Returns are a drain on profitability, and such a pervasive problem for retail that some companies have begun taking a more combative approach, banning customers for bad behavior and charging fees for returned goods. Is there a better way?

Part of the problem stems from learned behavior. Over the years, many retailers trained shoppers to expect fast shipping, long return windows and free refunds. Those policies helped drive conversion, particularly as e-commerce accelerated during the pandemic. But the cost of maintaining that level of flexibility has risen sharply, from logistics and restocking to fraud and lost full-price sales.

The 2025 holiday season, followed by the January returns wave, offered another stress test. Salesforce reports a 10% year-on-year increase in returns from the end of the 2024 holiday season to the same period in 2025. “Returns are a balance sheet issue, and we’re still in the thick of it,” says Caila Schwartz, director of consumer insights at Salesforce. At the same time, consumer behavior is shifting, with many using generative AI tools like ChatGPT and Gemini to research products, and social media for discovery, Schwartz notes. This can lead to impulse and last-minute shopping, she says, which feeds into the returns landscape: Salesforce has found that purchases made in the last four days before Christmas are most likely to be returned. The US National Retail Federation predicts customers returned $850 billion worth of merchandise in 2025.

While a flexible returns policy can rope customers in, a bad experience can turn customers off. Almost all — 93% — of consumers say a bad return experience makes them less likely to shop with a brand again, according to Salesforce. Against this backdrop, retailers are reassessing whether policy changes alone can meaningfully solve the problem. Experts say the best strategy is prevention, especially as emerging AI technologies begin to shape fit, sizing and product choice earlier in the customer journey.

Policy vs prevention

In the past few years, retailers have been tightening their previously liberal returns policies, from introducing shorter windows to charging fees, in an effort to curb abuse and recoup costs. “Consumers still desire fast shipping, free returns and lenient exchange policies, but most retailers are starting to pull back on these because they simply cannot afford to do it for all customers,” says Woodrow Levin, co-founder and CEO of returns and exchanges platform Extend. “Today, retailers do not have a way to segment their most valuable customers from policy abusers and therefore are forced to push stricter less favorable policies on everyone.”

However, this risks alienating customers. Last year, Asos unexpectedly deactivated the accounts of some shoppers with high returns rates. The backlash was swift, particularly from plus-size customers who said they often need to buy multiple sizes to find a good fit (a behavior called bracketing). Then, earlier this month, Asos introduced a more transparent tool, which shows clearly to customers that those with a return rate below 70% will continue to benefit from free returns. Meanwhile, those with a returns rate of over 70% will face a £3.95 returns fee if they keep less than £40 worth of items from their order, and customers with a returns rate of 80% or higher will also be subject to an additional £3.95 restocking fee, on top of the returns charge.

Salesforce’s data suggests that policy changes can work, but only if they are communicated clearly — and early. “Consumers are tolerant as long as they know before they make the purchase,” Schwartz says.

But both Schwartz and Levin point to prevention — rather than post-purchase penalties — as the more sustainable lever. That includes better fit tools, more accurate product data and more proactive guidance at the point of purchase. According to a Vogue Business survey of almost 700 consumers, 38% said they often return clothes because they are ill-fitting, while 91% said their clothing size changes depending on what brand they’re purchasing from.

Fit tech has long promised to solve these challenges, but most have come up short, often due to user adoption and inconsistent experiences. Plug-ins that ask customers to enter their height, weight and size, or to upload a photo to scan their body, introduce friction for the customers. However, experts suggest that AI innovations — particularly agentic AI — could be a better solution. “The innovation coming out of agentic tools allows the consumer to ask more nuanced questions about the fit, sizing and ultimately help prevent behaviors like bracketing,” she says. “We even anticipate triggered support agents, where the customer gets to the cart and an agent appears to say, ‘You’re ordering this size, but based on your purchasing history, we would recommend this size.’”

Getting descriptions, sizing and imagery right may not be glamorous, but it pays off, says PWC UK partner Laura Morroll. AI tools only help if the underlying data is sound, she emphasizes. “If you get the descriptions, the dimensions and the quality of all that right, the cost it can save downstream in returns is huge,” Morroll says. “These tools are looking for patterns to make recommendations. They’re only as good as the consistency of the quality checks on that particular style, but it’s a guide.”

The Extend platform encourages brands to move beyond top-line revenue metrics. “We are talking to merchants about measuring not just the revenue a customer generates, but also considering the cost of returns, exchanges, restocking, lost packages, multiple accounts created and more associated with a customer,” Levin says. “Once you understand your customers through this lens, merchants are empowered to encourage spending, loyalty and trust with their best customers.”

Exchanges: A missed opportunity

While refunds dominate the returns conversation, online exchanges remain underused — and often poorly designed. For shoppers, the process can be slow and fragmented. “It’s a really clunky process, isn’t it?” Morroll says. “You have to return, then buy again. You have to print out the label, or go to the parcel shop, then you wait for the other one to be picked and sent to you.”

Amazon has reset expectations by issuing refunds as soon as items are scanned at drop-off. “That speed of refund is really a key factor for customers,” Morroll says. “There are still a lot of retailers where it can take seven to 10 days for you to actually get the money back on your card.”

A worker at an Amazon fulfillment center.Photo: Michael Nagle/Bloomberg via Getty Images

From the retailer’s side, speed matters for a different reason: inventory recovery. “If I take the full length of time before I return it, that’s two weeks of the season where they could be selling it at full price,” she says. “So not only are they refunding me, but they’ve reduced their ability to sell it, and it’s more likely to be marked down.”

However, most brands won’t have the resources that Amazon does. Many now rely on third-party reverse logistics specialists to handle returns quickly and recondition items to be resold. For luxury brands, this also includes authenticity checks. Others are redesigning warehouses so returned products can be slotted dynamically rather than sent back to specific pickup locations.

Despite the friction, exchanges offer a chance to preserve revenue and deepen relationships — if handled well. “Often, retailers don’t offer exchange as the first option,” Schwartz says. “Even if the exchange is impossible and the size is sold out, how do you upsell into a different product? ‘Shoppers that bought this also bought this product over here.’ How do you keep the conversation going instead of just ending the transaction at the return?”

Salesforce data shows timing could play a role. “Most returns happen mid-week,” Schwartz says. “How can the retailer save that sale or encourage an exchange? How can they lean into agentic customer service, or train mid-week sales associates to redirect that return into something positive for both the brand and the consumer?”

In luxury, she adds, that conversation is often best handled by people, supported by technology. “It’s a delicate balance. It’s a high-touch, highly personalized experience,” Schwartz says. “Maybe that store associate has a better view of what the customer was searching for online and what we would recommend based on their purchase history. So you still have that human touch, but it’s powered by a much more personalized experience combining those journeys.”