---
title: "Ecommerce return rate: formula, benchmark and ROAS effect | Product Metrics"
description: "Return rate = returned ÷ sold, by orders, units or value. See the 2025 benchmark, what a return costs and how returns raise break-even ROAS."
canonical: "https://www.productmetrics.io/blog/ecommerce-return-rate"
pageType: article
language: en
publisher: "Product Metrics"
author: "Berend Vrakking"
datePublished: 2026-10-07
dateModified: 2026-10-08
image: "https://www.productmetrics.io/og/blog/ecommerce-return-rate.png"
---

> Content index: https://www.productmetrics.io/llms.txt

# Ecommerce return rate: formula, benchmark and ROAS effect

## Key takeaways

- Return rate = returned ÷ sold. Count it by value for ads and profit: the same month reads 8.0% by orders, 8.8% by units and 9.2% by value.
- Benchmarks count different things. US retailers expected 19.3% of online sales returned in 2025 (NRF); DACH retailers report 6% to 10% of articles, 26% to 50% in fashion (EHI).
- A return costs DACH online retailers €5 to €10 per article to handle (EHI), plus the whole product if it cannot be resold. The ad click is already paid for.
- Returns raise break-even ROAS because Google Ads counts the sale before the return. Divide break-even by (1 − return rate): 2.90 becomes 3.62 at 20% returns.

Your ecommerce return rate is the share of what you sold that comes back: **returned ÷ sold**. For ads and profit, measure it by value (euros refunded ÷ euros sold), because that is the basis your revenue, margin and ROAS are on. Counted by orders or units, the same month reads a different number.

The examples are made up but checkable by hand, and the break-even figures come from the same code as the [break-even ROAS calculator](https://www.productmetrics.io/break-even-roas-calculator), so you can reproduce every one of them there.

## 1. How do you calculate your return rate?

Divide what came back by what you sold, for the same group of orders, on one basis. Each basis answers a different question:

- **By orders:** returned orders ÷ orders placed. This is the load on customer service.
- **By units:** returned units ÷ units sold. The warehouse and stock planning live with this one.
- **By value:** refunded revenue ÷ revenue sold. This is what returns do to profit and ROAS.

Take a month with 1,000 orders, 1,250 units and €100,000.00 sold. 80 orders, 110 units and €9,200.00 come back: 80 ÷ 1,000 is 8.0%, 110 ÷ 1,250 is 8.8% and €9,200.00 ÷ €100,000.00 is 9.2%.

**Figure: One month's return rate (%), measured three ways.** Each bar is the same month's return rate on a different basis. It reads 8.0% by orders, 8.8% by units and 9.2% by value, because the returned orders were larger than average. ROAS and profit use the value basis.

| Measured by | Return rate |
| --- | --- |
| By orders | 8.0% |
| By units | 8.8% |
| By value | 9.2% |

_Source: Illustrative data: 1,000 orders (80 returned), 1,250 units (110 returned), €100,000.00 sold (€9,200.00 refunded)_

Value reads highest here because the returned orders were bigger than average. [Shopify's returns guide](https://www.shopify.com/enterprise/blog/ecommerce-returns) divides items returned by items sold. That fits stock planning, but ROAS is a ratio of euros, so the matching return rate is too.

Count exchanges by what is refunded. A like-for-like exchange moves stock and handling but refunds no euros, so it belongs in the order and unit rates and not in the value rate.

Group returns by the order date, not the refund date. A return arrives days or weeks after the sale, so a month's refunds include last month's orders, and a growing shop reads too low if it divides this month's refunds by this month's sales. Count a month's orders after their return window has closed.

## 2. What is the average ecommerce return rate?

Sources disagree, mostly because they count different things. The US benchmark is the NRF and Happy Returns [2025 Retail Returns Landscape](https://nrf.com/media-center/press-releases/consumers-expected-to-return-nearly-850-billion-in-merchandise-in-2025): retailers estimated that 19.3% of online sales would be returned in 2025, and 15.8% of all annual sales including stores.

