Ecommerce returns: the half of your return rate you caused

What an ecommerce return is

An ecommerce return is an item a customer sends back after buying it online, whether they ship it, hand it to a return point or carry it into a store, plus the refund, exchange or credit that settles it. Almost every guide on returns is about a customer changing their mind, and that's only part of the rate. The rest is orders you broke: the wrong SKU picked, the parcel damaged, the delivery that missed its promised date. No product photograph will stop one of those. I pulled the numbers on one brand's bundle: 157 units sold, 156 sent back, because a bundling app was splitting it into separate line items for the warehouse. Those returns were filed next to size exchanges.

Whatever starts it, a return runs four steps.

1. The request: where the customer asks and the brand issues a return authorization.
2. The journey back: by label and carrier, drop-off point or store counter.
3. The inspection: where the item is received, graded and routed.
4. The settlement: a refund, an exchange or credit.

How big the return rate really is

The National Retail Federation puts total US retail returns at $849.9 billion in 2025, 15.8% of annual sales, and it puts 19.3% of online sales in the returned column. Set that against the channel it lands on. The Census Bureau put first-quarter 2026 e-commerce at $326.7 billion, 16.9% of total retail, up 9.8% year on year. The channel with the higher return rate is the one still taking share, so returns eat more of your revenue every year even when your own rate holds flat.

This is also where most writing on returns stops being reliable. Quoted rates run from 15.8% to 30% depending on which page you open, and the higher ones have no study behind them, only another blog that has none either. Benchmarking against a number with no origin is benchmarking against nothing: take 19.3%, and make anything above it show its method.

Why online returns run higher

The comparison worth quoting comes from a government auditor. The USPS Office of Inspector General puts e-commerce return rates at about three times those of physical retail, and its later review of reverse logistics carries the NRF estimate that 17.6% of US online purchases by revenue were returned by 2023. The gap splits three ways.

• You can't touch it: fit, color and scale are guessed at checkout and confirmed on arrival.
• Bracketing is rational: ordering two sizes and sending one back is what a free returns policy asks the customer to do.
• The order didn't arrive as promised: wrong item, damaged parcel, late delivery, half the order.

Two of those are about the customer. One is about you, and it's the one your merchandising team can't fix.

The returns you caused

A return that starts with a broken order isn't a preference and can't be treated as one. It's an operational failure that took two weeks to surface in the returns report, where it looks identical to a return from somebody who ordered the wrong size.

Every returns dashboard reports one rate, so almost no brand knows which part of its number it built. Fraud takes a slice too, and that's its own subject.

One piece of it is measured, and again by an auditor. The Postal Service handled more than 11.6 million undeliverable Parcel Select packages scanned as return to sender in FY2024, carrying about $138 million in return postage, and only 39% reached the sender. Not one of those is a customer who changed their mind, and every one began as a delivery exception somebody could have acted on. Read it as a floor, not an estimate: it's one carrier's undeliverable traffic.

• The return the customer chose: the item arrived as promised and didn't suit. You reduce it before the order, with better product information, better sizing and a policy that doesn't invite bracketing.
• The return you caused: the item arrived wrong, damaged, late or not at all. You reduce it only before the return starts, by catching the break while the order is still fixable.

At almost every brand the default owner of a broken order is a helpdesk agent reading the customer's email, by which point the parcel is traveling backwards. A helpdesk is a system for replying about problems, not resolving them. Answering the customer and fixing the order are different jobs, and only one stops the return. The second is proactive e-commerce operations: a different function, not a better inbox.

What a return costs once it is moving

The same USPS auditor put the cost best: reverse logistics isn't forward logistics run backwards. It's more complex and more expensive. Volumes are unpredictable, every item is its own decision, and the work lands on the fulfillment operation already shipping today's orders. The bill arrives in four parts.

• The inbound leg: the return label and the carrier movement, usually subsidized or free to the customer.
• Handling and inspection: receiving, grading and testing the item before anything can be decided about it.
• Lost value: the gap between what the item sold for and what it can sell for now, which widens every day it sits in intake.
• Disposal: what it costs when the item can't be sold at all.

