Strategy
Return policies have shifted from a race to the most generous toward thoughtful designs that reduce abuse without alienating good customers. Done well, return policy becomes a brand signal and competitive moat, not just a cost line.
For most of the last decade, return policy in online fashion was an arms race toward the most generous offer: free returns, longer windows, no questions asked, no fees. The implicit logic was that conversion drives growth and friction kills conversion. The market rewarded retailers who removed friction, and the market punished retailers who reintroduced it. That model is breaking.
The shift is partly economic (free returns are expensive when they reach 30 to 40% of orders), partly logistical (return processing capacity has not scaled with volume), and partly behavioural (a small but persistent share of customers have learned to game generous policies). What has emerged in the last two years is a different conversation: return policy as a thoughtful design surface, where the goal is not maximum generosity but maximum clarity, fairness, and brand alignment. Retailers who do this well are using return policy as a customer-experience differentiator and a brand signal, not just managing it as a cost.
This article covers how the conversation has changed, what customers actually want from a return policy (which is different from what they say they want), and the design choices that turn return policy from a cost centre into a competitive moat.
The original generous-policy model emerged in the 2010s under specific economic conditions: cheap logistics, venture-funded growth, and consumer adoption of online fashion that was still climbing. Free returns were a customer-acquisition tool more than a cost line, because the marginal cost was small and the marginal customer was scarce.
Three things changed. Logistics costs rose, particularly cross-border. Venture funding for retail tightened, requiring proven unit economics rather than market-share investment. Consumer behaviour matured, with customers ordering multiple sizes of the same item with no intention of keeping more than one ("bracketing"), shifting return rates from a side effect to a structural feature.
The industry response, through 2024 and 2025, has been a quiet but broad reintroduction of friction. A substantial majority of large European fashion retailers now charge for returns in some form, restrict return windows to 14 to 30 days, or use store-credit defaults to discourage refund-seeking. Zara, H&M, and most of the major chains have moved in this direction. Customer reaction has been less negative than predicted, partly because the changes were gradual and partly because the new norm was easier for customers to accept than retailers feared.
The useful conclusion is that the previous decade's calibration was wrong, not that retailers have to revert to consumer-hostile policies. The right calibration is more thoughtful: friction in the right places, generosity where it earns trust, clarity throughout.
The most common misreading of customer research is that customers want generous return policies. They do, marginally, but what they actually want is clarity. Clarity outranks generosity once a baseline of fairness is in place. Surveys consistently show that customers tolerate fees, restrictions, and shorter windows when expectations are clear at checkout. They tolerate them poorly when expectations are unclear, when policies feel arbitrary, or when changes happen mid-relationship.
A few specific findings worth noting:
In short: customers respond to predictable systems. Generosity is one input. Clarity, consistency, and fairness matter at least as much.
Return policy is, in customer-experience terms, a behavioural design problem. Small choices about defaults, framing, and friction have outsized effects on outcomes.
Defaults matter. A return form that defaults to "refund to original payment" produces different behaviour than one that defaults to "store credit (with 10% bonus)". The legal options are the same in both cases; the customer choice distribution is different. Many retailers now use defaults to nudge toward outcomes that are better for both customer (faster value back) and retailer (lower cash refund pressure).
Time windows. A 14-day return window encourages prompt action: the customer decides whether to keep the item soon after receiving it, rather than letting it sit in the closet for months. A 90-day window encourages forgetting, then a wave of returns when the customer finally remembers. The shorter window is often better for both parties, particularly for fashion where seasonal timing matters.
Photo evidence requirements. Asking customers to photograph items before shipping the return reduces fraud (item swap, used and returned) without alienating honest returners. Most customers find it a reasonable request once it is explained. The retailers who avoid the requirement out of fear of friction tend to absorb the fraud cost instead.
Friction in the right places. Some friction is good. A two-click return process that makes returning trivially easy correlates with higher return rates than a five-click process. The five-click process is not customer-hostile; it is calibrated. The right level of friction filters borderline returns without rejecting genuine ones.
