Ecommerce Returns Report 2026: Rates, Reasons and Rules

The table lists the published return rates in the research that count goods, for the US, Germany, the Netherlands, the UK and the EU, with the year the data covers. There is no single average: NRF's latest US figure is 19.3% of online sales for 2025, Bamberg's latest measured German rate is 24.2% of parcels for 2021, and Dutch parcel statistics show 11.4% for 2025^1,2,31.

A typical retailer is not the same as the market. In EHI's 2026 survey of 108 DACH retailers, 66.7% reported an item return rate of 10% or less, while 23.6% reported more than 20%¹². The market rate is pulled up by high-volume fashion sellers: fashion accounted for 83% of returned shipments and 91% of returned items in Germany in 2021².

Countries where surveys count shoppers

In the Nordic countries and Poland the published figures are shares of people who returned anything within a time window, not shares of goods.

The National Retail Federation now publishes its returns report with Happy Returns, a UPS company. For 2025, merchants estimated that 15.8% of annual sales and 19.3% of online sales would be returned, a total of $849.9 billion, and expected 17% of holiday sales to come back^1,5. The survey covered 2,006 consumers and 358 professionals at merchants with revenue over $500 million, in summer 2025⁵.

Before 2024, NRF worked with Appriss Retail. The 2021 and 2022 editions surveyed 57 and 70 retailers and called their findings directional rather than scaled to the industry^34,33. The 2023 edition switched to Appriss returns data from customers that include 60 of the top 100 US retailers, paired with a survey⁸. Later NRF tables print a dash for 2023 because the method differs⁹.

Is there a 2026 NRF returns report?

Not as of 17 October 2026. NRF's sitemap, with entries up to 16 October 2026, lists no 2026 returns report⁶. The Happy Returns resource index, sitemap and press page list no 2026 edition^7,50,51. Past editions came out between October and January, so a 2026 edition may still appear^{35,34,48,49,32,1}.

Other US estimates

Appriss Retail's 2026 Total Retail Loss report puts 2025 US returns at $706 billion, built from Appriss data, external benchmarks and a survey of 1,020 consumers¹⁰. It splits that value into 52% bought and returned in store, 29% bought online and returned in store, and 19% bought and returned online¹⁰. Coresight Research found an average online apparel return rate of 24.4% of orders from a March 2023 survey of 100 US apparel decision-makers³⁶.

Germany has the most detailed public return data in Europe, mostly from the Forschungsgruppe Retourenmanagement at Bamberg University. Its 2021 European Return-o-Meter is the latest measured market rate; the parcel counts for 2024 to 2026 apply that rate to newer parcel volumes and are not new measurements^2,11.

The Return-o-Meter compares Germany with a pooled rest-of-Europe average and does not publish rates for other individual countries². Bamberg links the gap to retailer policy and payment habits: in 2021, 88.7% of German retailers offered free returns against 52.4% elsewhere, the average return window was 53.5 days against 29.6 days, and 28.8% of German orders were paid by invoice against 9.9%^2,52. An earlier Bamberg survey of 538 people aged 14 to 29 found fashion parcel return rates of 55.65% when paying by invoice and 30.15% with prepayment³⁸.

German shoppers' own estimate is much lower. In a Bitkom survey published in February 2023, online shoppers said they send back about one in ten online purchases: women 12%, men 8%, and 26% never return anything³⁹.

The EHI Retail Institute publishes distributions rather than averages. In its 2026 survey of 108 DACH retailers, 18.7% reported rising return rates, up from 14.7% a year earlier¹². Just over 80% of the 124 retailers in the 2025 survey did not expect return volumes to improve over the next three years⁵⁴.

Netherlands (retourpercentage)

Thuiswinkel.org, the Dutch ecommerce association, reports 11.4% of 615 million B2C parcels returned in 2025 and 10.9% in 2023³¹. For 2022 it cites the ACM parcel monitor: 12.3% of 473 million parcels⁴⁰. Its 2024 figure is inconsistent: it states 11.7%, but its own counts (57.8 million of 609 million) give 9.5%³¹.

EU

Eurostat publishes no return-rate indicator. The closest proxy is that 10% of EU online shoppers received wrong or damaged goods in 2025⁵⁵. The European Environment Agency estimates that about 20% of clothing bought online is returned and about 30% of footwear (not only online), and that online return rates can be up to three times those of physical stores⁴⁴.

United Kingdom

IMRG's most recent retailer survey in the research, of 41 retailers, found an average return rate of 13% for January to June 2022¹⁶. Clothing retailers ranged from 5% to 45%, while most home and garden retailers were below 5%¹⁶. Retail Economics and ZigZag forecast that UK non-food online returns will ease from 21% in 2024 to 19.5% in 2025, worth GBP 25.1 billion against GBP 26.7 billion, without stating whether the rate counts value, units or orders⁴¹. In an IMRG and nShift survey of 1,000 UK consumers in 2026, 85.6% said the returns policy matters when deciding to buy online⁵⁶.

Nordic countries

PostNord's 2026 survey of 4,000 consumers found that 28% had returned something in the past three months¹⁵. Young shoppers return far more: in 2024, 54% of Swedes aged 18 to 29 had returned something, against 17% of those aged 50 to 64⁴⁵.

Poland

In the 2026 E-commerce w Polsce survey, 46% of online buyers had returned at least one item in the previous 12 months and 29% had returned more than once¹⁴. Women (51%) returned more often than men (42%), and 72% of returners used parcel lockers^14,57.

France

FEVAD publishes no return rate on its public pages. Its 2026 key figures report EUR 196.4 billion of ecommerce revenue in 2025 without a return figure⁵⁸.

Fashion has the highest rates wherever online categories are compared, but the figures below use at least five different metrics. Read the metric column before comparing two rows.

