it before it costs you a customer.
Manual responses win under 20 percent of the time. A repeatable process is how you flip that.
Macros silently break. Customers notice before you do. Here's the audit that catches broken macros before CSAT slides.
A refund starts in your helpdesk and ends in your store. The seam between them is where refund procedures quietly break.
We use essential cookies for sign-in and a small amount of analytics to improve the product. Privacy policy.