Ecommerce Return Rates in 2026: Benchmarks by Category, and the Metric That Actually Matters

The benchmark, current as of 2026

Online returns run about 20%. Apparel runs closer to 40%.

If you searched for "ecommerce return rates," here is the number you came for, with provenance and a category breakdown. Then read past the table for why the headline rate is the least useful number on it.

The most recent industry data comes from the National Retail Federation's 2025 Retail Returns Landscape, its latest annual report and the benchmark heading into 2026. It found U.S. retailers expected about 15.8% of all sales to be returned, totaling $849.9 billion, down slightly from 16.9% and $890 billion the year before.^[1] Online returns run higher than that blended figure, because shoppers cannot touch the product before they buy: the NRF puts the ecommerce return rate at about 19.3% of online sales,^[1] and independent aggregators that weight toward DTC put the practical 2026 number right around 20%. Either way, roughly one in five online orders comes back, two to three times the in-store rate. (The NRF's next annual update is expected in late 2026; until then, these remain the current figures.)

That average hides almost everything that matters, because return rate is mostly a function of what you sell:

Category	Typical online return rate
Apparel	20–40%
Footwear	17–30%
Auto parts	~19%
Home & furniture	15–23%
Accessories & jewelry	12–15%
Electronics	8–15%
Beauty & personal care	4–12%

Ranges compiled from 2025–2026 returns-industry benchmark datasets; figures vary by source and by how each one defines a "return."^[3]

A few drivers explain most of the spread. Fit and sizing is the single largest cause of returns, responsible for up to 70% of apparel returns, which is why apparel and footwear sit at the top.^[3] Bracketing, ordering several sizes or colors intending to keep one and send the rest back, is now a mainstream habit practiced by a majority of online shoppers, up sharply from roughly 40% in 2018.^[3] And categories that cannot be resold once opened, like beauty, stay low because the return is often refused by policy, not declined by the customer.

The honest diagnosis

Return rate is the number you can least control.

Most "reduce your return rate" advice treats the rate as a dial you can turn down with better sizing charts and clearer photos. Those help at the margin. They do not move the rate much, because the two biggest inputs sit outside your reach.

Your category sets the floor. If you sell apparel, you will live in the 20–40% band no matter how good your product pages are. A beauty brand at 6% and an apparel brand at 30% are both healthy. Comparing your rate to a cross-category "average" tells you almost nothing.

Your best customers return the most. This is the counterintuitive part. Zappos has said for years that its highest-returning customers are also its highest-spending and most loyal.^[3] A high return rate often means customers trust you enough to buy freely, knowing they can send back what does not work. Suppress returns aggressively, with restocking fees, short windows, or a hostile process, and you suppress that buying behavior along with it.

The data backs the risk: 71% of consumers say they are less likely to shop with a retailer again after a bad returns experience, up from 67% a year earlier, and 82% now name free, easy returns as a major factor in where they buy, up from 76%.^[1] The backlash against fees is sharper still: 57% will not shop with a retailer again after being charged for a return, up steeply from 40% a year earlier.^[1] (Gen Z returns the most, averaging 7.7 online returns a year.^[1]) The returns experience is a retention lever. The return rate is mostly a category constant.

So stop optimizing the number you can't move, and start optimizing the two you can.

What to measure instead

Track refund rate and cost-per-return instead.

These two numbers are inside your control, they map straight to margin, and most brands track neither one cleanly.

1. Refund rate: how much of every return leaks cash

A return is not automatically a loss. It is a loss only when it ends in a cash refund. If the same return ends in an exchange (a different size or product) or store credit, the revenue stays on your books and often grows. Your refund rate is the share of returns that leave as cash. It is the number that actually hits your bank account, and it is far more movable than the return rate.

The mechanism that moves it is an exchange-first flow: when a customer starts a return, show them the right replacement before you ever show them a refund button. Wrong size becomes the next size up. A color that did not work becomes one that does. Done well, a large share of would-be refunds convert into retained revenue.

2. Cost-per-return: how much support labor each one burns

The cost of a return is not just reverse shipping and restocking. For many brands the bigger, hidden cost is support labor. When every return starts as an email ("how do I return this?") and an agent has to toggle into Shopify, look up the order, generate a label, and reply, you are paying a human for work a portal should do. Reverse logistics alone can run $30–65 per item in electronics, and for furniture it can exceed the product's margin.^[3] Add the agent minutes on top, and a "20% return rate" can quietly mean a meaningful share of your support team's week.

Cost-per-return is what self-service attacks. A return the customer completes alone, with no agent touch, costs a fraction of one that becomes a ticket.

The math that actually moves margin

An exchange keeps the revenue. A refund gives it back.

Two brands can post the identical 25% return rate and completely different P&Ls. The whole difference is what happens after the customer clicks "return."