No source in the research gives a numeric online return rate from 2023 or later for beauty, toys, jewellery or footwear on its own. The closest figures are IMRG's 2020 UK benchmark (health and beauty 4%, footwear 20%) and the EEA footwear estimate^42,44.

Size and fit leads in the fashion-focused sources (Narvar, Coresight's apparel survey and Zalando)^3,36,17. Across all categories the picture is less clear: Optoro gives a figure only for items that differed from their description, and Klarna's top two reasons are within one point of each other^18,40.

Bracketing

Bracketing means buying several sizes or colors to return most of them. Merchants told NRF that 23% of 2025 purchases included a bracketed item, rising to 26% in apparel and footwear¹. Happy Returns found that 51% of Gen Z shoppers bracket, against 24% of Baby Boomers⁶³. Narvar reports that 72% of shoppers always or sometimes buy online knowing they will return something⁶⁴.

Size advice

Fraud and abuse

Consortium data from Happy Returns and Wyllo show 1 in 64 orders over $500 involving return fraud, and a fraud rate for mail-in returns 6 times that of box-free in-person returns, a method Happy Returns sells⁷⁰.

Under the EU Consumer Rights Directive, consumers have 14 days to withdraw from an online purchase without giving a reason, counted from physical possession of the goods⁷¹. The trader refunds within 14 days of being told, including standard delivery, but may wait until the goods or proof of sending arrive⁷¹. The consumer pays only the direct return cost, and only if the trader said so beforehand and did not agree to pay⁷¹.

The withdrawal function

Directive 2023/2673 adds a labelled function to withdraw from online contracts⁴. In Germany, § 356a BGB requires a function labelled "Vertrag widerrufen" followed by a confirmation step, with a receipt on a durable medium⁷⁷. Denmark uses the labels "Fortryd aftale" and "Bekræft fortrydelse"⁹⁷. French official guidance describes fines of up to EUR 75,000 for companies that fail to inform consumers of the withdrawal right⁸³.

Destruction of returned goods

The Ecodesign for Sustainable Products Regulation (ESPR) defines an unsold product to include items returned under the withdrawal right²⁹. Ten derogations apply, including damage that cannot be repaired at reasonable cost and donation offers refused after at least eight weeks¹⁰⁵. None of the 12 simplification omnibus packages published up to 24 June 2026 names the ESPR¹⁰⁶.

Several large fashion retailers charge for postal or drop-off returns: Zara charges EUR 2.95 at drop-off points in Germany, the Netherlands and France, Uniqlo EUR 2.95 in Germany, Next GBP 2.50 and boohoo GBP 1.99 in the UK^{24,112,113,114,25,26}. The marketplaces bol and Otto and the electronics retailers Coolblue and MediaMarkt offer free returns within 30 days in the Netherlands and Germany^{27,28,115,116}.

How many retailers charge

What retailers disclose

The annual reports checked give no absolute return rate. Zalando's 2025 annual report gives no rate, and ASOS and Next report only changes or margin effects^142,143,144.

Return transport is the cost driver most often named by DACH retailers, by 70.6% of them¹². In the US, NRF found $145 million of returns for every $1 billion of sales in 2023⁴⁹.

For the US, Optoro and Environmental Capital Group estimate (around 2021) about 16 million tonnes of CO2 and 5.8 billion pounds of landfill waste a year from returns, without a published method¹⁵¹.

Published return rates disagree for seven main reasons.

1. Different bases. A rate can count parcels, items, orders or sales value. Bamberg's definitions show the same orders giving 66.67% on a parcel basis and 50% on an item basis, and a third definition counts value^152,153,154.
2. Shoppers or goods. Nordic, Swedish and Polish surveys count the share of people who returned anything over 1, 3 or 12 months^46,15,14.
3. Who reports. NRF's 2024 and 2025 rates are estimates by large merchants, while its 2023 rate came from transaction data plus a survey^1,32,8.
4. Typical retailer or whole market. EHI reports what share of retailers fall in each band; Bamberg's cluster rates are merchant means, while its 24.2% headline is market-weighted^12,2.
5. Method changes inside one series. NRF changed method in 2023 and 2024^8,9. Sweden's E-barometern monthly share went from 8% (2021) to 13% (2024), probably because the question changed^155,46.
6. Extrapolation. German parcel counts for 2024 to 2026 hold the 2021 rate constant¹¹.
7. Errors and restatements. Thuiswinkel's 2024 Dutch rate does not match its own counts, and an NRF blog cites 17.3% online for 2023 against 17.6% in NRF's own release^31,156,49.

The research combined twelve lanes: US studies, European studies, category figures, return reasons, legal framework, retailer policies, cost and environmental impact, and five gap-filling lanes. The dataset holds 495 sourced values from 175 sources; this page cites the subset listed in the sources.

• Primary first: we use the publisher's own report or the statute wherever we could reach it. For example, we cite NRF's 9% fraud figure rather than a co-publisher's 8%, and the German Federal Law Gazette rather than a commercial law site^1,157,78.
• Verification: every row was re-fetched and checked against its source on 17 October 2026. 421 values were confirmed as stated and 74 were corrected, mostly dates, scope and attribution. Rows that could not be verified are not shown.
• Vendor research from Salsify, Akeneo, Optoro, Narvar, Happy Returns and Coresight's sponsored study is labelled as such.
• Several full reports are paid or form-gated (EHI, PostNord country data, FEVAD, IMRG, Narvar), so some figures come from the publisher's press release or summary.
• Refresh: a full re-check every quarter, and monthly checks from October to January for the next NRF and Happy Returns edition. Past editions appeared between 11 January 2021 and 15 October 2025^35,1.