Returns management for apparel brands: the 2026 guide.

Why apparel returns are different

Online apparel returns run roughly 20-30% of orders, against under 10% for most other retail categories. The driver is structural, not fixable by policy: customers buy fit they cannot try. Sizing varies between brands, between categories, and - when grading discipline slips - between styles within one brand. Bracketing, ordering two or three sizes with the stated intention of returning the rest, has moved from edge behaviour to normal shopping pattern; for some brands a third of “demand” is provisional.

That rate turns returns from a customer-service afterthought into a core operational flow. At a 25% return rate, a brand shipping 200,000 D2C units a year is receiving 50,000 - each needing transport, inspection, a disposition decision, a refund event and a restock or write-off. The brands that treat that as a designed process protect margin twice: on the cost of processing, and on the recovered value of the returned unit. The brands that treat it as exceptions handling leak on both.

The returns loop, end to end

Every apparel brand runs this loop. The difference between running it well and badly is whether each step is systematised - and whether the loop is one designed flow or six teams’ improvisations.

D2C returns: the fit problem at scale

The D2C flow is a volume problem: thousands of small parcels, consumer expectations of fast refunds, and fit as the dominant reason code. The operational targets are speed and data.

Speed, because a returned unit is capital asleep: until inspected and restocked it can’t be resold, and in apparel it’s depreciating toward the markdown calendar while it waits. Best-run operations re-list returned stock in days; a returns bin that takes three weeks to work through is a stockout generator for the exact sizes selling best - because the best-selling sizes are also the most-returned ones.

Data, because the return reason captured at inspection is the cheapest product intelligence a brand collects. “Too small” clustering on one style is a grading flag. “Colour not as pictured” is a photography flag. Reason codes rolled up by SKU, fed back to product and ecommerce teams each season, are how return rates actually move - none of which happens if the inspection step records only “received.”

Size exchanges deserve their own path: an exchange kept is revenue saved, so offering exchange-first (with instant re-shipment of the new size) before refund converts a meaningful share of would-be refunds - but only if inventory visibility is accurate enough to promise that size now, which loops back to how fast returns restock.

Wholesale returns: RAs, claims and deductions

The wholesale flow is contractual, not consumer. Retailers return under negotiated terms - end-of-season stock balancing, damage and defect claims, compliance failures - and the mechanics run through return authorizations: the brand issues an RA specifying what may come back, the retailer ships against it, and receiving matches physical goods to the RA line by line.

The financial exposure is different too. Wholesale returns interact with the deduction system: retailers commonly short-pay invoices for claimed returns, damages or compliance violations, and the brand’s job becomes matching those deductions on the 820 remittance against actual authorized returns - then disputing the gap. Brands that don’t systematically match deductions to RAs are, in practice, accepting every claim; the industry pattern is that a meaningful share of deductions are disputable, and disputes with documentation win.

Which is why running wholesale returns through the D2C portal-and-refund process fails: the volumes are lower but the values are higher, the paperwork is contractual, and the money moves through credit memos and remittance deductions rather than card refunds. It’s a different workflow that happens to share a receiving dock.

Returns and inventory accuracy

Between the carrier’s first scan and the restock decision, a returned unit exists in limbo: sold, refunded (or credited), physically present, but not yet sellable. Get the systems handling of that limbo wrong and inventory is wrong in one of two directions.

Count returns as available on receipt - before inspection - and the brand sells units that turn out damaged, generating a second, angrier return. Leave returns out of the count until someone processes the bin, and in-season availability is understated by whatever the bin holds - missed sales concentrated in the best-selling sizes. The correct model is an explicit returns-in-process state, visible but not sellable, cleared by inspection with a disposition - which requires the returns flow to live in the same system as inventory, not in a returns app reconciled monthly.

Disposition and restocking

Every inspected return gets one of a handful of fates, and the margin difference between them is the economics of the whole returns operation:

• Restock as new - full recovery; requires condition standards and often re-pressing/re-bagging
• Rework then restock - minor repairs, re-ticketing; recovery minus labour
• Outlet / off-price - imperfect but sellable units into the secondary channel at planned margin rather than ad-hoc
• Donate or recycle - value recovery near zero, compliance and sustainability value real
• Dispose - the fate to minimise by every upstream decision

The practice that separates disciplined operations: disposition rules, not disposition judgment calls. Category, condition grade and value band map to a default fate; inspectors apply rules and flag exceptions. It’s faster, it’s consistent across staff, and it makes recovered-value reporting meaningful because the same unit gets the same fate on Tuesday as on Friday.

Reducing returns, not just managing them

Managing returns well caps the cost of each return. Reducing them attacks the count - and in apparel the levers are concrete:

• Grading consistency. The single biggest brand-controlled driver. If a medium fits like a medium across every style, bracketing falls; if grading drifts by style, size guidance can’t save it. This is a product-development discipline enforced through the tech pack and spec process.
• Fit content that reflects reality. Model height/size worn, garment measurements (not just body measurements), fit notes per style (“runs small, size up”) - unglamorous and measurably effective.
• Colour-accurate photography. A returns reason code that photography budgets can directly retire.
• SKU-level reason analytics. The loop-closer: return rate by style × size × reason, reviewed by product teams each season, so the three styles with genuine fit problems get re-graded instead of everything getting a generic size chart update.

Brands that work these levers run below category-average rates - not to single digits, because bracketing and fit uncertainty are structural, but the gap between a 30% operation and a 20% one is entire points of net margin.

Best practices for 2026

1. Self-service authorization with clear windows and condition policy - friction belongs in policy design, not in making customers email.
2. Inspect before restock, always - with reason capture at the inspection station, coded, not free-text.
3. Days-not-weeks restocking - measure returns-bin dwell time as a KPI alongside return rate itself.
4. Exchange-first for size issues - instant re-ship of the new size, saving the sale the return would have killed.
5. Separate wholesale RA workflow - RA issuance, line-level receiving match, credit memos, and deduction matching against remittances as standard practice.
6. Returns-in-process as a first-class inventory state - visible, unsellable, cleared by disposition.
7. Disposition rules by category/condition/value - judgment for exceptions only.
8. Season-close returns review with product teams - the analytics exist to change next season’s grading, not to decorate a dashboard.

Frequently asked questions