**Figure: Share of US retail sales returned (%), all sales and online.** Each bar is the share of sales US retailers expected to come back, according to NRF and Happy Returns. Online it was 19.3% in 2025, against 15.8% of all sales including stores (16.9% in 2024). These are shares of sales, not of orders.

| Sales measured | Share of sales returned |
| --- | --- |
| All sales, 2024 | 16.9% |
| All sales, 2025 | 15.8% |
| Online sales, 2025 | 19.3% |

_Source: National Retail Federation and Happy Returns, 2025 Retail Returns Landscape, press release of 15 October 2025 (checked 7 October 2026). Retailer estimates, US_

The retailer survey covered 358 ecommerce professionals at US merchants with over $500 million in revenue, so a smaller shop with a different mix may sit far from it. For Europe, the EHI Retail Institute surveyed online retailers in Germany, Austria and Switzerland ([dpa-AFX report, 5 December 2023](https://www.boerse-online.de/dpa-afx/studie-retouren-kosten-onlinehaendler-im-schnitt-fuenf-bis-zehn-euro-334018.html)). It counts articles, not euros.

| Source | Counts | Figure |
|---|---|---|
| NRF, US 2025 | Sales | 19.3% |
| EHI 2023 | Articles | 6% to 10% |
| EHI, fashion | Articles | 26% to 50% |
| Bamberg 2021 | Packages | 1 in 4 |

A newer EHI survey of 124 DACH online retailers (summer 2025, [TextilWirtschaft, 27 November 2025](https://www.textilwirtschaft.de/business/news/neue-ehi-studie-so-haben-sich-die-retourenquoten-entwickelt-253155)) counted returns as a share of orders. 24.5% of fashion providers reported that 36% to 50% of orders come back, and 12.2% reported more than half. The Bamberg row is "almost every fourth package" from a [University of Bamberg press release of 7 September 2022](https://www.uni-bamberg.de/presse/pressemitteilungen/artikel/erste-europaeische-haendlerbefragung-retourenmanagement/).

Use the table to see how wide the spread is, then calculate your own rate by product group. Other category averages often come without a definition or a year, so this guide leaves them out.

## 3. What do ecommerce returns cost?

A return costs the margin you expected, the cost of handling it, and the ad click that brought the buyer, which was already paid for. EHI found that DACH online retailers pay €5 to €10 per returned article on average and €10 to €20 for home and furnishing, with inspecting and quality-checking the returned goods the biggest cost ([dpa-AFX, 5 December 2023](https://www.boerse-online.de/dpa-afx/studie-retouren-kosten-onlinehaendler-im-schnitt-fuenf-bis-zehn-euro-334018.html)). Add the product itself if it cannot be sold again. Your own cost per return is total return costs ÷ returned orders, read from those invoices.

Return shipping is mostly the retailer's bill. In the same study only 14% of retailers made customers pay it and almost two thirds paid it themselves. In the NRF report, 82% of consumers said free returns are an important consideration when shopping online ([NRF, 15 October 2025](https://nrf.com/research/2025-retail-returns-landscape)).

For example, an order of €100.00 excl. VAT has a product cost of €60.00 and costs €6.00 to handle when it comes back, a figure inside EHI's range. The refund cancels the revenue, so a product that goes back on the shelf costs €6.00. One that cannot be resold costs €66.00.

**Figure: What one returned €100.00 order costs in cash (€).** Each bar is the cash one €100.00 return costs. If the product is resold as new, that is the €6.00 of handling. If it cannot be resold, you also lose the €60.00 product: €66.00 in total.

| What happens to the returned product | Cost of the return |
| --- | --- |
| Resold as new | €6.00 |
| Cannot be resold | €66.00 |

_Source: Illustrative data: price €100.00 excl. VAT, product cost €60.00, handling €6.00; before ad spend and outbound shipping_

In the Bamberg survey, 93.2% of returned items were sold as new (German online merchants, data for 2020 and 2021, [same press release](https://www.uni-bamberg.de/presse/pressemitteilungen/artikel/erste-europaeische-haendlerbefragung-retourenmanagement/)), so for some shops the €6.00 case is the common one. Check your own resale share before you assume either.