Good return-policy design treats these levers as tools, not problems. The goal is not minimum friction; it is the right friction for the brand and the customer mix.
A quietly powerful tactic, increasingly visible at brands confident in their products, is to publish return data alongside product listings. "94% of customers keep this item". "Most returns of this dress cite sizing, consider the size guide". "Average customer kept this for 18 months before reviewing".
This sounds counter-intuitive: surely surfacing return information depresses conversion? In practice, it does not. It reduces returns, increases trust, and signals confidence. Customers who know they are buying from a retailer that publishes uncomfortable data assume good things about the rest of the operation. Customers who self-select better with this information return less.
The practical implementations vary. Some retailers surface aggregate return rates. Some surface specific reason categories ("runs small" flagged at the size selector). Some publish customer fit reviews directly on product pages, alongside ratings. The common thread is treating customer-volunteered data as an asset to surface rather than a liability to hide.
The brand signal here is significant. A retailer who hides return data signals one thing about itself; a retailer who surfaces it signals something different. The customers who notice and value the difference are typically the higher-lifetime-value cohort.
Most retailers think of their highest-LTV customers as a homogeneous group. They are not. Within any loyalty cohort, return behaviour varies widely. Customers with low return rates are disproportionately profitable, and recognising that explicitly is increasingly common.
The operational pattern is differential return policies for different loyalty tiers, calibrated against return behaviour rather than purely against spend. A high-loyalty customer with a sub-15% return rate gets longer return windows, free returns, and other accommodations. A high-loyalty customer with a 70% return rate gets the standard policy, or in some cases account flags that limit further bracketing.
This is more sophisticated than the older approach of punishing serial returners. It rewards good behaviour explicitly, makes the relationship feel more genuine for the customers who deserve it, and removes the awkward conversation around "why am I being limited?" by simply not extending the perks rather than removing them.
The data infrastructure to do this well is not trivial. Most retailers using this approach run it through their CRM with return-rate fields integrated alongside spend and frequency. The technical work is real but well-understood at this point.
For most retailers, the return is the final brand interaction in any given purchase cycle. The customer has paid, received, decided, and is now wrapping up the transaction. Whether the experience is dignified, smooth, and on-brand at that point determines a lot about whether they come back.
The high-leverage areas:
Communication during processing. Refund cycles often run two to four weeks. "Where is my refund?" is the most common returns-related customer-service contact. Proactive status updates ("your return has reached the warehouse", "your refund has been issued") reduce inbound contacts and signal that the retailer is on top of the process.
Refund speed. Faster refunds are nearly always better than slower ones, even when slower is the operational default. Some retailers issue partial or instant refunds on receipt of the return label scan, before the warehouse has processed the item. The fraud risk is real but smaller than feared.
Tone of policy language. Return policies are often written in a defensive corporate tone that sounds hostile even when the actual terms are reasonable. Plain-language policy, written for clarity rather than legal CYA, lands much better. "We will refund items returned within 30 days in unworn condition" is fine. "Returns will be evaluated at our discretion and may be refused if criteria are not met" is unnecessarily corporate.
Last-mile experience. Drop-off locations, label printing, packaging materials. Each of these can either feel friction-free or feel like the retailer pushing cost onto the customer. Investing in the last mile of returns is not glamorous, but it is one of the highest-leverage places to differentiate.
The broader principle is that the return is part of the brand, not a separate function. Treating it that way is increasingly what differentiates retailers customers come back to from retailers customers buy from once and forget.
Return policy is one of the few customer-experience surfaces almost every retailer touches and almost none use as a strategic differentiator. The retailers who treat it as design, not just policy, increasingly stand out.
Fashion retailers face return rates of 30 to 40% in Europe. Here are the most effective ways to bring those rates down: better sizing data, accurate photography, honest descriptions, smarter review surfacing, and operational changes to the return process itself.
Last updated:
Most retailers undercount what a return actually costs. Once depreciation, opportunity cost, write-off, and customer-service overhead are properly attributed, returns are typically 2 to 3 times more expensive than spreadsheet defaults suggest.
Last updated:
