How to reduce your ecommerce return rate — without punishing customers

Returns are the post-purchase cost lever most brands attack last — usually after a scary quarter — and then attack wrong, by making returning harder. That backfires: in the EU the 14-day withdrawal right is mandatory anyway, and friction at the return stage shows up as lost conversion and angry reviews, not fewer returns.

Here’s the honest version: what’s normal, why returns really happen, and the levers that work — in order of impact.

What a normal return rate looks like

Return rates are category physics before they’re anything else. Rough benchmarks:

• Fashion & shoes: 20–50%. Sizing is the structural cause; bracketing (ordering two sizes to keep one) is rational customer behavior.
• Consumer electronics & hardware: 5–15%. Driven by “doesn’t do what I expected” and genuine defects, not fit.
• Home & living: 8–20%. Color/material expectations and transport damage.
• Beauty & consumables: under 10%. Hygiene exceptions limit returns legally.

Two consequences. First: compare yourself against your category, not against “ecommerce average”. Second: if your rate is normal for your category, the bigger win is usually cutting the cost per return, not the rate itself.

Why returns actually happen

Every returns-reason dataset we’ve seen — including our own two brands’ — boils down to a short list:

1. Fit / size (fashion’s #1, irrelevant for most hardware)
2. Expectation mismatch — the photos oversold, the dimensions weren’t listed, the color reads different in daylight
3. Arrived too late — customer bought elsewhere in the meantime; the parcel comes back unopened
4. Damaged in transit or genuinely defective — which is not a “return” at all but a warranty case, and mixing the two up costs you money (more on that in legal guarantee vs. warranty)
5. Bracketing and serial returning — a small customer segment with an outsized share of volume

Note what’s missing: “customers are lazy/dishonest” is not on the list. Treating returns as customer misbehavior leads to policies that punish the 95% for the 5%.

The levers, ranked by impact

1. Fix the product page before anything else

The cheapest return is the one that was never ordered wrong. Exact measurements (garment measurements, not just S/M/L; product dimensions for hardware), photos in honest light, video where texture matters, size guidance built from your own returns data (“runs small — 68% of buyers size up”), and reviews that surface fit feedback. Brands that do this rigorously routinely shave several points off their rate.

2. Deliver when you promised

“Arrived too late” returns are pure logistics. Show realistic delivery windows in checkout, and when a parcel gets stuck, tell the customer before they notice — proactive delay handling is the same mechanism that kills WISMO tickets (the WISMO playbook covers it), and it also prevents the buy-elsewhere-return-yours pattern.

3. Capture the return reason — structurally

A returns flow that asks “why?” with fixed options plus a free-text field turns every return into product data. Route it weekly to whoever owns the product page and purchasing. In practice a handful of SKUs drive most of the excess rate — measure per SKU, not per store — and a recurring reason on one SKU (“strap breaks”, “smaller than pictured”) is an instruction, not a statistic.

4. Make the exchange easier than the refund

A customer who wants a different size or a replacement for a damaged unit is trying to stay your customer. If your process funnels everyone toward “refund” because it’s the only smooth path, you’re converting savable revenue into refunds. Exchange-first flows, instant store credit as an option (never forced), and for low-value items the occasional “keep it, we’ll refund you” — which is often cheaper than round-trip logistics on a €12 product.

5. Handle the outlier segment individually

Serial returners are real but small. The answer is individual — a friendly conversation, in extreme cases refusing future orders — not sitewide friction. Blanket measures (return fees, shorter windows, interrogation-style forms) tax every honest customer for the behavior of a few.

What not to do

• Don’t make returning hard. In the EU, consumers have 14 days regardless — the legal baseline is covered in our right-of-withdrawal guide. Friction doesn’t reduce the legal obligation; it reduces repeat purchase.
• Don’t hide the policy. An unclear returns policy costs conversion at the exact moment of highest purchase intent.
• Don’t treat defect claims as returns. A defective product is a warranty case with different rules and different costs — misrouting these means you pay return shipping you didn’t owe, or deny rights the customer does have.

The cost side: process beats rate

Say you ship 2,000 orders a month at a 12% return rate — 240 returns. Getting the rate to 10% (a solid outcome for a year of PDP work) saves 40 returns. Cutting the handling cost of all 240 — label generated automatically, refund or exchange executed in the shop without a human copying data between portal, helpdesk and backend — usually saves more, faster. We measured this on our own brands: a returns case that took eleven, twelve minutes across three systems takes about two when one system reads the case, checks the window, creates the label and writes the result back to the shop (the full story).

If you want to see that process running on your own store’s cases: the live demo resolves one in front of you — no signup required.