Dropshipping returns and refunds: the EU 14-day rule in practice

The short answer

If you sell to a consumer in the EU, they can send the goods back within 14 days of delivery without giving you a reason, and you have 14 days from being told to refund them. That is not your policy, and you cannot write it away in your terms. It applies to you even if your company is outside the EU, because the rule follows the customer.

The part that actually decides whether your store survives is not the rule. It is that your goods shipped from a supplier warehouse you do not control, and the return comes back to you. The law gives your customer a right that is cheap for them to use and expensive for you to honour, and the gap between those two things is the whole problem. This guide covers the rule exactly, the change that landed on 19 June 2026, what one refund really costs, and the arithmetic that tells you which products can carry it.

This is general information rather than legal advice, and national implementations differ. Where a rule comes from a specific article or court judgment, it is named in the sentence so you can check it.

The rule, article by article

The EU Consumer Rights Directive (2011/83/EU) sets out a small number of deadlines that all run in the same direction. Most stores get one or two of them wrong, usually the ones that cost money.

Read those together and a pattern appears: three separate provisions punish the same failure, which is not telling the customer their rights before they buy. Getting the returns page right is the cheapest compliance work in ecommerce, and skipping it is the most expensive.

What changed on 19 June 2026: the withdrawal button

This is new enough that most guides on this query still do not mention it. Directive (EU) 2023/2673 inserted Article 11a into the Consumer Rights Directive, covering the exercise of the right of withdrawal from distance contracts concluded through an online interface, and the obligation applies from 19 June 2026. If a customer can buy from you in a browser or an app, they must be able to cancel the same way.

On the reported reading of it, a compliant setup needs a clearly labelled withdrawal function that is prominently displayed and easily accessible throughout the whole withdrawal period, a separate confirmation step so nobody cancels by accident, and an automatic acknowledgement of the withdrawal sent to the customer on a durable medium such as email, recording what they submitted and when. Article 11a does not create any new withdrawal rights and does not touch the exceptions list, so nothing below changes because of it. It only changes how the customer exercises a right they already had.

Two practical notes. It reaches traders outside the EU who direct their activities at EU consumers, so a US or UK registered store selling into Germany is in scope. And the reported consequence of getting it wrong is the familiar one: the withdrawal period does not start running properly, which puts you back in Article 10 territory with a year of open cancellations. Sourcing honesty here, because it matters: we could not load the EUR-Lex text of Article 11a in this session, so the detail above comes from law firm and industry commentary published on the change rather than from the directive itself. The date and the existence of the obligation are consistent across every source we checked. Confirm the implementation detail with your own legal advice or your platform’s compliance documentation before you build it.

The exceptions, and the two that dropshippers over-claim

Article 16 lists the cases where there is no withdrawal right. The ones that come up in dropshipping are goods made to the customer’s specifications or clearly personalised, goods liable to deteriorate or expire rapidly, sealed goods which are not suitable for return for health protection or hygiene reasons and were unsealed after delivery, goods inseparably mixed with other items after delivery, and sealed audio or video recordings or computer software unsealed after delivery.

The hygiene exception is the one stores reach for, and it is much narrower than the average returns page pretends. In Case C-681/17, slewo v Ledowski, decided on 27 March 2019, the Court of Justice held that a mattress whose protective film the customer had removed after delivery was still returnable. The exception is read strictly because it restricts a consumer right, and it covers goods that genuinely cannot be put back on sale for health or hygiene reasons, not goods that merely need cleaning. If a mattress does not qualify, your product probably does not either.

Personalisation is the other one. A name printed on the item is a specification. A choice between the black one and the white one is not, because picking from options you already stock is ordinary shopping. Claiming an exception you do not have is worse than having no exception at all, because it is exactly the kind of clause that counts as failing to inform the customer correctly.

The part the law does not solve: the return address

Nothing in the directive requires you to have a warehouse. It requires you to take the goods back and refund them. That is where dropshipping and consumer law pull hardest against each other, because the goods came from a supplier who has no interest in receiving them.

Sending a parcel from Europe back to China typically costs somewhere in the range of $15 to $40, which is often more than the item is worth, and that figure comes from fulfilment industry write-ups rather than a carrier rate card, so treat it as an order of magnitude. Suppliers know this. CJ Dropshipping states plainly on its own policy pages that it does not recommend returning products to its warehouse, because international shipping is expensive and the parcel takes months to arrive, by which point the goods are often damaged anyway. Its process is a refund or a resend on the supplier side, opened as a dispute, rather than a physical return.

So the honest picture is three separate transactions that people confuse for one. Your customer withdraws under EU law and gets their money back from you. You separately claim against your supplier under the supplier’s own dispute policy, which has different deadlines, different evidence rules and no obligation to match what you just paid out. The physical goods usually stay wherever they are. Whether the supplier pays is a commercial question about your relationship with them, not a legal consequence of your customer’s right. Budget for the first and hope for the second.

