Why customers abandon your checkout — ten causes, ranked.

Cart abandonment is the most quoted number in e-commerce and one of the least useful, because the headline figure bundles people who were never going to buy together with people you actively lost. This guide separates the two: what the 70% actually contains, the ten reasons shoppers give for leaving a checkout they had already started, and — for anyone selling into the EU — which of those reasons describes a practice the Consumer Rights Directive does not permit in the first place.

Why shoppers abandon a checkout they had started

Baymard Institute’s most recent quantitative study. Respondents could give more than one reason, so the column does not total 100%, and shoppers who left because they were just browsing are excluded from it entirely.

What the 70% actually contains

The figure everyone quotes is 70.22%, calculated by Baymard Institute across fifty separate studies. It is real, it is stable, and on its own it tells you almost nothing about your shop — because most of what it measures is not a failure of your checkout.

Separately, 42% of online shoppers report having abandoned a cart simply because they were browsing and not ready to buy. Adding items to a cart is how people compare prices, check delivery costs and save things for later; treating that as lost revenue leads to spending money chasing customers who were never in the market. The reasons below deliberately exclude those people. They are the answers given by shoppers who wanted to buy and then stopped.

So the number to move is your own, over time, not the gap between you and 70%. The same applies to conversion generally — the seven causes in why your website isn’t converting are the equivalent list for a site that takes enquiries rather than orders.

Cause 1: the price changes at the last step

Extra costs are the largest single reason at 40%, and "could not see the total cost up front" adds another 12%. Together that cluster is bigger than the next two causes combined, and it is one problem wearing two hats: the shopper reached a number they had not been shown, at the point where they had already decided to buy.

Note what it is not. It is not that shipping is expensive — shoppers pay for delivery constantly. It is that the price appeared late, which reads as a trick even when it is only an oversight in how the cart was built.

In the EU, the biggest cause of abandonment describes something the law already requires you not to do. The Consumer Rights Directive obliges a trader to give the consumer, before they are bound by the contract, "the total price of the goods or services inclusive of taxes" together with "all additional freight, delivery or postal charges". The same directive requires the order button itself to read "order with obligation to pay", or an unambiguous equivalent — and states that where a trader fails to comply, the consumer is not bound by the contract at all. Pre-ticked boxes that add a paid extra are prohibited outright; consent has to come from a positive action.

The practical fix is ordinary shop-building work: show delivery cost in the cart rather than after the address step, publish a free-delivery threshold if you have one and show the distance to it, and if your rates genuinely depend on destination, put a postcode estimator on the cart page. None of that is a design decision. It is the difference between a compliant checkout and one that is both illegal and losing you two shoppers in five.

Cause 2: you asked them to register

Eighteen per cent leave because the shop required an account. The account exists for your benefit, not theirs — an order does not need one, and every shop platform in existence can attach an order to an email address.

Offer guest checkout, and offer account creation after the order is placed, on the confirmation page, with the fields already filled in from what they just typed. You get a meaningfully higher opt-in rate than you would have got at the gate, and the people who decline still bought something.

Cause 3: delivery is slow — or you never said when

Twenty per cent abandon over delivery speed, and it is worth being precise about which half of that you can influence. Making delivery genuinely faster is a logistics investment. Telling people when their parcel arrives is a text change.

This is not only a checkout problem. Eurostat’s 2025 survey of EU online shoppers found 35.4% encountered a problem of some kind, and the most common one across the whole union was delivery being slower than expected, at 19.9% — the single largest source of dissatisfaction in European e-commerce, ahead of both site usability and damaged goods.

Show an estimated delivery date on the product page, not a shipping duration at the end of checkout. "Arrives Thursday 17 September" and "3–5 working days" describe the same service, and only one of them lets someone decide whether it is soon enough. Where you cannot promise a date, offer a faster paid option and let the shopper choose — the choice itself removes most of the objection.

Cause 4: they did not trust you with a card

Nineteen per cent said they did not trust the site with their card details. That is not a statement about your encryption, which no shopper has ever inspected. It is a reaction to accumulated small signals, and it peaks at exactly the moment you are asking for the number.

What reads as risk: no company name, registration number or physical address anywhere; a returns policy that cannot be found; a checkout that looks visually different from the shop, as though the flow has been handed to a stranger; broken layout on a phone; and contact details that lead nowhere. What reads as safe is dull specificity — a real address, a phone number that connects to a person, order support named on the page, and a checkout that looks like the rest of the shop.

Trust is not a badge row above the pay button. It is the absence of anything that makes a cautious person hesitate while their card is in their hand.

Cause 5: too long — and mostly, too many fields

Seventeen per cent leave because the checkout was too long or too complicated. The useful part of that finding is the diagnosis underneath it: Baymard puts the average checkout at 11.3 form fields as of 2024, down from 12.7 in 2019, while most shops need only eight.