## 4. Returns raise break-even ROAS

Google Ads counts the revenue before the return, while you keep the margin only on what stays. **Break-even ROAS, as reported before returns, = (1 ÷ contribution margin on kept sales) ÷ (1 − return rate).**

Take a product with a 40% margin, a 2.5% payment fee and 3% for reverse logistics. That leaves a [contribution margin](https://www.productmetrics.io/glossary/contribution-margin) of 34.5% of the revenue you keep. With no returns, break-even ROAS is 1 ÷ 0.345 = 2.90. At 20% returns it is 2.90 ÷ 0.8 = 3.62. The [margin calculator](https://www.productmetrics.io/margin-calculator) runs the margin part for one product, with its own return rate, payment fee and reverse logistics.

**Figure: ROAS needed to break even as the return rate rises.** The line is the ROAS the same product needs to break even at each return rate, with price and costs unchanged. It rises from 2.90 with no returns to 4.14 at 30% returns, 43% higher.

| Return rate (share of revenue) | ROAS needed to break even |
| --- | --- |
| 0% | 2.9 |
| 5% | 3.05 |
| 10% | 3.22 |
| 15% | 3.41 |
| 20% | 3.62 |
| 25% | 3.86 |
| 30% | 4.14 |

_Source: Illustrative data, calculated with the model of the break-even ROAS calculator: product margin 40%, payment fee 2.5%, reverse logistics 3%, returned products resold_

The calculator gives the same figures for €10,000.00 of ad revenue. Contribution is 34.5% of the revenue you keep: €3,450.00 at 0% returns and €2,760.00 at 20%.

| Returns | Kept revenue | Break-even ROAS |
|---|---|---|
| 0% | €10,000.00 | 2.90 |
| 5% | €9,500.00 | 3.05 |
| 10% | €9,000.00 | 3.22 |
| 15% | €8,500.00 | 3.41 |
| 20% | €8,000.00 | 3.62 |
| 25% | €7,500.00 | 3.86 |
| 30% | €7,000.00 | 4.14 |

> **Sidenote:** The model assumes returned products go back into stock and are resold. It also charges reverse logistics as 3% of the revenue you keep, so that cost does not grow with the number of returns, as handling does in practice. Enter the percentage that matches your own return rate. If products cannot be resold, the real break-even is higher than shown.

In profit terms the same test is POAS, which is ROAS × contribution margin, so break-even POAS is 1.0 at any return rate. [POAS vs ROAS](https://www.productmetrics.io/blog/poas-vs-roas) explains when to use which.

## 5. Your ROAS counts the sale before the return

Google Ads reports the conversion value your tag recorded at purchase. A later return changes it only if you upload a conversion adjustment. [Google Ads Help](https://support.google.com/google-ads/answer/7686280) describes two kinds. A retraction withdraws the conversion completely. A restatement changes its value.

Both need the order ID that was recorded with the original conversion. Google lists returned purchases and partial returns as use cases ([About conversion adjustments](https://support.google.com/google-ads/answer/7686447)). A retraction sets the value to 0.00 and lowers the conversion count. A restatement leaves the count unchanged.

Google's page also sets deadlines: you have up to 7 days after a conversion is first recorded for autobidding readability, and overall adjustments can be made within 54 days. A return that arrives later stays in your reported value.

So check your own account. If returns are uploaded as adjustments, the ROAS you see is already after returns, and the break-even to compare it with is the one on kept sales (2.90 in the example). If they are not, the ROAS is before returns, and you compare it with the raised break-even (3.62 at 20% returns).

## 6. Track it per product group

A shop-wide return rate hides the products that drive it. Take two groups with €3,500.00 of ad revenue each: shoes come back at 25% and accessories at 5%, so the shop reads 15%.

