Return Rate: Three Denominators, Category Benchmarks, and the ROAS Killer

Your Meta report says 4.2x ROAS on the apparel campaign. What it does not know is that 28% of those dresses came back, some of them worn once to a party. The revenue the ad "drove" partly reversed itself two weeks later, but the ad platform booked the sale and never saw the refund. Return rate is the metric that closes that loop, and for physical-product brands it is one of the most under-watched numbers on the P&L.

The catch is that "return rate" is not one number. Depending on whether you divide by units, orders, or revenue, the same month can read 12%, 18%, or 22%, and each is correct for a different question.

Below: the three denominators spelled out with a worked example, benchmarks by category, and the mechanism by which returns quietly gut the ROAS your dashboards celebrate.

---

What return rate actually measures

Return rate is the share of what you sold that customers sent back, over a given period. The definition is simple; the denominator is where it gets interesting, because "what you sold" can be counted three ways, and brands routinely quote whichever number looks best without saying which one they used.

Three denominators, three different numbers

This is the part almost every "average return rate" article skips. The same returns, divided three ways:

Basis	Formula	Answers
Unit-based	Units returned / Units shipped	How much product physically comes back
Order-based	Orders with a return / Total orders	How many customers initiate a return
Revenue-based	Refunded revenue / Gross revenue	How much money actually reverses

They diverge because returns are not evenly distributed. A customer who orders three sizes of one dress intending to keep one inflates the unit-based rate (two of three units come back) while the order-based rate counts a single returned order. And if the returned items are the cheap ones while the kept item is the expensive one, the revenue-based rate stays low even as the unit rate spikes.

A worked example

A brand ships a month like this:

• Units shipped: 5,000, of which 900 are returned
• Orders placed: 2,000, of which 300 contain at least one returned item
• Gross revenue: $200,000, of which $32,000 is refunded

The three rates:

CopyUnit-based    = 900 / 5,000   = 18.0%
Order-based   = 300 / 2,000   = 15.0%
Revenue-based = $32,000 / $200,000 = 16.0%

All three are true. If you are planning warehouse and restocking labor, the unit rate matters. If you are measuring customer experience and friction, the order rate matters. If you are correcting profitability and ROAS, the revenue rate is the one that hits the P&L. Quote the wrong one to the wrong question and you will mis-size the problem.

Benchmarks by category

Return rate is one of the most category-dependent metrics in ecommerce. Fit-and-feel products get sent back constantly; consumables almost never. Rough ranges on a broadly units/orders basis:

Category	Typical return rate	Driver
Apparel & footwear	20–30%	Fit, sizing, bracketing; higher for occasion wear
Accessories & jewelry	10–20%	Look-vs-expectation mismatch
Electronics	10–20%	Compatibility, buyer's remorse, defect returns
Home & furniture	5–15%	High shipping cost suppresses casual returns
Beauty & skincare	Under 10%	Hygiene rules limit returnability
Consumables & supplements	Under 5%	Used up, rarely returnable

Online apparel sits far above the all-retail average precisely because customers cannot try before they buy and often order multiple sizes on purpose. If you sell apparel, a 25% return rate is normal, not a crisis; a 25% return rate on supplements would mean something is badly wrong.

Why return rate quietly destroys ROAS

Here is the mechanism that catches operators out. Your ad platform measures ROAS at the moment of purchase. Returns happen days or weeks later, on a different timeline the pixel never sees. So the ROAS you optimize toward is a gross number, and your actual, returns-adjusted ROAS is lower, sometimes dramatically.

Walk it through. A campaign spends $10,000 and drives $42,000 in reported revenue:

CopyReported ROAS = $42,000 / $10,000 = 4.2x

Now apply a 25% revenue-based return rate for that apparel campaign:

CopyNet revenue after returns = $42,000 × (1 − 0.25) = $31,500
Returns-adjusted ROAS = $31,500 / $10,000 = 3.15x

The campaign you thought ran at 4.2x actually ran at 3.15x. If your break-even ROAS is 3.0x, your real margin of safety just went from comfortable to razor-thin, and you never saw it because returns processed after the reporting window.