Take a brand doing $2M a year online at a 25% return rate. That is $500K of returned merchandise in motion. The return rate is fixed by the category. The outcome is not:

Refund-default flow. Most returns end in cash back. Say 80% refund, 20% exchange. You hand back roughly $400K and retain $100K.
Exchange-first flow. The portal steers each return toward a replacement first. Move the mix to 50% exchange or store credit and you retain $250K instead of $100K. Same return rate. $150K more revenue kept.

Those percentages are illustrative, not a promise; your real conversion depends on catalog and policy. But the shape is the point: the lever is the refund-to-exchange ratio, not the return rate. A brand fighting to shave two points off its return rate is chasing maybe $40K of avoided returns while leaving $150K of convertible revenue on the table.

What this looks like in production

A footwear brand we run returns for processes roughly 7,500 returns a month and converts about 40% of them into exchanges rather than refunds.

At that volume, every point of refund-to-exchange conversion is real money kept on the books. The portal does the steering automatically, by return reason, so the gain costs nothing in extra agent time.^[4]

From the brands we run returns for

What the brands winning on returns actually do.

Four moves, drawn from anonymized patterns across high-volume Richpanel returns deployments.^[4] Notice that none of them try to lower the return rate.

Lead with the exchange, not the refund

The portal shows replacement options based on the stated return reason before a refund is ever offered. "Too small" routes straight to the next size up. This is the single biggest lever on refund rate.

Make store credit the better deal

A small bonus tips the choice. One jewelry brand offers a 10% bonus when a customer shops a new item with their return credit, and 5% for plain store credit, nudging cash refunds toward retained revenue.

Sometimes, just let them keep it

For low-value, high-return-shipping items, "keep it and take a partial refund" is cheaper than paying for reverse logistics. Several brands offer keep-it refunds below a price threshold, scoped by return reason.

Resolve the return without an agent

Self-service for the standard cases, AI for the exceptions. The customer completes a size swap themselves; only edge cases (damage, warranty, a bundle) reach a human, and they arrive with full context. Cost-per-return drops.

The common thread: every one of these is a decision about what happens after the return starts, not an attempt to stop returns from happening. That is the whole reframe. We built Richpanel's exchange-first returns portal around it, and it lives inside the same platform your support team already uses, so a return that does turn into a question gets answered by the AI agent in the same inbox, not bounced to a separate tool.

The honest caveat

When the return rate does matter.

Reframing away from the rate does not mean ignoring it. There are real cases where the rate itself is the problem worth attacking:

Fraud and abuse. The NRF estimates about 9% of returns are fraudulent (wardrobing, empty-box scams, counterfeit returns), and 85% of retailers now deploy AI to detect it.^[1] Broader measures that fold in policy abuse run higher: Appriss Retail and Deloitte put return fraud north of $100 billion a year, roughly 15% of all return volume.^[2] Here the rate, specifically the serial-returner tail, is worth attacking directly with policy, tracking, and blocklists.
A rate climbing against your own baseline. The cross-category average is noise. Your own trend is signal. A rising rate on a specific SKU usually means a real product problem: a sizing run that shifted, a description that overpromises, packaging that arrives damaged. That is fixable, and worth fixing.
Margin-thin or oversized goods. When reverse logistics can cost more than the item's margin (furniture, heavy goods), every avoided return is pure savings, and stricter final-sale rules can be rational.

And the honest counter to our own pitch: if your return volume is genuinely low and rarely turns into support tickets, a dedicated standalone returns app does the portal job perfectly well. The case for unifying returns with your helpdesk gets strong specifically when returns keep becoming conversations, which is exactly when cost-per-return is eating your team.

What to do with this

What to do with this, starting this quarter.

Six moves, in rough priority order. The first two are measurement, because you cannot manage what you are not tracking.

1. Instrument refund rate

Split returns into refunds vs. exchanges vs. store credit. If you only track "return rate," you are blind to the number that actually hits your bank.

2. Instrument cost-per-return

Reverse shipping, plus restocking, plus the agent minutes each return consumes. The labor line is the one most brands never count.

3. Turn on an exchange-first flow

Show the right replacement before the refund button, keyed to the return reason. This is the highest-leverage change you can make on refund rate.

4. Add a store-credit incentive

A 5–10% bonus on credit or shop-now reliably shifts the refund-to-exchange ratio without touching the return rate.

5. Move standard returns to self-service

Let customers complete size swaps and standard returns themselves. Reserve agents for the exceptions, handed off with full context.

6. Audit your rate by SKU, not in aggregate

Find the specific products driving an above-baseline rate and fix the real cause (sizing, photos, packaging) instead of chasing a portfolio average.

Frequently asked

Ecommerce returns, answered plainly.

What is a good ecommerce return rate?

There is no single good number, because return rate is mostly set by category. The latest industry data, the NRF's 2025 Retail Returns Landscape (its most recent annual report, the benchmark heading into 2026), puts the overall ecommerce rate at about 19 to 20% of online orders. Apparel runs 20 to 40%, footwear 17 to 30%, electronics 8 to 15%, and beauty 4 to 12%. Compare your rate to your own category and your own history, not to a cross-category average. And treat the rate as a weak target: refund rate and cost-per-return tell you far more about margin.

