The biggest theft problem in retail does not run out the door. It walks in through it, stands in line, and asks for a refund.
In 2025, US retailers took back USD 849.9 billion in merchandise — 15.8% of all sales. Nine percent of those returns were fraudulent (NRF / Happy Returns, 2025). That is roughly USD 76 billion, and it is the industry's own low estimate.
Two years earlier, NRF and Appriss Retail measured fraud and abuse at nearly 14% of returns — USD 101 billion, up 20% in a single year. The measurement moves. The scale does not.
Either figure is several times any defensible estimate of organized retail crime (Edition #03). One of those problems has task forces, federal legislation, and conference keynotes. The other has a policy PDF and nobody's name on it.
The counter is a register running in reverse
The fraud is not exotic. It is industrialized versions of behavior your returns desk approves every day:
71% of retailers report increases in overstated-quantity returns, 65% in empty-box returns, and 64% in decoy returns — counterfeit or substitute items handed back in place of the product (NRF / Happy Returns, 2025).
Claims and appeasements are the fastest-growing abuse channel — "item not received," "arrived damaged" — refunds issued with no merchandise handled at all, and bad actors accumulating credits at scale (NRF / Appriss Retail).
19.3% of online sales come back, and shoppers aged 18–30 made 7.7 online returns in twelve months, the most of any generation (NRF / Happy Returns, 2025).
The culture has moved with the numbers: 45% of consumers now say "bending the truth" to make a return is acceptable, and close to two-thirds admit to at least one costly returns behavior — wardrobing, bracketing, returning a different item (NRF / Happy Returns, 2025). This is not a crime ring. It is a norm.
Nobody owns the number
A USD 76–101 billion loss category with no owner is an organizational choice. Look at the wiring:
Nearly half of retailers run separate departments for in-store and online returns (NRF / Appriss Retail) — while the fraud deliberately crosses channels: buy online, return in store, claim online. The silo is the vulnerability.
AP is staffed against the exit. Apprehensions, ORC cases, exception reports at the register. The returns desk reports to operations or customer experience — functions whose KPI is speed, not scrutiny.
85% of retailers now point AI at return fraud. Only 45% find it effective (NRF / Happy Returns, 2025). Tooling without ownership produces dashboards, not outcomes.
Compare the treatment. A shoplifter walking out with USD 200 of product gets a case file, possibly a prosecution. A serial refunder extracting USD 20,000 a year in appeasement credits gets an apology and a coupon.
The policy is the accomplice
The loss is not leaking around your controls. It is flowing through your policies, as designed:
82% of shoppers call free returns a major purchase consideration, up from 76% a year earlier (NRF / Happy Returns, 2025). CX won the policy argument, and the loss line pays the invoice.
Return fraud never books as fraud. It disappears into returns reserves, appeasement expense, and damaged-goods write-offs. The P&L structure hides the crime better than any fraudster could.
64% of merchants say updating their returns process is a priority in the next six months (NRF / Happy Returns, 2025). That redesign window is when AP either gets into the room or spends five more years outside it.
A perspective from the field
The most expensive dishonest customer I have encountered in this business never concealed an item. She returned things. Store credit, gift receipts, different branches, always inside policy — approved every time by our own people, in our own system, at a counter we had staffed for speed.
By the time the pattern surfaced, the total was a multiple of what our best-documented theft crew had taken. The exit had cameras, sensors, and a guard. The counter had a smile and a service-time target. We had built a fraud machine and given it a customer-service logo.
Three practical moves for the next 90 days
Merge the data before you merge the org. Build one identity-keyed view of returns, claims, and appeasements across store and online. Cross-channel fraud is invisible in siloed reports, and nearly half the industry is siloed by design. The first output — top 50 identities by refund value — usually pays for the project.
Put returns fraud on the AP scorecard with a named owner. One leader, a dollar target, and exception reporting on refund-to-sales by store, return rate per identity, and claims rate by channel. This is the same discipline exception-based reporting brought to sweethearting twenty years ago, pointed at the counter.
Price the policy, not just the fraudster. Quantify what no-receipt returns, instant refunds, and appeasement thresholds cost per quarter, and hand merchandising and CX the price tag. Policies change faster than criminals do — but only when someone shows what they cost.
Closing note
None of this argues for hostile returns. Free, fast returns win customers, and 71% of consumers say one bad returns experience drives them away (NRF / Happy Returns, 2025). The policy is a commercial weapon. The question is who audits the weapon.
Retail spent this decade's outrage on the front door. The larger number was behind the industry the whole time, at the counter, moving politely through a process nobody owned. Claim it before the next budget cycle prices it as normal.
I want to hear how your organization assigns returns and claims fraud today — who owns the number, if anyone does. Reply with anything you can share, anonymized if you prefer.
Forward this to one LP or AP leader who should be reading it.
— Gabriel
The LP Brief is a weekly intelligence read for senior loss prevention and asset protection leaders. Free. No vendor noise.
Not yet subscribed? thelpbrief.com