What one return actually costs

Here is a €59.90 product with free shipping, sourced from outside the EU at €18.00 landed including the supplier’s shipping, carrying the €3 flat import duty that has applied to low value parcels since 1 July 2026. Card processing is at 2.1% plus €0.30, the figure most 2026 breakdowns quote for Shopify Payments on the Basic plan in the EU, though rates differ by country and France is reported lower at 1.5% plus €0.25.

A kept order contributes €59.90 minus €1.56 in fees, minus €18.00 of goods, minus €3.00 of duty, which is €37.34 before you have paid for a single ad. Now refund it. The customer gets €59.90 back. The processing fee is not returned to you: Shopify, Stripe and PayPal all keep the fee from the original transaction when you issue a refund, and PayPal says so explicitly. The goods are with the customer or in a parcel nobody wants. The €3 duty was paid at import and is gone in practice, since EU customs law does have returned goods relief but it needs re-export and documentation that is not realistic per parcel at this value. Cash in €59.90, cash out €82.46, so the order that looked like €37.34 of contribution is €22.56 of loss.

The useful part is what happens when you spread that across a return rate, because the arithmetic collapses into one line. Every order either keeps €37.34 or loses €22.56, and the difference between those two outcomes is exactly the €59.90 the customer paid. So your average contribution per order is €37.34 minus your return rate times the sale price. One point of return rate costs you one percent of your sale price, every time, whatever your cost structure looks like.

That table assumes the returned goods are worth nothing to you, which is the realistic dropshipping case and not the retail one. If you hold stock locally and can resell a return, the loss per return falls by whatever you recover and the whole picture improves. It also assumes a €30 cost per purchase, which is optimistic against the benchmarks in dropshipping profit margins. Run your own numbers on the profit margin calculator and the breakeven ROAS calculator, and add the duty properly with the EU landed cost calculator.

Return rate is chosen at product selection, not fixed later

The benchmarks are consistent enough to plan with. Roughly one in five online orders comes back, against about a third of that rate in physical stores. By category, Richpanel’s 2026 compilation of returns industry datasets puts apparel at 20% to 40%, footwear 17% to 30%, home and furniture 15% to 23%, accessories and jewellery 12% to 15%, electronics 8% to 15% and beauty and personal care 4% to 12%. Those are ranges compiled from commercial benchmark sets rather than an audited census, so treat the band as real and the exact figure as approximate. The National Retail Federation estimate that around 9% of returns are fraudulent sits on top of all of it.

Put that next to the table above and the conclusion is uncomfortable but clear. Sizing categories do not have a returns problem you can fix with better photography. They have a return rate that is two to five times another category’s, applied to the same sale price, and at that level it decides profitability before your ads do.

Shapewear is the clearest example on our own radar, because it combines the worst of both: it is a sized garment, so it sits in the apparel return band, and it is intimate apparel, so the hygiene exception is the first thing every store in the niche tries to claim. After slewo, a returns page that refuses all shapewear returns outright is on thin ice. The workable version is sealed packaging plus an honest, prominent statement of what unsealing forfeits, and a sizing guide good enough that fewer people unseal the wrong size in the first place.

Earbuds sit at the other end and are worth studying because the category is genuinely favourable on the return rate and genuinely awful on the cost of the ones you do get. Electronics run 8% to 15%, but reverse logistics on an electronics return is quoted at $30 to $65 per item in the same 2026 compilation, which is more than the margin on most dropshipped earbuds. This is the category where refunding without asking for the goods back is usually the cheaper policy, not the generous one.

Skincare tools land in the 4% to 12% beauty band, which is the friendliest set of numbers in the table, and they are the clearest legitimate case for sealed hygiene packaging. Seal it, say what unsealing means, and you have an exception you can actually stand behind. For how to weigh this against demand when you are choosing what to sell, see best dropshipping niches and how to validate a product.

There is a quiet signal in what advertisers actually pick. As of August 2026, exactly one of the twenty largest niches on our radar is a sized garment. Everything else is gadgets, home goods, beauty, accessories and pet products. Nobody wrote that rule down; it is what survives when people run the numbers with real money.

Refunding without asking for the goods back

Once the arithmetic above is in front of you, the standard advice to always demand the item back stops making sense. You are allowed to waive the return. The directive gives you a right to withhold the refund until the goods move, not an obligation to collect them.

Compare the two paths on a €59.90 order. Demand the return and the customer pays €15 to €40 to ship it to an address in another hemisphere, waits weeks, and you receive an item you cannot resell, having lost the same €22.56 plus the goodwill. Refund without the return and you lose €22.56, immediately, with a customer who is not angry and not calling their bank. The second path is cheaper on any item where reverse logistics costs more than the recoverable value, which for dropshipping is most items under about €40.

The obvious objection is that people will exploit it, and some will. Set a threshold rather than a blanket policy: waive the return under a value where the shipping exceeds the recovery, require it above that, and watch for repeat claimants on the same address or card. That threshold is a business decision about your own numbers, and it is much easier to make once you have priced returns into the product properly using the pricing calculator.

