What is return fraud? Types, cost and how retailers reduce it

Return fraud happens when someone deceives a retailer to obtain a refund, exchange or store credit. An item comes back used, swapped or stolen, a refund is claimed for goods that were delivered, or a return reason is falsified to win a refund the policy would refuse. It happens inside the returns process, and that is where retailers can catch it.

Return fraud often comes to light after the refund has been issued, when the warehouse opens the box and finds an item that is worn, incorrect or missing. The only record is the return reason the customer gave. Losing less starts with asking better questions earlier in the returns process, recording the answers, and issuing refunds when the evidence checks out, inside the deadlines consumer law sets.

What is return fraud?

Return fraud, also called returns fraud, occurs when someone intentionally uses the returns process to obtain money or goods they are not entitled to under the retailer's policy. The customer knows the item does not qualify, or that nothing is coming back at all, and uses the process to get paid anyway. Intent to deceive separates fraud from an ordinary return that happens to be inconvenient.

The difference between return fraud, refund fraud and return abuse

Refund fraud, often confused with return fraud, is a subset of it: the refund without the return. The customer claims the parcel never arrived, reports an empty or partial box, or raises a chargeback through the card issuer while keeping the goods. Because no genuine item comes back, the warehouse has nothing to inspect. Proof of delivery and claim approval are the controls that work here.

Return abuse is different in kind: it follows the letter of the return policy but goes against its intent, and it involves no deception. Examples include buying four sizes and returning three, ordering a full outfit to try at home and keeping none of it, or returning items from the same category month after month. Each return is allowed on its own; the pattern is the problem, and the retailer decides where to draw the line through policy design and return-rate monitoring.

A retailer that offers free returns for 60 days without checking item condition has set very few boundaries around what counts as abuse, while one that records item condition, requests photos for "faulty" claims and inspects occasion wear when it arrives has drawn a line it can enforce. Where abuse starts is the retailer's own call, made through its policy and its return-rate thresholds, and good returns management is the main control it owns for both fraud and abuse.

The main types of return fraud and abuse (and what each looks like in your returns data)

Each type leaves different clues in data the returns team already has: return reasons, condition codes, timing, SKU return rates, claim frequency and the warehouse receipt scan. Two of these types, bracketing and serial returning, are abuse rather than fraud, and are included because they appear in the same data and respond to the same policy levers.

Wardrobing

Wardrobing is buying an item, using it once, and returning it as unworn. It usually relates to clothing, according to the University of Portsmouth report.

What you see: certain SKUs with return rates well above the category average, spikes in returns after holidays, wedding season or long weekends, and items returned as "changed my mind" that arrive with scuffed soles, missing hygiene strips or a trace of perfume. Google searches for the term spiked in July 2024, a reasonable prompt for when to step up inspection of occasion wear.

What reduces it: recording item condition when the return is requested, requiring photos for the return reasons that justify it, inspecting items in the affected categories when they arrive, and holding the refund for those categories until the item has been inspected. When a customer cancels under the statutory withdrawal right, European consumer law also makes them liable for the loss in value on goods handled beyond what is needed to check them, provided the retailer gave the withdrawal information.

Bracketing and serial returning

Bracketing is ordering several sizes or colors with the intention of keeping one. Serial returning describes a customer who consistently returns a much higher share of purchases than other customers. Both are forms of return abuse that cost money to serve.

What you see: orders containing multiple sizes where most items are returned, customers with return ratios in the top few percent, and return reasons related to fit or color rather than defects.

What reduces it: these are return policy and product information issues before they are anything else. Fees for return shipping where the law allows and the customer was told before purchase, free returns only above a minimum order value, and shorter windows outside campaign periods change the economics for the heaviest returners. Using return-reason data to improve product pages (size guides, color accuracy, fabric descriptions) removes one of the reasons customers bracket in the first place. It will not remove all of them, since some shoppers will always prefer trying three sizes at home to reading a chart.

Receipt, price and label fraud

The customer returns an item against a different order, returns a sale item for a full-price refund, or sends back a cheaper lookalike while keeping the more expensive one. With label tampering, the customer redirects the parcel to a false address so that a carrier scan triggers the refund before anyone opens the package.

What you see: value mismatches when returns arrive, a returned SKU that does not match the order line, carrier scans with no matching warehouse receipt.

