Returns Fraud

Returns fraud is the deliberate abuse of a retailer's returns and refund process to obtain money, goods, or credit a customer is not entitled to. It is the loss-prevention layer of Returns Management and overlaps with — but is distinct from — returns abuse (policy-stretching behaviour like Bracketing (Fashion Returns)|bracketing and Wardrobing|wardrobing that may be technically permitted). Sources disagree on where "abuse" ends and "fraud" begins, which is the single biggest reason headline loss figures diverge. Fashion and footwear are repeatedly named as the highest-exposure categories because of sizing-driven return volume, high resale value, and worn-once-then-returned behaviour.

Firewall: every claim is what a source reports. See ../../CONTEXT.md Rule 1. Most quantitative figures below come from vendors selling returns/fraud software (Riskified, Signifyd, Loop, ZigZag, Appriss, Happy Returns) — flagged inline; each has an incentive to make the problem look large.

Scale & cost

NRF and Happy Returns project total US retail returns of $849.9B in 2025, with 19.3% of online sales returned, and found that 9% of all returns are fraudulent (as-of 2025-10-15). The same survey found 45% of shoppers think it acceptable to "bend the rules" when returning. A Deloitte/Appriss Retail survey puts fraudulent returns and claims at a $103B loss to US retailers (as-of ~2024–2025; Appriss primary not directly fetched).

Riskified (vendor), analysing >1M refund claims, reports refunds run at 1–2% of total sales dollars with nearly 1 in 4 refund dollars linked to abuse (as-of ~2026-01).

Fraud taxonomy

What sources report retailers track:

• Empty-box / "box of rocks" — returning a package with bricks or worthless items; Appriss (vendor) reports a 65% increase in retailers noting this (as-of ~2026, via search summary).
• Overstated quantity — claiming more items returned than were sent; Appriss reports 71% of tracking retailers noting increases (as-of ~2026, via search summary).
• Decoy / counterfeit returns — returning a fake in place of the genuine item; Appriss 64% (as-of ~2026, via search summary).
• Item-Not-Received (INR) / false claims — Riskified (vendor) names INR the most-abused category, 25% more likely to be fraudulent than "Missing Items" claims (as-of ~2026-01).
• Receipt manipulation — stolen, forged, or AI-generated/altered receipts; abuse of receipt-free policies to "return" stolen goods. [Appriss/Happy Returns explainers, vendor]
• Gift-card fraud — purchase with promotional/stolen gift cards, then return for cash or store credit. [vendor explainers]
• Cross-retail / cross-channel returns — buy at one retailer and return to another with a laxer policy; Buy Online Return In Store (BORIS)|BORIS and BORO combined account for >$362B in returns (Signifyd, vendor, as-of 2025).
• Policy abuse — Wardrobing (wear-and-return) and Bracketing (Fashion Returns)|bracketing sit on the abuse↔fraud boundary; UK self-report ~16% wardrobing, ~42% bracketing (via search summary, low confidence).
• AI-generated damage claims (emerging) — submitting AI-generated photos of "damaged" goods that actually arrived intact; flagged by Boll & Branch and described as "spreading fast." [PYMNTS, 2026-03-03]

Riskified (vendor) also reports velocity/value risk signals: claims filed in the first 7 days are >20% more likely to be abusive; orders over $1,000 are 33% more likely; orders over $2,000 show 2.5× higher claim rates than orders under $100 (as-of ~2026-01).

UK / EU & fashion-specific

Per Retail Economics, UK retailers lose £1.3B/yr to returns fraud, with apparel/footwear abuse rates reaching up to 15% of total returns (via FashionUnited/Loop, vendor-authored, as-of 2025). 64% of UK brands told Loop rising returns fraud had the biggest financial impact on their business (as-of 2025).

Serial Returners: ZigZag (vendor) found UK serial returners fell 12%→8% of consumers in 2025 — a group that cost retailers £6.6B — saving a forecast £1.7B (as-of 2025-10).

UK legal constraints (stable): Consumer Contracts Regulations 2013 (14-day cancel + 14-day return ≈ 28-day minimum window), Consumer Rights Act 2015 (30-day reject-faulty window). Retailers cannot refuse statutory returns but may inspect and deduct for wear/misuse; AI fraud-detection is constrained by UK GDPR / Data Protection Act 2018 (transparency, access, appeal). [FashionUnited/Loop, 2025-05-16]

Detection & prevention

What sources report retailers doing:

• Return fees / withheld delivery refunds — 76% of the UK's 100 largest retailers charge fees or withhold delivery refunds, mostly under £3 (Retail Gazette, as-of 2025-10). 66% of fashion retailers charge for at least one return type (via search summary, low confidence).
• Risk-based / differentiated returns — Riskified "Dynamic Returns" treats trusted vs high-risk shoppers differently in real time; 85% of retailers reportedly deploy AI to detect return fraud and 84% of executives changed return policy in the past year (via search summary, low confidence).
• AI inspection at the warehouse — a UPS subsidiary deployed AI-based inspection to spot counterfeit/fraudulent returns in the 2025 holiday season [Reuters 2025-12-18].
• Account closures / blocklisting — Asos closed accounts under its fair-use policy (and faced backlash); Next introduced high-frequency-returner measures (Retail Gazette, as-of 2025-10).
• Returnless Refunds — refund-without-return to cut reverse-logistics cost on low-value items; a returnless-refund fraud-detection market is sized at $380M (2025) → $1.5B (2036), 13.5% CAGR (FMI, vendor sizing, low confidence).

Key terms

Term	Meaning
Returns abuse	Policy-stretching that may be technically permitted ([[Bracketing (Fashion Returns)
Empty-box	Returning a package with worthless contents to claim a refund
INR claim	"Item Not Received" — Riskified's most-abused false-claim category
Wardrobing	Wearing an item once then returning it as unworn
BORIS/BORO	Buy-online-return-in-store / -online; cross-channel return paths exploited for fraud
Returnless refund	Refund issued without requiring the item back

Benchmarks (as-of 2025–2026)

Figure	Value	Source	Confidence
US returns total 2025	$849.9B	NRF/Happy Returns	high
Online return rate 2025	19.3%	NRF/Happy Returns	high
Returns that are fraudulent	9%	NRF/Happy Returns	high
US fraudulent return/claim loss	$103B	Deloitte/Appriss (via secondary)	med
Refund dollars linked to abuse	~1 in 4 (25%)	Riskified (vendor)	high (vendor)
UK returns-fraud loss	£1.3B/yr	Retail Economics (via Loop)	med
UK apparel/footwear abuse rate	up to 15% of returns	BRC (via Loop)	med
UK serial returners	8% (down from 12%)	ZigZag (vendor)	high (reported)

Gaps

The two practitioner streams were both down this run: Reddit MCP not connected and Apify YouTube transcript actor not connected — so no operator counter-narrative on real return-counter fraud, blocklist effectiveness, or staff burden. No 2024–2026 source tying RFID / hidden anti-wardrobing tags to fraud prevention was fetched. No primary data on returnless-refund abuse rates or fashion-specific practice. EU (ex-UK) figures absent — context is UK-only. Several headline figures (Appriss 2026 percentages, Signifyd $362B, FMI sizing, NRF/Deloitte 85%/84%/66%) are from search summaries, not direct fetches. Primary Appriss Retail and McKinsey reports not directly fetched. Running UNIQLO-Europe gap — no UK/EU fashion-primary worked returns-fraud case with internal metrics.