Is a high return rate always bad?

No. Zappos has long noted that its highest-returning customers are also its highest-spending and most loyal, because a generous return experience is what makes them buy freely in the first place. Aggressively suppressing returns with fees, short windows, or a painful process tends to suppress that buying behavior too. The cost of returns lives in cash refunds and support labor, not in the rate itself.

How do I convert returns into exchanges instead of refunds?

Use an exchange-first returns flow. When a customer starts a return, the portal shows the right replacement (the next size up, a different color) based on the stated return reason, before it ever shows a refund button. A small store-credit or shop-now bonus, commonly 5 to 10%, shifts more customers toward keeping the revenue with you. A footwear brand we run returns for converts about 40% of its returns into exchanges this way.

Does a self-service returns portal just create a support ticket?

A good one does not. The common failure of weak self-service is that it collects a request and then hands it to an agent, which is just a ticket with extra steps. A real exchange-first portal lets the customer complete a standard size swap or exchange themselves, end to end. Only genuine exceptions, such as a damaged item, a warranty claim, or a complex bundle, route to a human, and they arrive with full context.

How much does processing a return actually cost?

More than the reverse shipping. Reverse logistics and restocking can run $30 to $65 per item for electronics, and for furniture or oversized goods it can exceed the product's own margin. The cost most brands never count is support labor: when every return starts as an email an agent handles manually, a 20% return rate can consume a meaningful share of the team's week. That hidden labor cost is what self-service removes.

Should returns live in my helpdesk or a standalone returns tool?

Both models work. A dedicated returns app handles the portal job well, and if your return volume is low and rarely becomes a conversation, it is sufficient. The case for running returns inside your helpdesk gets strong when returns regularly turn into support tickets, because keeping returns and support in one platform, one bill, and one place for your data and agents directly cuts cost-per-return.

Sources & references

Where the numbers come from.

Inline citations [1]–[4] map to the entries below. Public benchmarks link to their primary source; the first-party figures are anonymized aggregates, with the methodology stated.

How the first-party numbers are derived

What [4] refers to: The exchange-conversion rate, store-credit-bonus structures, and keep-it-refund patterns are observed across live Richpanel returns deployments for high-volume ecommerce brands (footwear, jewelry, apparel, supplements). Figures are reported as anonymized aggregates; per-brand data is bound by NDA.
Illustrative vs. measured: The $2M / 25% worked example is illustrative arithmetic to show the shape of the lever, not a customer result. The ~7,500 returns/month and ~40% exchange-conversion figures are real, anonymized, drawn from a single deployment.
Why category rates are ranges: Category-level return rates vary by source and by definition (some count exchanges as returns, some do not). We publish ranges rather than false-precision point estimates, and cite the benchmark class rather than a single study.

National Retail Federation & Happy Returns (a UPS company), “2025 Retail Returns Landscape” (released October 2025), with the prior-year “Consumer Returns in the Retail Industry.” The NRF's most recent annual returns report and the current benchmark heading into 2026 (its next update is expected late 2026). Source for the 15.8% overall rate ($849.9B), the 16.9% / $890B prior-year figures, the ~19.3% ecommerce rate, the 9% fraud share, the 85% AI-adoption figure, and the returns-experience statistics (71% up from 67%, 82% up from 76%, 57% up from 40%, Gen Z averaging 7.7 online returns a year). Based on surveys of 2,006 consumers and 358 ecommerce professionals at merchants over $500M in revenue. nrf.com/research/2025-retail-returns-landscape
Appriss Retail & Deloitte, “Consumer Returns in the Retail Industry.” Source for the broader return-fraud-and-abuse magnitude (north of $100B, roughly 15% of return volume). apprissretail.com
Compiled 2025–2026 ecommerce returns-industry benchmarks. Category-level return-rate ranges, the fit-and-sizing share of apparel returns, bracketing prevalence, and reverse-logistics processing costs are synthesized from multiple 2025–2026 returns-platform datasets and industry analyses; ranges reflect variation across sources. The Zappos “best customers return the most” observation is a long-standing public statement from Zappos leadership.
Richpanel returns deployments (anonymized aggregate, 2025–2026). Exchange-conversion rates, store-credit incentive structures, and keep-it-refund patterns from live Richpanel returns customers, per the methodology above. Aggregate ranges are publishable; underlying tenant data is NDA-bound. Methodology questions: amit@richpanel.com.

Version history, v1.1 (2026-05-21): refreshed framing to 2026 and flagged the data vintage (NRF's 2025 Landscape remains its most recent report); added NRF consumer-behavior stats (82% up from 76%, 57% charged-for-return up from 40%, Gen Z 7.7 returns/yr, 85% AI fraud-detection), widened the electronics and beauty category ranges to current datasets, and tightened the fraud provenance. v1.0 (2026-05-21): initial publication, replacing the legacy /blog/ecommerce-return-rates post.

One in five online orders comes back. The brands that protect their margin don't chase a lower return rate.