**Figure: Return rate (%) of shoes, accessories and both together.** Each bar is the share of a group's revenue that comes back. Shoes return 25% and accessories 5%, so the shop-wide 15% describes neither group.

| Product group | Return rate |
| --- | --- |
| Shoes | 25% |
| Accessories | 5% |
| Both together | 15% |

_Source: Illustrative data: each group has €3,500.00 of ad revenue, so the combined rate is the average of the two_

Each group has its own break-even. With the same 40% margin and costs, shoes break even at 2.90 ÷ 0.75 = 3.86 and accessories at 2.90 ÷ 0.95 = 3.05. Both run at a ROAS of 3.5 on €1,000.00 of ad spend. Accessories clear it and earn €147.13 after ads. Shoes miss it and lose €94.38.

**Figure: Every group gets a ROAS of 3.5: how much does each one need?.** For each group, one bar is the ROAS it needs to break even and the other is the 3.5 it gets. Accessories need 3.05, so they make a profit; shoes need 3.86, so they lose money. Together the shop needs 3.41 and clears it, so the account hides the shoes.

| Product group | ROAS needed to break even | ROAS each group gets |
| --- | --- | --- |
| Shoes (25% returns) | 3.86 | 3.5 |
| Accessories (5% returns) | 3.05 | 3.5 |
| Both together (15%) | 3.41 | 3.5 |

_Source: Illustrative data, calculated with the model of the break-even ROAS calculator: €1,000.00 of ad spend and €3,500.00 of revenue per group, margin 40%, payment fee 2.5%, reverse logistics 3%_

Together they earn €52.75 after ads, and the shop's 3.5 sits above its blended break-even of 3.41. The account report looks fine, and the shoes lose money inside it. [Break-even ROAS per product](https://www.productmetrics.io/blog/break-even-roas-per-product) shows the same mechanism for margin differences.

Use product groups, or products with enough orders. One product with ten orders and three returns reads 30% largely by chance, while a group of a few hundred orders is steadier.

The same 20% returns multiply break-even ROAS by 1.25 at any margin, so in absolute terms the thinnest margin moves furthest: +1.02 at 30% against +0.56 at 50%.

**Figure: ROAS needed to break even at three margins, with and without 20% returns.** Each pair of bars is the ROAS a product needs to break even at that margin, without and with 20% returns. Returns multiply it by 1.25 at every margin, so the thinnest margin suffers most: +1.02 at 30% (4.08 to 5.10) against +0.56 at 50% (2.25 to 2.81).

| Product margin | ROAS needed, no returns | ROAS needed, 20% returns |
| --- | --- | --- |
| 30% margin | 4.08 | 5.1 |
| 40% margin | 2.9 | 3.62 |
| 50% margin | 2.25 | 2.81 |

_Source: Illustrative data, calculated with the model of the break-even ROAS calculator: payment fee 2.5%, reverse logistics 3%, returned products resold_

For each product group, the test is **break-even ROAS = 1 ÷ (margin on kept sales × (1 − return rate)).** Above it, increase the group's priority. Below it, lower the priority, or fix the cause first, such as size information, photos or packaging.

In Product Metrics, Product Score (0 to 100) can weight Return Rate as one of six metrics, so a 25% return rate like the shoes' can count in a product's score if you give it weight. [Product Segmentation](https://www.productmetrics.io/product-segmentation) works from ad clicks and ROAS or POAS, not from this score.

## 7. How do you reduce your return rate?

Most returns trace back to the product, its size or its description. In an EHI survey of 500 German consumers for KPMG (press release of 11 December 2025), quality defects, wrong size and damage each came up for about 70% of respondents, and answers could overlap.

**Figure: Reasons German consumers give for returns (% naming each).** Each bar is the share of 500 German consumers who named that reason. Quality defects (73.8%), wrong size (72.6%) and damage (71.2%) lead, and 46.2% admit to deliberate double orders. People could name several reasons, so the shares add up to more than 100%.

| Return reason | Share of consumers naming it |
| --- | --- |
| Quality defects | 73.8% |
| Wrong size | 72.6% |
| Damaged product | 71.2% |
| Description or image mismatch | 60.4% |
| Deliberate double orders | 46.2% |

_Source: EHI Retail Institute survey of 500 German consumers for KPMG, press release of 11 December 2025 (checked 7 October 2026). Multiple answers possible_

Retailers read it the same way. In EHI's retailer study, 86% called detailed shop information with precise descriptions and images the most important way to lower the rate, 74% tried to avoid returns on purpose and 70% recorded why items came back. Both surveys are German, so check the reasons in your own market.

