Returns now cost retailers nearly $890 billion, with $100 billion in fraud alone. As holiday orders surge, real-time data and automation are retailers’ best defense.
Retailers are bracing for tariffs, higher freight rates, and continued pressure on labor as they head into the holiday season. However, there’s a quieter cost center that has grown just as fast: returns.
According to NRF data, nearly 17% of all retail sales were returned in 2024, totaling almost $890 billion in merchandise. And not every return represented an honest mistake or a legitimate issue. At least $100 billion was tied to fraudulent or abusive behavior, ranging from item-not-received scams to deliberate “wear and return” tactics; the actual figure is likely higher, given that few retailers have complete visibility into the true scope of return fraud and abuse.
The holiday season amplifies the risk—between flash promotions, extended return windows, and thinned-out review teams, bad actors know the system is strained. Leading retailers must utilize real-time data, automation, and more intelligent risk segmentation to prevent fraud before peak-season orders begin to arrive.
Also Read: How Explainable AI Builds Trust in Data Decisions
The Scope of The Problem
Return losses aren’t driven by professional fraudsters alone; they’re also being fueled by ordinary shoppers adapting to a changed shopping environment.
- Automation has expanded the attack surface: self-serve portals, instant refunds, and QR drop-offs have removed friction for honest customers while opening the door for fraudsters who know refunds often arrive before verification is complete.
- Policy leniency is being exploited: In one NRF study, 45% of shoppers said it’s acceptable to “bend the rules” when returning items. Once seen as loyalty builders, free returns and extended windows have changed expectations entirely, transforming “no questions asked” policies into perceived permission for wardrobing, cross-account returns, and “keep it” manipulation.
- The resale loop is fueling arbitrage: Peer-to-peer marketplaces have made it easier for refunded goods to reappear for sale within hours, turning convenience into a gray-market economy. In the UK, for example, a report found that serial returners (representing just 11% of shoppers) are responsible for nearly 25% of total online returns.
Why Retail’s Old Playbook No Longer Works
For years, retailers have relied on policies and manual checks to manage fraud. Unsurprisingly, those defenses have begun to buckle under their own weight.
- Static rules don’t scale: Legacy fraud systems are built on binary logic: flag or approve, refund or deny. But one-size-fits-all policies inevitably miss the nuance, frustrating loyal customers while clever abusers learn how to skirt the edges of policy language and timing. Retailers also can’t treat fraud like a finite threat with a static list of “bad actors” when, in reality, return abuse is fluid and evolves with every policy update and the changing economy.
- Manual reviews drain resources: Every suspicious claim demands outreach between customer service, logistics, and finance before a refund can be cleared. Retailers can spend up to 21% of an item’s value just to process a return — a figure that can reach 1.5x the order’s worth once shipping, labor, and restocking are included.
- Data silos block visibility: Fraud detection can’t be effective when the signals are scattered. Payment data lives with finance, order data with operations, and customer histories with support. That separation hides repeat offenders who use multiple accounts, addresses, or payment methods to avoid detection.
- The refund-speed paradox: Consumers expect near-instant refunds, particularly during the holiday season. But the faster the refund, the narrower the verification window. Every day shaved off a processing timeline may improve customer satisfaction, but it also accelerates losses when a return turns out to be fraudulent.
How Retailers Can Fight Back Before Peak Season
With the holidays right around the corner, there’s no time to waste. Even small steps taken now can make a significant difference once the return season gets into full swing. Here are four practical ways retailers can get ahead.
Also Read: The Hybrid Office: Why APAC’s Tech Is Failing Workers
Segment Risk, Not Customers
Static policies treat every shopper the same, but real-time fraud scoring brings nuance back into the process. By analyzing variables such as order value, return frequency, and fulfillment history, retailers can categorize shoppers into dynamic trust tiers (e.g., loyal, neutral, or high-risk) and adjust policies accordingly.
A loyal customer, for example, might qualify for instant refunds year-round, while higher-risk profiles are routed through extra verification. The point isn’t to block returns, but to apply trust more intelligently, so your best customers never feel the guardrails that protect the business.
NEST NEW YORK, a fragrance brand with extreme gifting spikes, is an example of why adaptive trust tiers are worth it — by approving more good orders while filtering unauthorized reseller patterns, the brand saw 50% fewer chargebacks last Q4.
Automate trust decisions
Once those trust tiers and rules are defined, automation brings them to life. Manual reviews can’t keep pace with peak-season volumes. Still, AI verification tools can apply those decisions instantly — cross-checking order, payment, and tracking data before a refund ever leaves the account. These tools flag anomalies in real-time, from repeated ‘item-not-received’ claims to SKU swaps and mismatched addresses, while automatically greenlighting legitimate requests.
Caraway implemented this shift when manual fraud checks were straining support and delaying shipments as demand grew. After automating approval and review workflows, monthly fraudulent transactions were reduced by half, more than $75,000 in chargebacks were prevented, and the support team saved over 115 hours to focus on genuine customer needs.
Close the loop with fulfillment and CX data
Not every red flag signals fraud. Late shipments, missing packaging, or unclear product instructions can all trigger the same refund disputes that fraud teams are trained to identify.
That’s why the smartest retailers now connect their fraud engines with fulfillment and CX data to spot the difference between deception and dysfunction. Fixing these root causes upstream reduces friction for customers and false positives for fraud teams.
Also Read: The Code Problem: Why Connected Cars Are Prime Cyber Targets
Audit your blind spots before the rush
Even the strongest fraud models rely on clean data, but most retailers discover their weak points only after volumes surge. Before the holidays, run a quick audit across order, refund, and fulfillment systems to identify broken links, such as SKUs without tracking visibility, carriers missing delivery confirmations, or policies that automatically approve refunds after a set number of days.
Tightening these gaps now means fewer losses later. The key is having technology in place that can automatically carry out these defenses, so teams aren’t left chasing edge cases by hand when volume spikes.