That last line has a landfill attached to it. The EPA put US textile generation at 17 million tons in 2018, of which 14.7% was recycled and 11.3 million tons was landfilled.

What the law actually requires

There's no general federal right in the US to return merchandise you simply don't want, which surprises people. The return window is the retailer's own promise. What federal law regulates is the shipping clock and the refund clock. Under the FTC's Mail, Internet, or Telephone Order Merchandise Rule a seller must ship within the window it advertised, or within 30 days if it advertised none, and must otherwise offer the buyer a delay or a cancellation with a prompt refund. The Rule defines a prompt refund: seven working days for cash or check, one billing cycle where the buyer paid on credit. That clock attaches to your failure to ship, not to a customer sending an item back.

• Federal: no general right to return unwanted goods. The FTC Rule governs shipping and refund timing, not whether a return is allowed.
• California: Civil Code 1723 requires the refund policy to be conspicuously posted. A retailer that doesn't post it's liable for the purchase price if the buyer returns the goods within 30 days.
• New York: General Business Law 218-a does the same and names online retailers explicitly, requiring the policy to be shown before billing information is requested.

In both states, the penalty for not posting a policy is being handed the most generous one on the market.

Where returns are actually reduced

The share you caused isn't reduced by a better policy page or a faster portal: by then the return already exists. It's reduced upstream, in the order. Detect the break from signals your systems already emit: the order that hasn't moved, the tracking number with no scan, the stock that's gone. Decide against the promise the order made at checkout, then act: reship, correct the address, re-route, or tell the customer before they ask.

Keeyu isn't a returns portal, not a returns management platform, not a carrier, not a 3PL and not a helpdesk. We don't authorize a return, print a label, grade an item or restock it. We work on the order, before the return exists.

If your return rate is one number, you're pulling one lever at two problems, and the share you built yourself is the share nobody is separating. Every order is a promise. Keeyu keeps the promise: we watch each order against what it was promised, detect the break, decide what should happen and act on it, usually before the customer knows anything went wrong. See what the Keeyu platform does after checkout.

• National Retail Federation and Happy Returns. 2025 Retail Returns Landscape. An estimated 19.3% of online sales returned in 2025, and returns at 15.8% of annual retail sales.
• National Retail Federation. Returns projected at $849.9 billion in 2025. The $849.9 billion headline figure behind the returns rate.
• US Census Bureau. Quarterly Retail E-Commerce Sales. First quarter 2026 e-commerce at $326.7 billion seasonally adjusted, 16.9% of total retail sales, up 9.8% year over year.
• USPS Office of Inspector General. Riding the Returns Wave (RARC-WP-18-008). E-commerce return rates at roughly three times those of physical retail, and reverse logistics as more complex and more expensive than forward logistics.
• USPS Office of Inspector General. Sending It Back (RISC-RI-24-005). Carries the NRF estimate that 17.6% of US online purchases by revenue were returned by 2023.
• USPS Office of Inspector General. Protecting Revenue for Returned Parcel Select Packages (25-053-R25). More than 11.6 million undeliverable Parcel Select packages scanned as return to sender in FY2024, about $138 million in return postage, and only 39% reaching the sender.
• US Environmental Protection Agency. Textiles: Material-Specific Data. US textile generation of 17 million tons in 2018, 14.7% recycled and 11.3 million tons landfilled.
• Cornell Legal Information Institute. 16 CFR 435.2, Mail Order Rule. The advertised shipping window, the 30 day default, and the seller's duty to offer a delay option or cancellation.
• Cornell Legal Information Institute. 16 CFR 435.1, definitions. Prompt refund defined as seven working days for cash or check and one billing cycle where the buyer paid on credit.
• California Legislative Information. California Civil Code 1723. Conspicuous posting of a refund policy, and liability for the purchase price where it is not posted and goods are returned within 30 days.
• New York State Senate. General Business Law 218-a. The refund policy posting duty, applied explicitly to online retailers and required before billing information is requested.