Chargebacks: the price of making returns difficult

A hard returns process does not stop the refund. It changes who processes it and adds a fee. A customer who cannot get their money back from you calls their bank, and a chargeback is strictly worse than the refund you were avoiding: you lose the money anyway, you pay a dispute fee, and the dispute counts against you even if you win.

On Shopify Payments the fee is reported at $15 for US merchants, in a $15 to $30 band depending on region, with the disputed amount and the fee taken from your balance straight away and returned if you win, though Shopify itself says the fee refund depends on your country. The number to watch is your dispute rate. Industry writeups describe merchants being flagged for elevated dispute rates around 1.1%, and card network monitoring programmes sit in that neighbourhood, so roughly one dispute in a hundred orders is the zone where your payment processing becomes the problem rather than the returns.

In the EU there is an extra wrinkle worth knowing. If you never told the customer about the withdrawal right, you are not defending a chargeback on a policy dispute. You are defending it against a legal right the customer unambiguously has for the next twelve months, and you will lose.

What your returns page needs to say

1. That the 14 day right exists, before they pay. Not buried in terms. On the product page or in checkout, because three separate penalties in the directive hang on having informed them.
2. Who pays return postage, explicitly. One sentence stating that the customer bears the direct cost of return. Leave it out and the bill is yours by default.
3. Where returns actually go. An address the customer can use, and honest costs. If you handle small value returns by refunding without collection, say so, because it is a selling point.
4. A model withdrawal form and a working withdrawal function. From 19 June 2026 an online store needs the clearly labelled in-interface route described above, with a confirmation step and an emailed acknowledgement.
5. Any exception you rely on, product by product. Sealed hygiene goods, personalised items, perishables. Name the product category and what forfeits the right, and do not claim an exception you cannot defend.
6. Delivery estimates that are true. The withdrawal window starts on delivery, so slow shipping extends your exposure, and an unrealistic estimate is what turns a patient customer into a chargeback. That is the strongest argument for EU stock in how to find EU suppliers.

For the rest of the store setup around this, see how to build a store that converts, and for the broader EU picture including VAT and product compliance, dropshipping in Europe and the €3 duty explained.

Good returns are a growth lever, not just a cost

The cost framing is only half of it. Sendcloud’s 2026 delivery report, which surveyed 8,000 shoppers across eight European markets and was published on 20 May 2026, found that 70.73% say they are more likely to buy from a store that offers easy online returns, rising to 81.60% in France, and that 41% of European shoppers made at least one return in the past three months. Another 13.48% named complicated returns as a barrier to buying from an international store at all, and 18.24% said they would order more internationally if returns were easy with a local drop-off point. It is a vendor survey by a company selling shipping software, so read the direction rather than the decimal places, but the direction is consistent: returns policy is read before checkout, and it moves conversion.

That is the resolution of the apparent contradiction in this guide. Easy returns raise conversion and raise return rate at the same time. Whether that trade is good depends entirely on the table above, which is why the decision belongs at product selection rather than in a policy meeting. A category at 8% can afford to be generous and will win customers by being so. A category at 35% cannot buy its way out with a better policy.

Where SpotPeaks fits, honestly

We do not handle your returns, hold stock or give legal advice. What we can tell you is which products other people are still paying to advertise after a long time, and that is a better returns signal than it sounds, because an advertiser who has run the same ad for a year is someone whose refund rate did not eat the campaign. As of August 2026 the longest continuously running ad in our data has been live past 1,000 days. Whatever that product’s return rate is, it is survivable at their price.

The SpotPeaks radar currently tracks 640 products with live ads across 80 niches, and that coverage is Facebook-weighted at the moment, so read it as a Facebook-first view rather than a complete picture of every platform.

Browse the niches free at winning products and Facebook products, and every calculator linked on this page works without an account. The full radar with ad longevity history is part of the paid product at $39 per month after a 14-day free trial.

The verdict

The EU 14-day rule is not the hard part. It is a short list of deadlines, and the expensive failures all trace back to one omission: not telling the customer their rights before they pay, which triggers a twelve month window, hands you the return postage and cancels their liability for damage. Fix the returns page, add the withdrawal function that became mandatory on 19 June 2026, and that entire class of problem goes away for roughly an afternoon of work.

The hard part is that a return costs you the full sale price in contribution, so your product choice sets your return rate and your return rate quietly sets your margin. We cannot guarantee profit, and no returns policy can: a category at 30% returns is a category where a good policy makes you lose money slightly more politely. Next step: take the product you are about to test, find its category band in the table above, subtract that percentage of your sale price from your contribution, and check on the breakeven ROAS calculator whether what is left still buys a customer. If it does not, you learned that for free instead of over three months. Then read why a store gets traffic but no sales, which is the same arithmetic applied one step earlier in the funnel.

FAQ