Return fraud in apparel: wardrobing, bracketing and how brands fight back
Nine percent of US retail returns are fraudulent, and apparel absorbs more of that loss than almost any other category, because a garment can deliver its full value in a single wear and still come back claimed as unworn. This guide covers what return fraud costs, the main fraud and abuse types, wardrobing, the difference between fraud and legitimate abuse like bracketing, how portals and warehouse inspection split the detection job, the wholesale version of the problem, and the controls that cut losses without punishing honest customers.
What return fraud costs apparel brands
The NRF and Happy Returns 2025 Retail Returns Landscape report, published October 2025, puts numbers on the problem. US retail returns reached $849.9 billion, a 15.8% overall return rate, rising to 19.3% for online purchases. Within that, 9% of all returns were fraudulent, and 45% of consumers admitted to “bending the truth” on at least one return, with roughly two thirds admitting to some form of abusive returns behavior.
Apparel sits at the sharp end of those averages for two structural reasons. First, fit uncertainty drives the highest legitimate return rates in retail, which gives fraudulent returns a large, normal-looking stream to hide inside. Second, clothing is the rare product that can deliver its complete value in one use and come back looking sellable. A blender that has made one smoothie is visibly used. A dress that has been to one wedding often is not.
The direct refund is only part of the loss. A fraudulent return also consumes return shipping, receiving labor, inspection time and, when it slips through, a restocked unit that is not actually sellable at full price. For brands running returns through a 3PL, each of those steps is a billed touch.
The main fraud and abuse types
Return fraud is not one behavior. The table below separates the common apparel patterns, marks which are actually fraud and which are policy abuse, and names the primary control for each. The distinction matters because the responses are different: fraud is stopped with verification, abuse is managed with policy and merchandising.
| Type | Fraud or abuse | What happens | Primary control |
|---|---|---|---|
| Wardrobing | Fraud | Item is bought, worn for an event or a photo, and returned as unworn, sometimes with the tags reattached. | Inspection at receiving: wear, odor, deodorant marks, missing or retagged labels. |
| Substituted item | Fraud | A different item comes back in the box: an older garment, a cheaper lookalike or a counterfeit, while the original is kept. | Scan and match at receiving: SKU, barcode and condition checked against the RMA line. |
| Empty-box return | Fraud | The parcel arrives with nothing inside, or stuffed with filler at the right weight, while the refund claim proceeds. | Weight capture and open-box verification at receiving, tied to the RMA before credit. |
| Item-not-returned claim | Fraud | The customer claims a refund for a return that was never actually shipped, or disputes delivery of the outbound order. | Carrier scan events matched to the RMA; refund held until first scan or receipt. |
| Stolen-goods return | Fraud | Shoplifted or fraudulently obtained merchandise is returned for cash, credit or gift cards, often without a receipt. | Receipt or order match required for refunds; no-receipt returns limited to identified exchange. |
| Bracketing | Abuse, not fraud | The same style is ordered in several sizes or colors with the intention of returning most of it. | Fit guidance, size charts and reviews reduce it; per-customer bracketing history informs policy. |
| Serial returning | Abuse | A small share of customers generates an outsized share of returns, some legitimate, some borderline. | Return-rate scoring per customer; selective friction such as inspection before credit. |
| Returnless-refund farming | Fraud | Customers learn which claims trigger keep-it refunds on low-value items and repeat them across orders. | Caps by order value, claim type and per-customer history; periodic claim audits. |
Two things stand out from the table. Almost every fraud type is caught at one of two points, either a pattern visible before the item ships back or a physical discrepancy visible at receiving. And the single highest-leverage control appears in the right column repeatedly: receive and inspect against an authorization before money moves. That is the discipline the rest of this guide keeps returning to.
Wardrobing: the apparel signature
Wardrobing is buying a garment intending to wear it once and return it: a dress for an event, an outfit for a photo, a jacket for a trip. The item comes back worn, sometimes with the tags carefully reattached, and often cannot go back on sale at full price. It is the fraud type most specific to apparel, common enough that some shoppers do not consider it fraud at all.
Detection is physical. Portals cannot see wear; receiving can. Effective wardrobing control means the returned unit is inspected against its authorization before the refund settles: tags present and original, no wear indicators, no odor, no marks. Brands that refund on carrier first-scan give that inspection up in exchange for customer experience, which is a legitimate trade for low-value basics and a poor one for occasion wear and high-ticket styles. The workable pattern is to segment: instant credit where wardrobing risk is low, inspection before credit where it is high or where the account is flagged.
Policy language helps at the margin. Explicitly stating that worn items are not refundable, and that returns are inspected, deters the casual wardrober. It does nothing against the determined one, which is why the inspection step, not the policy page, is the control.
Fraud vs abuse: why bracketing is not fraud
Bracketing, ordering the same item in several sizes or colors and returning what does not fit, is not fraud. It is legitimate use of the returns policy, and for many shoppers it is a rational answer to inconsistent sizing. It is also expensive: every bracketed order manufactures guaranteed returns, each carrying full shipping and processing cost, and bracketing has grown steadily as shoppers treat the bedroom as the fitting room.
The distinction matters operationally. Treating bracketing like fraud, with account blocks or refund friction, punishes some of your best customers. Treating fraud like bracketing, with a shrug and a restock, funds the fraudsters. Bracketing is reduced with merchandising: accurate size charts, fit notes, model measurements, reviews that talk about fit, and size recommendation tools. Serial returning sits between the two, a small share of customers generating an outsized share of returns, and is best handled with per-customer return-rate scoring that applies friction selectively rather than policy that applies it to everyone.
