Reducing the ecommerce return rate: three brands with numbers

🇹🇷 Türkçe: Bu yazının Türkçesini oku →

The uncomfortable line for any commerce director in a quarterly review is not the traffic slide or the conversion slide. It is the one that shows how much of what got shipped came back. In the US, the National Retail Federation and Happy Returns forecast that consumers would return roughly $849.9 billion of merchandise in 2025, or 15.8% of retail sales. Online is worse: the same research puts the internet return rate at 19.3%. That is not a warehouse problem. That is a fifth of the top line, moving in the wrong direction, before you account for the cost of processing it.

Boards are starting to notice, and the reasonable question they ask a supplier is not “what is the industry doing about returns” but “what would you actually change on our store to move the number”. I have watched a dozen versions of that conversation in the last year, and the honest answer is that reducing the ecommerce return rate is not a single project. It is three projects that share a workshop, and most stores are only running one of them.

So rather than write a lecture, I want to walk through three brands that have each moved on one of the three levers and reported a real number afterwards, then say what a mid-market WooCommerce store would copy from each on Monday.

The three: ASOS, changing what the shopper sees about their own return history before checkout. Zalando, collecting better data from the shopper up front so recommendations get honest. Snipes, testing whether an in-page fit finder changes behaviour on a footwear catalogue where returns have historically been the price of doing business.

ASOS moved the return rate by telling shoppers what their own return rate was

ASOS’s first-half fiscal 2026 result carries a number that is easy to miss in the profit narrative. In the 26 weeks to 1 March 2026, ASOS reports a +160 basis point year-on-year improvement in the underlying returns rate, attributed to what it describes as a new returns policy underpinned by greater transparency and customer education. The interim results filing, published 23 April 2026 under Chief Executive José Antonio Ramos Calamonte, credits the policy with roughly one percentage point of underlying return-rate benefit for every month it has been active.

The mechanism is worth looking at because it is behavioural, not technological. In December 2025 ASOS launched a returns transparency tool inside the app that shows each customer their own historical return rate, alongside guidance on how to shop more consistently. There is no new fit AI in this story. There is no size scanner. There is a mirror.

• +160 basis points year-on-year improvement in the underlying returns rate, H1 FY26.
• Roughly +1 percentage point improvement in returns rate per month the policy has been active.
• 30% year-on-year increase in profit per order over the same window.
• Returns transparency tool launched in the ASOS app, December 2025.

Our first half shows continued progress on executing our strategic priorities across Relevant Fashion Product, Inspirational Shopping Experience and an Efficient Operating Model.

José Antonio Ramos Calamonte, Chief Executive Officer, ASOS plc (H1 FY26 interim results, 23 April 2026)

The transferable lesson is not that transparency is a returns strategy on its own. It is that ASOS accepted the political risk of telling a segment of their shoppers something uncomfortable about their own behaviour, and it converted, at scale, into profit-per-order growth. The 30% figure is doing a lot of work in that filing: it says that avoided returns are not lost sales, they are retained margin, and the finance function can see the difference. Any store that has been quietly building a “serial returner” flag in the CRM and doing nothing with it has this move available today.

Zalando moved the return rate by asking better questions before the sale

Zalando’s route to the same number is the opposite temperament. Where ASOS educates the customer after the fact, Zalando collects better information before there is a fact. On 20 July 2023 Zalando’s corporate site announced size recommendations based on the customer’s own body measurements: two photographs of the shopper in well-fitting clothing feed a model that estimates their measurements and matches them against the catalogue’s size grid, brand by brand.

The reported figure is narrow but real. Items purchased with size advice have a 10% lower size-related return rate than items purchased without it. That is not a marketing claim: Zalando gives it as a controlled comparison inside its own catalogue. The technology behind it came from Fision, the Swiss body-measurement startup Zalando acquired in 2020, plus its own internal fit modelling.

• 10% reduction in size-related returns for items with size advice vs items without.
• Recommendation model built on two shopper photographs plus brand-level size grids.
• Underlying body-measurement technology acquired with Fision (2020).
• Launched to European markets, 20 July 2023, corporate.zalando.com.

Helping customers find the right fit is an important part of deepening customer relationships. It is frustrating to wait for an item to then discover it does not fit.

Stacia Carr, VP Size & Fit, Zalando (corporate.zalando.com, 20 July 2023)

The transferable lesson is that the ceiling of a recommendation engine is set by the quality of the data going into it, not by the sophistication of the model. Zalando’s advantage here is not that they run a fancier algorithm than a mid-market retailer could buy off the shelf. It is that they invested in the front-end interaction that produces cleaner inputs. A store using a generic “S/M/L guessed from past orders” model is being outrun by a store that spent six weeks getting the size intake screen right.

Snipes moved the return rate by putting the fit conversation on the product page

The third lever is closer to the ground and, for a WooCommerce merchant, the most immediately copyable. Snipes, the German streetwear and footwear retailer, ran a two-month controlled A/B test on snipes.de with an in-page fit recommendation widget from Fit Analytics against a static size chart, published as a Fit Analytics case study. The test cohort was large — 645,000 shoppers — which matters because footwear return behaviour is noisy and small samples flatter whichever side you want to believe.