What reduces it: requiring customers to start returns from an existing order, generating labels through the returns portal instead of printing them in advance, and triggering refunds on the warehouse scan rather than the carrier's first scan. Consumer law sets the outer limit on that last point. In the UK, where the retailer has not offered to collect the goods, the refund is due within 14 days of the retailer receiving them back or, if earlier, the customer supplying evidence of sending them. In the EU, the 14 days run from the customer's withdrawal notice, and, unless the retailer offered to collect the goods, it may hold the refund until the goods arrive or the customer shows evidence of sending, whichever comes first. Either way, the inspection window is short. In the UK, the customer's evidence of sending, such as a carrier drop-off receipt, can bring the deadline forward; in the EU it ends the retailer's right to hold the refund.

Empty box and "item not received" claims

The returned parcel is unusually light or empty, or the customer claims the original delivery never arrived. Research by the University of Portsmouth for Cifas lists "claiming items did not arrive" and "reporting partial or empty boxes" among the most common refund-fraud methods.

What you see: claims with no associated return shipment, repeated claims linked to the same address or email, and delivery records showing a signature or a doorstep photo for a parcel the customer says never came.

What reduces it: attaching carrier proof of delivery and signature data to the claim, requiring manual approval on claims before any refund is issued, and checking package weight when high-value returns arrive.

Item substitution and counterfeit returns

The customer returns a different item, an older model, or a counterfeit in the original packaging.

What you see: the item's condition does not match the outbound record, its serial number differs from the one shipped, or an item marked "faulty" turns out not to be one you sold.

What reduces it: scanning each returned item against its original shipment record, and issuing refunds after inspection for the SKUs where substitution pays. A mismatch at receipt shows that something changed. Finding out when takes more records: for high-value items, the outbound serial record, the carrier's tracking events and the parcel weight at each scan let the retailer tell a substituted return from a parcel tampered with in transit, and decide whether the claim goes to the customer or the carrier.

Stolen goods and organized reshipping

Organized groups use stolen card details to make purchases, redirect parcels and submit fraudulent claims at scale. At first it looks like a delivery problem: addresses changed after dispatch, "not received" claims concentrated in one postal code, customers who become hostile when asked to verify. The companion piece on organized fraud disguised as delivery problems covers the delivery-side signals and the cross-team response in depth.

Friendly fraud and chargebacks

Friendly fraud is a chargeback raised by the real cardholder for goods they received. It goes through the card network rather than the returns portal, and your payment provider handles the dispute tools. The returns team provides the evidence: proof of delivery, any existing return record and the customer's stated return reason.

Other types you will see named

Fraud guides name several more variants of these types, each with its own control.

• Switch fraud: an older or broken unit is returned in the new product's box. Flagged by the serial-number check at receipt.
• Bricking: a device is returned with valuable components removed. Flagged by inspection before refund on electronics.
• Open-box fraud: an opened, used item is returned as new for a full refund. Reduced by condition codes and the customer's liability for loss in value under EU and UK consumer law, where the retailer gave the withdrawal information (in the UK the retailer may deduct it from the refund).
• Cross-retailer returns and retailer arbitrage: an item bought cheaper elsewhere is returned to you at your price. Reduced by accepting returns only against an existing order.
• Price-tag switching: a higher price tag is moved onto a cheaper item, which is then returned for the higher refund. Reduced by refunding the amount on the order record, whatever the tag says.
• Gift-card fraud: a refund is pushed onto a gift card that is then resold. Reduced by refunding to the original payment method, which UK and EU law requires for statutory withdrawals unless the customer agrees otherwise.
• Employee fraud: staff process refunds to themselves or to accomplices. Reduced by role-based user permissions, the activity log on each return, and keeping refund approval and refund issuing with different people.

What return fraud costs retailers

Refund fraud could cost UK online retailers between £317 million and £5.76 billion a year, with a median estimate of £2.3 billion, according to University of Portsmouth research funded by Cifas and published on May 21, 2026. The range is modeled: the researchers applied a range of return rates to UK online sales, then applied US fraud-and-abuse rates for returns, from Deloitte and Mastercard estimates, to those returns, which is why the spread is wide and the headline £5.76 billion is the upper bound. The same study analyzed nearly 500,000 posts across cybercrime forums and Telegram and found "refund-as-a-service" vendors selling guaranteed refunds for up to 30 percent of the refund value.

The National Retail Federation projected total US retail returns at $890 billion for 2024 in its December 2024 returns report. In the same report, 93 percent of the retailers surveyed called retail fraud and other exploitative behavior a significant issue for their business.

A finance team deciding whether to fund a fix needs the headline figure broken into the costs behind it:

• The refund paid for an item that can no longer be resold at full price, or that never comes back at all.
• Return shipping. A dress worn once and sent back needs two shipments and produces one sale, at best.
• Warehouse inspection and handling time, including the legitimate returns that now need checking because the fraudulent ones were never separated.
• Lost resale value when worn, damaged or incomplete items must be sold through outlet channels or discarded.
• Customer service time on disputes. A claim with no evidence behind it costs agent time to investigate and often still ends in a refund.

A £100 fraudulent refund can cost less than £100 once stock recovery, resale value and carrier claims are netted off, and more once handling and dispute time are added. The baseline a finance lead needs is the retailer's own: refunds paid on returns later found to be worn, substituted or missing, plus the handling time on those cases, over a quarter, set against the cost of the controls that would have caught them.

What is driving ecommerce return fraud

Free, hassle-free returns help win orders: the NRF report found 76 percent of consumers surveyed who had returned an online purchase in the past year consider free returns a key factor in deciding where to shop. Fraudsters exploit that same convenience, particularly when a refund is issued as soon as a carrier scans a parcel, with no questions asked. Where returns were made easier by removing questions at the request, the evidence needed to detect fraud went with them.

Fraud has also become more organized: the Cifas research describes forums and Telegram channels where fraudsters share techniques, sell refunds as a paid service, and discuss which retailers’ procedures are weak and target them. Of the forum users who stated an age (19 of them), all were 14 to 30, with an average of 19.

Cross-border returns make inspection harder when they pass through more hands and take longer to reach the warehouse, so market-specific rules and clear condition codes keep the standard consistent across countries.

Whether the total volume of return fraud is rising, or whether retailers are now seeing more of what was always there, is a question the available research does not settle. The forum data shows organized, shared methods and the loss estimates show a large problem, so the same controls apply whichever way that question is settled.

How retailers detect return fraud

Return fraud detection starts with the data the returns process already produces: return reasons, condition codes, return timing, SKU return rates, per-customer return ratios, claim frequency by address and the warehouse receipt scan. Each one marks a case for review, and the inspection shows what happened. The most common red flags:

• A customer's return ratio or claim count far above the rest of the base.
• Repeated "not received" or "damaged" claims to one address, email or payment card.
• A refund requested within hours of delivery, or a claim raised before the parcel could have arrived.
• A return parcel whose weight at the carrier's scan does not match the item ordered.
• A condition or serial number at the warehouse scan that does not match the outbound record.
• A return reason that does not match what arrives, such as "faulty" on a unit that works at the bench, or "changed my mind" on an item that comes back worn.

Collect the data

A returns portal that captures a reason and an item condition for every returned product, along with a photo where it helps, replaces guesswork with a record. Before Hunkemöller digitized its returns, the company didn't know in advance how many returns might come back each day, or what was driving them. After rolling out nShift Returns across six European markets, the company gained richer insights into return behavior, and its warehouse teams could see expected volumes days in advance.

Make it visible

Monitoring return rates by product shows which SKUs are being returned unusually often and whether that rate has risen over the past 30 days. Exporting return reasons and condition codes by SKU to your BI tools lets a returns manager filter for "worn" or "not as described" and see the categories, customers and dates behind each.

Act on it

Approval rules flag claims and high-value returns for manual review while everything else stays self-service. The measure of whether the rules work is the share of flagged cases confirmed on inspection, against the share of legitimate refunds delayed. Track both from the first month, or the rules will look successful because they hold more returns.

Specialized fraud-detection tools check identity, device and payment signals that a returns platform does not capture, and belong on the shortlist once the returns data is reliable. The guide to choosing returns management software covers what to ask a vendor about reason capture, approvals and reporting.

How to prevent and reduce return fraud

Most of the following controls can be configured in a modern returns management system, and most add no step for a legitimate customer returning an unworn sweater within the return period. The ones that do, such as a photo on a "faulty" claim or a short wait on a flagged category during peak, belong in the returns policy so the customer knows what to expect.

Set the policy per market and per flow

Policy is the first control because it defines what counts as abuse, and two kinds of rule apply at once. In the EU, retailers cannot contract out of the statutory withdrawal period or refund deadlines, and the UK regulations place the same duties on the trader; both leave room for more generous terms on top. Everything beyond the statutory rights is a commercial choice: a longer window during a campaign and a shorter one outside it, free returns above a minimum order value, return shipping fees where the law allows and the customer was told before purchase. Keep the rules separate for each market and each process (return, exchange or claim) so customers in Sweden and the UK see a policy that reflects their local law and expectations.

A five-question check on whether the current policy invites fraud:

1. Is the refund issued at the carrier's first scan, before anyone has seen the item?
2. Does the return request ask only for a reason, with no item condition?
3. Can a "faulty" or "damaged" claim be filed without a photo?
4. Does occasion wear carry the same window and the same checks as everything else?
5. Is there a weekly view of return ratio and claim count per customer?

Each "yes" to questions 1 to 4, and a "no" to question 5, marks a gap to close first.

Capture reasons, conditions and evidence at the request

Ask customers to give a return reason and an item condition for each product, with an optional comment, and require a photo for the reasons a photo can settle, such as "faulty" or "damaged on arrival". Make the photo required or optional depending on the reason, and show only the relevant reasons for each product category, so "too small" appears for clothing and never for a book. Suspicious returns are then visible before a label is issued, and the record exists if the claim later becomes a dispute. Return-reason data also helps product teams: repeated size-related returns on one SKU usually point at the size guide, and fixing that removes one reason customers order multiple sizes.

Review flagged claims and high-value returns

Require approval for claims and high-value returns before a return label or refund is issued. During peak periods, a retailer can widen manual pre-approval for a few weeks, for example from claims only to all returns in the market where wardrobing is highest, and switch it back afterward. Outside those windows and markets, returns stay self-service, so most legitimate customers see no extra steps. The review workload scales with what you choose to flag, so start narrow, measure the confirmed-fraud rate, and widen only where the data supports it.

Inspect and record at receipt

Equip the warehouse with a scanning tool that shows item-level return data next to the original order, and photograph or film anything that fails inspection at the returns bench, filed against the return number. For categories prone to substitution or wardrobing, issue the refund after inspection, within the statutory deadline. For everything else, issue it on the warehouse receipt scan. Keep restocking status visible so finance can see which returned items are still sellable. Decide in advance who rules on a disputed condition, such as a customer claiming the damage happened in transit, and how the customer can challenge that ruling. A written procedure makes a deduction much easier to defend if the customer later raises a chargeback.

Use the in-store channel

An in-store return is inspected by a member of staff at the counter, using labor the retailer already pays for. Hunkemöller saw a 15 percent shift from warehouse returns to in-store returns after moving to digital returns, creating more opportunities for assisted service, remarketing and repeat purchases. Robin Visser, Omni Channel Business Development Manager at Hunkemöller, described the result: "We've made returns part of a seamless omnichannel customer experience with increased returns control and insights. What was a historical pain point for the company and our customers has been changed into something that adds real value."

Shoppers are heading the same way: in Sweden, store returns rose from 10 percent to 16 percent of shoppers' most recent returns between 2022 and 2026, according to PostNord's E-barometern for the second quarter of 2026. Offering the store as a return option at the point of request moves more returns into a channel where someone can look at the item, provided store staff have the order record and a short checklist for what to look for.

Use delivery evidence for "item not received"

When a customer reports a missing delivery, start with what the carrier recorded. Where the carrier provides proof of delivery or a signature, attach it to the claim before a refund decision is made. A doorstep photo, a locker pickup scan or a signature resolves many "not received" claims quickly, whichever way the evidence points. It shows where the parcel went, while who took it in is a separate question, so misdelivery and theft after delivery still need a human decision.

Know the law on worn goods in Europe

Consumers in the EU and the UK who cancel within the statutory period are entitled to a refund. They are also liable for any loss in value caused by handling the goods beyond what was necessary to establish their nature, characteristics and functioning. The EU directive gives the test in its own recital: the consumer may try on a garment but may not wear it. In the UK, the retailer may deduct that amount from the refund, up to the contract price. In EU and EEA countries the directive sets the liability and national law sets how it is recovered; Sweden and Norway both write the duty to compensate into their consumer distance and off-premises contract acts.

Two limits apply. The amount is limited to the loss in value: cleaning or repair costs, or the gap between selling the item as new and as second-hand. And the retailer's right to deduct is lost if it failed to inform the customer about the right of withdrawal before the sale; the customer's own withdrawal right survives, and the withdrawal period extends by up to 12 months. The piece on the right of withdrawal covers those information duties. Separately, sealed goods that are not suitable for return for health or hygiene reasons, and were unsealed after delivery, fall outside the withdrawal right; that exception turns on the seal and the suitability for return, whatever the product category, and rights for faulty goods are a separate matter. Condition records and photos taken when returns arrive give the retailer the evidence to calculate a deduction and defend it.

Where nShift Returns helps

nShift Returns, our returns management software, provides the branded self-service portal where customers select items, give a reason and choose return, exchange or claim. It manages the rules for each market and flow, including return periods, free returns above a minimum order value and fees by refund method or customer group. It captures return reasons and condition codes by SKU, applies approval rules for claims and high-value returns, and exports the data to your BI tools. When returned items arrive, a warehouse scanning tool shows item-level data and restocking status, with refunds and credit notes synced to your ERP.

With these controls switched on, most returns still move through self-service, and the flagged ones arrive with a reason, a condition code, a photo and a scan record attached. Talk to a returns specialist about how these rules can be applied across your markets.