Easy returns are good for sales. Customers buy more when they know they can send an item back. Unfortunately, generous returns also invite abuse, and the cost lands on margins that are already thin once shipping, handling and resale discounts are counted. Industry surveys regularly put return fraud and abuse in the billions of dollars across retail, though estimates differ and should be treated with caution.
The aim of prevention is not to make returns hard. It is to keep the rare bad actor from exploiting a policy built for the honest majority. This guide describes the common schemes, the signals that suggest them, the controls that work and how to treat customers fairly. It builds on returns automation, refund rules to automate and order fraud rules for small stores.
The common schemes
| Scheme | What happens | Typical signals |
|---|---|---|
| Wardrobing | Buying, using once and returning as if unused | Returns shortly after events, signs of wear, tags removed |
| Item swap | Returning a different, cheaper or counterfeit item in the box | Weight or serial number mismatch |
| Empty or partial box | Claiming a return and sending nothing or less | Parcel weight below expected; refund requested before arrival |
| “Item not received” claims | Saying a delivered parcel did not arrive, then keeping it | Delivery proof contradicts; repeated claims across orders |
| “Not as described” abuse | Claiming a fault that does not exist | Pattern across many orders; vague reasons |
| Serial returners | Returning most of what they buy, repeatedly | High return rate over many orders |
| Stolen-goods returns | Returning shoplifted items for refunds or store credit | No matching order; in-store returns without receipts |
| Receipt or label fraud | Forged receipts, reused labels or fake returns labels | Mismatched order data |
| Price arbitrage | Buying at a sale price, returning where the item cost more | Return value above paid price |
| Refund plus chargeback | Getting a refund and also disputing the charge | Duplicate recovery; see chargeback prevention |
Some of these are organised fraud and some are opportunistic behaviour from otherwise ordinary customers. The same controls address both, but the response differs: a fair warning for a customer who returns a lot of clothing, a hard stop for an item swap with evidence.
Keep it in proportion
Before adding controls, measure. What share of returns look abusive, and what do they cost? Take a sample of recent returns, classify them honestly and put a cost on each: lost goods, shipping both ways, handling, lost resale value. If abuse accounts for 1 percent of returns, heavy controls could cost more in lost goodwill than they save. If it is 8 percent, they will pay. Without a measurement, the risk is either overreacting or doing nothing.
Prevention at each stage
| Stage | Control | Notes |
|---|---|---|
| Policy | Clear return window, condition rules, who pays for shipping, restocking fees where lawful | Check local consumer law before restricting returns |
| Before purchase | Accurate descriptions and size guides | Prevents honest “not as described” returns |
| Return request | Require an order match, reason code and, for high-risk cases, photos | Verify the customer is the buyer |
| Return label | Provide tracked, unique labels | Reduces fake labels and gives proof of return |
| In transit | Hold the refund until the carrier scan, or until receipt for risky cases | Instant refunds suit trusted customers |
| Receiving | Check item, serial or batch, weight and condition; photograph | Evidence for disputes |
| Refund decision | Rules plus human review above value or risk thresholds | Partial refund for used goods where allowed |
| After | Update customer risk record and learn from patterns | Feed back into rules |
Verify at receiving
Receiving is the best control because it deals with the physical item. Compare what arrived with what was sent: product, size, colour, serial number or batch code where applicable, accessories and packaging. Weigh the parcel against the expected weight. Take photographs of the contents and condition, linked to the return record. Where the item is valuable, a short inspection checklist, with a clear pass, partial or fail outcome, speeds decisions and gives evidence. Make sure staff know what to do with a suspected swap: stop, photograph, record and escalate, rather than confront.
