About 11.4% of return value was flagged as high risk for fraud in 2026, according to Loop Returns’ Global Ecommerce Report, an analysis of 23.4 million returns across more than 4,000 Shopify merchants collected between November 2024 and October 2025. Two other widely cited studies land on different numbers for the same question: Appriss Retail and Deloitte measured 15.14% of returns as fraud or abuse in 2024, and NRF’s 2025 survey found 9%.
None of the three figures is wrong. They measure different things, on different populations, with different instruments. Knowing which one applies to a given claim matters more than picking a single number to repeat.
What percentage of returns are fraudulent?
There is no single answer, only three credible ones. Loop Returns’ 2026 Global Ecommerce Report flags 11.4% of return value as high risk for fraud, with an average flagged return worth $120, drawn from 23.4 million returns across 4,000-plus Shopify merchants. Appriss Retail and Deloitte’s Consumer Returns in the Retail Industry report, now in its seventh year, measured confirmed fraud and abuse at 15.14% of returns in 2024. NRF’s 2025 Retail Returns Landscape survey, based on shoppers and ecommerce professionals reporting their own experience, put the figure at 9%.
Figure 1: Three separate measures of return fraud, none directly comparable to the others. Source: Appriss Retail and Deloitte (2024 data), Loop Returns (2026 report), NRF (2025 survey).
A journalist quoting “the” return fraud rate should name which study and which year the number comes from. The gap between 9% and 15.14% is not evidence that fraud fell or rose by six points, it is evidence that three organizations asked three different questions.
Why do return fraud rate estimates range from 9% to 15%?
Each figure answers a different question about a different group. Loop Returns’ 11.4% is the share of return value its AI system flagged as high risk, a pre-decision signal, not a confirmed outcome. Appriss Retail and Deloitte’s 15.14% combines confirmed fraud with policy abuse (such as wardrobing) inside point-of-sale data from 60-plus of the top 100 US retailers, paired with a survey of 150 retail executives and 1,000 consumers. NRF’s 9% comes from self-reports: a summer 2025 survey of 2,006 consumers who had returned an online purchase and 358 ecommerce professionals at merchants with $500 million-plus in revenue.
| Source (year) | Rate | What it measures | Basis |
|---|---|---|---|
| Loop Returns (2026) | 11.4% | Return value flagged high risk by an AI system | 23.4M returns, 4,000+ Shopify merchants |
| Appriss Retail / Deloitte (2024) | 15.14% | Confirmed fraud and policy abuse | POS data, 60+ top-100 US retailers |
| NRF (2025) | 9% | Self-reported fraud, shopper and merchant survey | 2,006 shoppers, 358 ecommerce professionals |
Warning
Do not add or average these three figures, and do not treat a lower number as proof fraud is declining. Flagged risk, confirmed fraud, and self-reported behavior sit at different stages of the same funnel: a return can be flagged by Loop’s system, never confirmed as fraud by a retailer, and never show up in NRF’s shopper survey at all. Each figure is correct for what it measures and unreliable as a stand-in for the other two.
Has the return fraud rate risen or fallen?
Under Appriss Retail and Deloitte’s consistent point-of-sale methodology, the only apples-to-apples comparison available, the rate climbed from 13.7% of returns in 2023 to 15.14% in 2024. NRF’s own 2023 press release warns that its 2022 figures cannot be compared to later years because of a methodology change, and Loop Returns’ 11.4% is new for 2026 with no prior-year figure under the same measurement, so this article does not chart a false trend line across all three sources.
Figure 2: Return fraud reporting milestones by source and year. Sources: Appriss Retail/Deloitte (2023, 2024), NRF (2025), Loop Returns (2026). Each year uses a different measurement, not one continuous series.
The direction that holds across the comparable years is upward: fraud and abuse grew as a share of returns from 2023 to 2024 even as total return volume held roughly flat. Whether Loop’s 11.4% flagged-risk figure continues climbing in 2027 will only be answerable once the same report publishes a second year under the same method.
Fraud versus bending the rules: where is the line?
Not every gray-area return is fraud, and not every flagged return is confirmed fraud either. NRF’s 2025 survey found 45% of consumers say it is acceptable to “bend the rules” when returning an item, and nearly two-thirds admit to at least one costly return practice such as wardrobing or bracketing. That is a much larger group than the 9% to 15.14% fraud-rate figures above, because attitude and self-reported habit are not the same measurement as a confirmed fraud finding.
Figure 3: How a broad attitude toward “bending the rules” narrows down to a confirmed fraud finding. Source: synthesized from NRF (2025) attitude data, Loop Returns (2026) flagging data, and Appriss Retail/Deloitte (2024) confirmation data.
A shopper who wardrobes one outfit for a single event is a different risk profile than an account that repeats the pattern on every large purchase, which is exactly the distinction AI fraud-scoring tools like Loop’s are built to draw before a human reviewer gets involved.
Which return tactics drive the fraud rate most?
Retail executives and shoppers describe the problem from opposite sides of the counter. Appriss Retail and Deloitte’s 2024 survey of 150 retail executives found wardrobing, wearing or using an item then returning it as new, cited as a significant concern by 60%. NRF’s 2025 survey of shoppers themselves found overstated return quantity reported most often, at 71%.
| Tactic | Reported by | Share |
|---|---|---|
| Wardrobing | Retail executives (Appriss/Deloitte, 2024) | 60% |
| Overstated return quantity | Shoppers (NRF, 2025) | 71% |
| Empty box / “box of rocks” | Shoppers (NRF, 2025) | 65% |
| Decoy or counterfeit item returned | Shoppers (NRF, 2025) | 64% |
The two lists do not describe separate problems, they describe the same behaviors from the retailer’s side and the shopper’s side. A tightened return policy with a clear condition-of-return clause gives fraud-scoring systems, and the staff who review their flags, a documented standard to check either version of the claim against.
Does the fraud rate vary by channel?
Neither Appriss Retail nor NRF publishes a fraud rate broken out by channel, but the return rate feeding into that fraud pool is far from even. Appriss Retail’s 2024 data puts the online return rate at 24.52%, against 8.72% in-store, and buy-online-return-in-store transactions now account for 52% of return dollars. The broader online return picture, including year-over-year trend and category benchmarks, is covered in ecommerce return rate statistics for 2026.
Figure 4: Return rate by purchase channel, the pool fraud-detection systems screen. Source: Appriss Retail, 2024 data.
A higher return rate is not automatically a higher fraud rate. What the online-to-in-store gap does mean is that fraud-scoring systems built mostly on in-store, receipt-based signals have to work harder on mail-back and buy-online-return-in-store returns, where proof of purchase and item condition are both easier to fake.
How are retailers turning these numbers into action?
Retailers have moved from broad policy tightening toward targeted detection. 85% now use AI-based fraud scoring, 83% tightened their return policies, and 67% require a receipt or proof of purchase, according to Appriss Retail and Deloitte’s 2024 data. That detection work sits inside a larger cost picture: Loop Returns’ same 23.4-million-return dataset that produced the 11.4% flagged-risk figure also found the average return fee has climbed to $9.04, one of several cost levers covered in how much returns cost retailers.
Figure 5: Illustrative positioning of reported return-fraud tactics by prevalence and detection difficulty, synthesized from NRF (2025) and Appriss Retail/Deloitte (2024) survey data. Not a single measured dataset, treat as directional.
A published policy is the detail that turns a flagged return into a defensible decision instead of a dispute. If your current terms do not state a return window, a proof-of-purchase requirement, and what counts as an acceptable condition, you can generate a refund policy that spells out exactly what a fraud-scoring flag, or a human reviewer, can point to.
The Bottom Line
The honest answer to “what percentage of returns are fraudulent” is that it depends which study you cite: 11.4% of return value flagged high risk (Loop Returns, 2026), 15.14% of returns confirmed as fraud or abuse (Appriss Retail and Deloitte, 2024), or 9% self-reported (NRF, 2025). All three point the same direction: return fraud is a real, growing, measurable cost, not a rounding error inside total returns. What changes between the numbers is the stage of the funnel being measured, flagged risk, confirmed fraud, or self-report, and a retailer citing any single figure should say which one and why. The practical fix does not depend on picking the “right” percentage: a clear return window, a proof-of-purchase requirement, and a stated condition clause give both AI fraud scoring and human reviewers a documented standard to check a claim against.
Frequently Asked Questions
What percentage of returns are fraudulent? It depends on which study you read. Loop Returns’ 2026 Global Ecommerce Report flagged 11.4% of return value as high risk for fraud, based on 23.4 million returns across 4,000+ Shopify merchants. Appriss Retail and Deloitte measured 15.14% of returns as fraud or abuse in 2024. NRF’s 2025 Retail Returns Landscape survey put shopper-reported fraud at 9%. The three studies measure different populations with different methods, so none of the figures should be swapped in for another.
Why do return fraud rate estimates range from 9% to 15%? Each figure answers a different question. Loop Returns’ 11.4% measures the share of return value an AI system flagged as high risk, not confirmed fraud. Appriss Retail and Deloitte’s 15.14% combines confirmed fraud with policy abuse in point-of-sale data from 60-plus large retailers. NRF’s 9% comes from a summer 2025 survey of shoppers and ecommerce professionals reporting on their own experience. Flagged risk, confirmed fraud and abuse, and self-reported behavior are three different measurements of the same underlying problem.
Has the return fraud rate gone up or down? Under Appriss Retail and Deloitte’s consistent point-of-sale methodology, the rate rose from 13.7% of returns in 2023 to 15.14% in 2024. Loop Returns’ 11.4% flagged-risk figure is new for 2026 and has no prior-year baseline yet under the same methodology, so it should not be read as a decline from 15.14%.
Which return tactics drive the fraud rate most? Retail executives cite wardrobing, wearing or using an item then returning it as new, as the top concern at 60%, per Appriss Retail and Deloitte’s 2024 executive survey. Shoppers themselves most often report overstated return quantity (71%) and empty-box or “box of rocks” returns (65%), per NRF’s 2025 survey.
Sources and References
- Loop Returns. (2026). “2026 Global Ecommerce Report: Retention Benchmarks.” Analysis of 23.4 million returns across 4,000+ Shopify merchants, November 2024 to October 2025.
- Appriss Retail. (2024). “Appriss Retail Annual Research: Fraudulent Returns and Claims Cost Retailers $103B in 2024.” Seventh annual Consumer Returns in the Retail Industry report, with Deloitte, based on POS data from 60+ of the top 100 US retailers and a survey of 150 retail executives and 1,000 consumers.
- National Retail Federation. (2025). “2025 Retail Returns Landscape.” Survey of 2,006 consumers who returned an online purchase in the past 12 months and 358 ecommerce professionals at merchants with $500M+ revenue, conducted summer 2025.
- National Retail Federation and Appriss Retail. (2023). “NRF and Appriss Retail Report: $743 Billion in Merchandise Returned in 2023.” Explicitly notes 2022 figures are not comparable due to a methodology change.
- Digital Commerce 360. (2025). “Retailers continue battling fraudulent and abusive returns in 2024.” Channel-level return rate breakdown from Appriss Retail and Deloitte data.
Note: All figures verified as of August 2026. Return fraud figures are refreshed at least twice a year, and the methodology differences between the Loop Returns, Appriss Retail/Deloitte, and NRF tracks are flagged in-text and should be re-checked before each refresh.