Detection: what portals see, what inspection catches
Fraud detection in returns runs at two layers, and neither replaces the other.
The portal layer sees patterns. Returns portals and order management rules work before the item ships back: return velocity per customer, repeat claims of the same type, claim reasons that do not match purchase history, mismatched addresses, and serial-returner scores. This layer is good at deciding how much friction a given return deserves: instant label and instant credit for a clean account, label but credit-on-inspection for a flagged one, manual review for the worst.
The inspection layer sees the goods. No pattern model can see inside the box. Receiving catches the substituted item, the worn garment, the missing tags, the empty box and the quantity short of the claim, but only if the parcel is received against a known authorization with the expected SKU, quantity and condition on it, rather than opened blind. Scan the barcode, match it to the RMA line, grade the condition, and route the unit: restock, rework, outlet or write-off.
The failure mode is running the layers apart. If the portal issues refunds and the warehouse records receipts and the two never reconcile line by line, fraud lives in the gap: refunds settle for goods that never arrived, and worn goods restock because nobody compared condition to claim. Detection works when both layers feed one record and money moves only after the record closes.
The wholesale version: deduction abuse
D2C gets the headlines, but wholesale has its own return fraud economics, and the sums per incident are larger. Retailers return goods under an authorization, then settle by short-paying invoices through deductions. Where discipline is weak, brands eat deductions for returns that never physically arrived, arrived short, or arrived in worse condition than claimed, plus compliance chargebacks stacked on top.
The control is the same discipline at wholesale scale: no return accepted without an RA, every RA received line by line with quantities and condition captured at the dock, credit memos raised from the receipt rather than the retailer’s claim, and every deduction on the remittance matched back to a specific RA, receipt or chargeback before it is accepted. Unmatched deductions get disputed with the receiving record as evidence. Brands without that chain have no evidence, and unverifiable deductions simply become margin leakage that no one can even size.
Prevention that does not kill conversion
The same NRF and Happy Returns 2025 report found 82% of consumers treat free returns as a major factor when deciding to buy online. Blanket tightening, shorter windows, restocking fees for all, refund-after-inspection for everyone, trades fraud loss for sales loss, usually badly. The controls that work are selective:
- Authorization on every return, both channels. The RMA or RA is the reference that receiving verifies against and finance settles against. No authorization, no automatic refund.
- Segmented refund timing. Instant or first-scan credit for clean accounts and low-risk categories; credit on inspection for flagged accounts, occasion wear and high-ticket items.
- Receiving that grades condition. Every returned unit scanned to its authorization and dispositioned: restock, rework, outlet, write-off. Worn goods never restock silently.
- Returnless refunds with caps. By order value, claim type and per-customer history, with periodic audits of who is farming them.
- Deduction matching on the wholesale side. Every short-pay tied to an RA, receipt or chargeback before acceptance; the rest disputed on the receiving record.
- Fit content that shrinks bracketing. Size charts, fit notes and reviews reduce the legitimate return stream, which also shrinks the cover fraud hides under.
None of these require treating customers as suspects. They require the returns flow to run against records: an authorization before the goods move, a receipt before the money moves, and a match before a deduction stands.
Frequently asked questions
What is return fraud?
Return fraud is any return made under false pretenses to obtain a refund, credit or replacement the customer is not entitled to. Common forms in apparel include wardrobing (wearing an item and returning it as unworn), returning a different or counterfeit item in place of the original, empty-box returns, returning stolen merchandise, and refund claims for items that were never sent back. The NRF and Happy Returns 2025 Retail Returns Landscape report, published October 2025, found that 9% of all US retail returns are fraudulent.
What is wardrobing?
Wardrobing is buying a garment with the intention of wearing it once and returning it for a full refund, typically for an event, a photo or a trip. The item comes back worn, sometimes with tags reattached, and often cannot be resold at full price. It is one of the most common forms of return abuse in apparel because clothing can usually be worn once without visible damage.
Is bracketing return fraud?
No. Bracketing, which is ordering the same item in several sizes or colors and returning the ones that do not fit, is legitimate use of a returns policy, not fraud. It is still expensive, because every bracketed order generates guaranteed returns with full shipping and processing cost. Most brands treat bracketing as a fit and merchandising problem to reduce with better size guidance, not as fraud to punish.
What is a returnless refund?
A returnless refund is a refund issued without requiring the item to be shipped back, usually when the cost of return shipping and processing exceeds the recoverable value of the item. Used deliberately on low-value orders it saves money, but it is also a fraud target: customers learn which claims trigger a keep-it refund and repeat them. Brands typically cap returnless refunds by order value, claim type and per-customer history.
How do brands detect return fraud?
At two points. Returns portals and order management rules catch patterns before the item ships back: return velocity per customer, repeat claims, mismatched claim reasons and serial-returner scores. Warehouse inspection catches what patterns cannot: the wrong item in the box, signs of wear, missing tags, and quantity short of the claim. Fraud control works best when both layers feed one record, so the refund or credit is settled against what was actually received, not what was claimed.
Does tightening return policies hurt conversion?
It can. The same NRF and Happy Returns 2025 report found 82% of consumers treat free returns as a major consideration when buying online. The practical approach is to keep the policy generous for the majority and apply friction selectively: inspection before credit on flagged accounts, returnless refunds capped by history, and clear documentation requirements on high-value claims. Blanket policy tightening trades fraud losses for lost sales.