It compounds through the funnel. Returns inflate your effective customer acquisition cost (you paid to acquire a purchase that partly reversed), drag down net profit margin (you eat return shipping, restocking labor, and often the outbound shipping too), and distort conversion rate economics if a chunk of "converted" revenue is temporary. A high-ROAS, high-return campaign can be less profitable than a lower-ROAS campaign that keeps its sales.

When return rate misleads

• Timing lag distorts the current month. Returns arrive weeks after the sale, so a fast-growing month understates return rate (lots of recent, not-yet-returned orders) and a shrinking month overstates it. Cohort returns by order date, not refund date.
• Blended rate hides SKU-level disasters. A calm brand-wide average can conceal a handful of SKUs returning at catastrophic rates.
• Basis-switching flatters. Reporting the order basis when the unit basis is ugly, or vice versa, misrepresents the real cost.
• Not every return costs the same. A resellable return costs shipping and labor; a used or damaged return is a near-total loss. A flat return rate treats them identically.

Frequently asked questions

What is a good return rate for ecommerce?

It is entirely category-dependent. Apparel and footwear routinely run 20-30% and that is normal; electronics 10-20%; beauty under 10%; consumables under 5%. There is no universal 'good' number, only good-for-your-category. Compare to peers in your vertical, not to a blended all-retail average.

How do I calculate return rate?

Divide what came back by what you sold, but decide the basis first. Unit-based is units returned over units shipped; order-based is orders with a return over total orders; revenue-based is refunded revenue over gross revenue. The same month can read differently on each basis, so state which one you mean.

Which return rate denominator should I use?

It depends on the decision. Use the unit basis for warehouse and restocking planning, the order basis for customer-experience and friction analysis, and the revenue basis for profitability and ROAS correction. The revenue basis is the one that hits the P&L.

Why is my apparel return rate so high?

Because apparel buyers cannot try before they buy and often 'bracket' by ordering multiple sizes intending to keep one. That makes 20-30% return rates normal for online apparel. High rates concentrated in specific SKUs or sizes usually point to a fit or sizing-description problem you can fix.

How do returns affect ROAS?

Ad platforms book revenue at purchase and never see the refund that follows weeks later, so reported ROAS is a gross number. Applying your revenue-based return rate gives the real figure: a 4.2x reported ROAS at a 25% return rate is actually 3.15x. Scaling on gross ROAS alone can fund campaigns that are quietly unprofitable.

How do returns affect profit margin?

Returns hit net margin harder than the refund alone suggests. You lose the sale revenue, often eat both outbound and return shipping, pay restocking labor, and take a partial or total loss on items that come back used or damaged. A high return rate can turn a contribution-positive product into a net loss.

Should I count exchanges as returns?

Track them separately. A pure exchange keeps the revenue in the business and mostly costs you shipping and handling, while a refund reverses the revenue entirely. Lumping them together overstates your revenue-based return rate and hides that exchanges are far less damaging to margin.

Summary

Return rate looks like one number and is really three: units, orders, and revenue, each answering a different operational question. Get the denominator right before you benchmark, because apparel at 25% is healthy and supplements at 25% is a fire. Above all, remember that returns process on a timeline your ad platform never sees, so the ROAS you optimize toward is gross and your real, returns-adjusted ROAS is lower. For fit-sensitive brands, judging campaigns on returns-adjusted numbers is the difference between scaling profitably and scaling into a loss.

Keep exploring

Use Case

Ads ROI analysis

A ready-to-run analysis workflow.

Integration

Shopify integration

Connect this source in minutes.

Guide

Returns-adjusted ROAS calculator

A deeper, practical walkthrough.

Guide

How to run cohort retention analysis on Shopify

A deeper, practical walkthrough.

Guide

D2C returns playbook

A deeper, practical walkthrough.

Guide

D2C ecommerce benchmarks

A deeper, practical walkthrough.