• 645,000 shoppers in the A/B test on snipes.de.
• 2% reduction in the return rate for the Fit Finder cohort.
• 8% reduction in size sampling (multiple sizes of the same item added to basket).
• 3% uplift in conversion rate for shoppers using Fit Finder.

We were looking for a solution that could reduce size sampling while protecting growth. Fit Finder delivered on both counts.

Dennis Scheen, Head of Digital Platform Development & Operations, Snipes (Fit Analytics case study)

Two per cent on the return line does not sound like much next to Zalando’s ten or ASOS’s 1.6 points. Read the second number instead. An 8% cut in size sampling is a 8% cut in the “buy three, keep one, return two” pattern that eats footwear economics alive. That behaviour looks like conversion from the front of house and looks like a returns disaster from the back. The reason to care is that Snipes ran this against a control group, not against a memory of last year, which is a bar most retailers do not clear when they claim a returns win.

The transferable lesson is procedural. Snipes did not commit to fit technology as a strategy. They tested it on a subset large enough to be honest, measured three variables that could move independently, and let the numbers decide. The right question to ask a supplier proposing a fit widget is not “does it work” but “what happens to your case study when I ask for a 645,000-shopper control”.

Three brands, three levers, one thing they share

Look at the three interventions next to each other and the shared attribute is not the technology. ASOS runs a behavioural tool inside its app. Zalando runs a computer-vision measurement flow. Snipes runs a size recommender on a product page. What they share is that in each case the retailer accepted a small amount of friction at the front of the funnel — one more screen, one more question, one uncomfortable summary of the shopper’s own record — in exchange for a measurable change further down. Every store that has quietly agreed with itself that “any friction kills conversion” has taken this move off the table without deciding to.

The second shared attribute is that all three brands published a control. ASOS separates the returns-policy benefit from the wider trading result and gives a per-month figure. Zalando compares items with size advice against items without. Snipes ran an A/B test with a real cohort size. On a subject where vendor decks are wall-to-wall self-serving numbers, that discipline is doing more work than the technology.

What we would build inside a WooCommerce store this quarter

We work on WordPress and WooCommerce, so the practical translation matters. Here is what an equivalent programme looks like inside a real store rather than at an all-hands. None of it requires a platform migration.

• A returns intelligence view in the customer account, week one. Every WooCommerce order is already in the database with an order status and a set of line items. A small custom endpoint plus a My Account tab can show a customer their own 12-month return rate, the categories they most often return, and a plain-language nudge. This is the ASOS move at a WooCommerce store’s scale. Expect 3–5 working days for a first version if your data model is clean, longer if past returns were tracked in a shipping plugin rather than as WooCommerce order statuses.
• A serial-returner segment that leaves the store and enters the ESP, week two. The same underlying view feeds a Klaviyo, Mailchimp or ActiveCampaign segment. The segment is not a punishment list; it is an audience that gets fewer “restock alert” prompts and more “here is why this fits” content. Two hours of work if the ESP connector is already installed, most of a sprint if it is not.
• Product data quality before you buy any fit technology. The Zalando result is a data-quality story dressed as a technology story. Before you commission a fit widget, audit the WooCommerce product attributes: are the size tables consistent across brands, are the fit attributes structured or free-text, is there a canonical “measurements” section on every variable product. Fixing this is unglamorous, and it is why fit tools underperform on stores that skipped it.
• An honest A/B test on one intervention, one quarter. Pick one lever from the three above and run it against a control on 30–50% of the catalogue. Set the primary metric before you start and do not switch it once you have seen the data. This is where most agency engagements go soft; we would rather ship a smaller test that survives audit than a bigger one that does not.
• A single dashboard that finance and merchandising both read. Every one of these interventions is a returns-and-margin story, not a returns story. Wire the WooCommerce data to a shared view that pairs the return rate with the retained margin, so the number that shows up in the board pack is the one ASOS quoted: profit per order. If the internal argument ends up being “returns are marketing’s problem”, nothing moves.

A note on order of operations. Every buyer we speak to wants to start with the technology because it is the tangible thing on the vendor slide. The three examples above suggest the opposite order: start with the intake data quality, then the behavioural signal to the shopper, then the model on top. If we were quoting this work for a mid-market fashion or footwear brand on WooCommerce today, that is the order we would sequence it in, and we would tie payment to the metric your finance director already tracks.

The last thing worth saying is what did not move any of the three numbers above. None of these brands changed their return policy from generous to punitive. ASOS added transparency, not friction. Zalando added a measurement flow, not a restocking fee. Snipes added a recommendation, not a rule. The lever that is loudest in industry commentary — charging shoppers to return — is absent from three of the strongest published results of the last two years. If your board is being pitched a returns strategy that leads with a fee, this is worth putting in the room.

If any of this is a live problem inside your organisation, we would rather look at it against your actual return data than argue in the abstract. Get in touch and we will start with the intake screen, not the pitch.