Risk scoring returns
A return risk score combines signals into a simple indicator that decides how much friction to apply. It need not be sophisticated: a points system is enough to start.
| Signal | Points (example) |
|---|---|
| Return rate over 50% across several orders | 3 |
| Several “not received” claims | 3 |
| High-value item, new account | 2 |
| Refund requested before label scan | 2 |
| Mismatch between billing and shipping, with other flags | 1 |
| Long history of orders without issues | -3 |
| Verified customer with many kept orders | -2 |
Define actions by score band: low risk gets instant refunds or exchanges, medium risk gets refund on carrier scan, high risk gets refund after inspection and a person reviews. Weights should be tuned to your data, and scores should never be the only basis for refusing a legitimate return. Keep humans involved in anything adverse, and make sure the rules do not discriminate on protected characteristics. See AI ecommerce return fraud prevention and fraud screening.
Dealing with serial returners
- Define a threshold from your data, for example a return rate and a minimum number of orders.
- Review each case. Look at reasons: sizing issues in a size-sensitive category are different from repeated “faulty” claims.
- Start with a friendly message and help, such as a size guide or fit consultation.
- Then apply limits that are clearly stated in your policy: store credit instead of refunds, restocking fees, or requiring inspection before refund.
- As a last resort, decline future orders, following local law and your terms.
- Record the decisions and the reasons.
Fairness and law
- Consumer-protection law in many places gives customers rights to return or cancel in certain situations, which a store policy cannot remove.
- Do not accuse without evidence. Use neutral wording, such as “we were unable to accept this return”.
- Apply rules consistently and document decisions.
- Handle personal data lawfully when building risk profiles; tell customers in your privacy notice and allow review of automated decisions where required.
- Avoid blanket penalties on whole groups, such as a country, without justification.
Store credit, exchanges and other softer options
Not every control needs to be a barrier. Offering an exchange or store credit as the first choice in the return flow keeps revenue in the business and filters out some abuse, since credit is less attractive to a refund fraudster. Offering free exchanges for size problems reduces refunds from honest customers. For consumables, offering a replacement instead of a return avoids shipping low-value items back. Where law gives customers a right to a refund, make sure the refund option is clearly available; credit and exchange should be offers, not obstacles.
| Option | Effect on abuse | Effect on honest customers |
|---|---|---|
| Exchange first | Reduces cash refunds | Convenient for size and colour changes |
| Store credit | Less attractive to fraud | Acceptable if optional and fair |
| Keep-it refund for low-value goods | Removes shipping and handling cost | Pleasant surprise; use with care to avoid exploitation |
| Instant refund for trusted customers | Needs a good risk score | Fast and loyal-making |
| Refund after inspection | Strong for high-risk cases | Slower; explain clearly |
Measure the result
| Metric | Why |
|---|---|
| Return rate overall and by product | Baseline and problem products |
| Share of returns flagged and confirmed abusive | Accuracy of rules |
| Cost of abuse | Goods lost, shipping, handling |
| Cost of controls | Inspection time, friction, lost sales |
| Customer complaints about returns | Signal of unfair friction |
| Time to refund for trusted customers | Keep it fast where risk is low |
A worked example
A store with 2,500 orders a month has a 14 percent return rate, or 350 returns. A sample review suggests 5 percent of returns, about 17, are abusive, costing an average of $48 each in lost goods, shipping and handling, around $820 a month. It introduces tracked labels, a receiving checklist with photographs and a points score that holds refunds for the riskiest 6 percent until inspection. Abuse falls by about half, saving roughly $410 a month, against about 5 hours of extra inspection time, or $100. A number of customers on hold complain, and the store adjusts the thresholds. The net gain is about $300 a month, plus better evidence for disputes. The figures are illustrative; the main message is to size the problem before spending on it.
Mistakes to avoid
- Punishing everyone for the behaviour of a few.
- No inspection of high-value returns.
- Instant refunds for every account.
- Opaque rules that frustrate honest customers.
- Relying on a score alone for adverse decisions.
- Not recording evidence, which weakens any dispute.
Sources and further reading
Figures and rules in this guide come from the primary pages below. Platform limits and prices change, so open the page before you build on a number.
- FTC: Business Guidance, consumer-protection guidance for businesses
- Shopify Help Center: return rules and policies, managing returns in Shopify