So record the reason per product before you pick a fix: size information, photos and copy, packaging, or supplier quality. A bad return also costs repeat sales. In the KPMG survey, 36.6% of customers had not shopped with the same retailer again after a negative return experience.

## 8. Work out your own numbers

**Work out your return rate and its ROAS effect**

1. **Group orders by order date.** Take one month of orders per product group, and let the return window close before you count. A return arrives days or weeks after the sale.
2. **Divide refunded value by sold value.** Use revenue excl. VAT on both sides. Keep the order and unit rates for stock and warehouse planning.
3. **Divide your break-even by (1 − return rate).** Start from 1 ÷ contribution margin on the revenue you keep. At 20% returns, divide by 0.8.
4. **Check which ROAS you compare it with.** If returns are uploaded to Google Ads as conversion adjustments, your ROAS is after returns: compare it with the break-even on kept sales. If not, compare it with the raised one.

Order and refund reports in your shop platform hold the inputs. Begin where the ad budget is heaviest: a wrong break-even there costs the most.

## Try it on one product group

Enter your return rate, margin and fees in the [break-even ROAS calculator](https://www.productmetrics.io/break-even-roas-calculator) to see the break-even for one product group. [How to calculate ROAS](https://www.productmetrics.io/blog/how-to-calculate-roas) explains which revenue to count, and [target ROAS from margin](https://www.productmetrics.io/blog/target-roas-from-margin) turns a break-even into a target that leaves profit after ads.

## Frequently asked questions

### What is a good return rate for ecommerce?

No rate is good for every shop. NRF and Happy Returns report that US retailers expected 19.3% of online sales to be returned in 2025. EHI reports an article-based average of 6% to 10% across all product groups for online retailers in Germany, Austria and Switzerland, and 26% to 50% in fashion. A good rate for you is one your margin absorbs: calculate break-even ROAS at your own rate, per product group, and compare each group with its own history.

### How do you calculate return rate?

Divide what came back by what you sold, for the same orders. By orders: returned orders ÷ orders placed. By units: returned units ÷ units sold. By value: refunded revenue ÷ revenue sold. For example, €9,200.00 refunded on €100,000.00 sold is 9.2%. Use value when you want to know what returns do to ROAS and profit.

### Do returns affect ROAS?

Yes. Google Ads reports the conversion value your tag recorded at purchase, and a return only changes it if you upload a conversion adjustment against the order ID. Without adjustments your ROAS is before returns, so break-even ROAS has to be raised by dividing it by (1 − return rate).

### Should returns be counted in POAS?

Yes. POAS is contribution margin ÷ ad spend, and contribution margin should be measured on the sales you keep, after returns and the cost of handling them. A returned order brings in no margin but its ad click was already paid for.

### How much does an ecommerce return cost?

EHI found that online retailers in Germany, Austria and Switzerland pay €5 to €10 per returned article on average, and €10 to €20 for home and furnishing, with inspecting and quality-checking returned goods the biggest cost. If the product cannot be resold, add its product cost. Read your own figure from carrier and warehouse invoices.

### How do I reduce my ecommerce return rate?

Find the reason per product first, because each reason has a different fix. In EHI's study, 86% of retailers called detailed shop information with precise descriptions and images the most important measure, and 70% record why items come back. Typical reasons are quality defects, wrong size, damage and descriptions that do not match the product. Start with the product groups that have the highest return rate, since they cost the most.

---

Written by Berend Vrakking, founder of Product Metrics. Last updated 2026-10-08.

HTML version: https://www.productmetrics.io/blog/ecommerce-return-rate