Field count matters more than step count. The average checkout has been 5.1 steps since tracking began in 2012, and that number barely moves because it barely matters — a shopper does not resent clicking "continue", they resent typing. Combine address lines, drop the company field unless you sell B2B, derive the city from the postcode, and stop asking for a phone number you will not call.

Then make the fields that remain cheap to complete. Correct autocomplete attributes and input types let a browser fill an address in one tap and give a phone keyboard an @ key instead of the full alphabet; the same change is covered in more detail as cause 8 of why your website isn’t converting. It is an afternoon of work and nobody will ever compliment you on it.

Cause 6: it simply broke

Another 17% abandoned because the site errored or crashed. Checkout is usually the least tested flow in a shop — it is tedious to test, it needs real payment credentials, and it is the one page nobody browses casually — so it is also where breakage survives longest.

Eurostat found 11.5% of EU online shoppers in 2025 reported that a website was difficult to use or worked unsatisfactorily, making it the second most common problem after slow delivery. That is the same phenomenon measured from the other side, by a statistical agency rather than a usability lab.

Two things pay for themselves. One is an automated test that completes a real order end to end on every deploy, including a card decline and a validation error. The other is treating checkout as a performance page rather than a functional one: interaction delay on a payment form is abandonment, and the thresholds in Core Web Vitals explained apply here more sharply than anywhere else on the shop.

Cause 7: the returns policy

Thirteen per cent left because the returns policy was unsatisfactory — and a good share of those had simply not been able to find it. In the EU this is largely a communication problem rather than a commercial one, because the entitlement already exists whether you advertise it or not: a consumer buying at a distance has fourteen days from receiving the goods to withdraw, no reason required.

There is a direct penalty for staying quiet about it. If the trader fails to inform the consumer of the right of withdrawal, the period does not stay at fourteen days — it extends to twelve months, and only resets to a fresh fourteen days if the information is eventually supplied. A shop that hides its returns terms to discourage returns has, in law, granted a year of them.

So say it plainly on the product page, not only in a footer policy: fourteen days, who pays return postage, and how long a refund takes. It converts, and it is the cheapest compliance you will ever do.

Causes 8 and 9: payment methods, and declines

Nine per cent abandoned because there were not enough payment methods, and 10% because the card was declined. These look like one problem and are two.

The method mix is a market question rather than a best-practice one. The combination that converts in one European country is not the combination that converts in the next — card dominance, bank transfer, wallets, buy-now-pay-later and cash on delivery all have very different shares depending on where the shopper lives, and a cross-border shop that ships the same three options everywhere is leaving orders in whichever market it guessed wrong about. Look at your own orders by country before adding anything.

Declines are a different discipline. Some are genuine, but a meaningful share are recoverable: an authentication step that fails silently on mobile, no retry offered, an error message that says "payment failed" and nothing else, or no alternative method presented at the point of failure. A declined card should end with the shopper looking at a working next option, not a dead end. If your shop’s payment stack cannot tell you why declines happen, that is the first thing to fix — see top 20 API integrations for an online shop for what that layer should be doing.

The order to fix them in

Total price first, guest checkout second, delivery dates third. Those three are the largest shares in the table, they are the cheapest to change, and none of them requires redesigning anything — they are copy, configuration and where a number appears. Field reduction comes fourth because it takes real work; trust, returns wording and payment mix after that; and the engineering items last, not because they matter least but because they are the ones you cannot do in an afternoon.

What should not be on the list at all is a redesign. Checkout abandonment is a diagnosable, itemised problem with published benchmarks for every line, and changing everything at once is the one approach guaranteed to teach you nothing about which change worked.

If you would rather have someone measure yours, that is what our e-commerce work starts with — the current numbers per step, against the benchmarks above, before anybody proposes building anything. And if the conclusion is that the platform itself is the constraint, a €2,000 shop vs a €50,000 shop sets out honestly what the money buys at each end.

Sources

The abandonment figures, the EU problem rates and the legal text above come from these, checked September 2026. The ranking of what to fix first is ours.

• Baymard Institute — Cart abandonment rate statistics ↗

  Average documented abandonment of 70.22%, calculated across 50 separate studies, plus the quantitative breakdown of reasons given by shoppers who abandoned during checkout — 40% extra costs, 20% slow delivery, 19% card distrust, 18% forced account, 17% complexity, 17% errors, 13% returns policy, 12% no visible total, 10% declines, 9% payment methods. Respondents could select more than one.

• Baymard Institute — Checkout flow average form fields ↗

  The average checkout carried 11.3 form fields in 2024, down from 11.8 in 2021 and 12.7 in 2019, against a practical target of 8 for most shops. Average checkout length has held at 5.1 steps since tracking began in 2012, which is why field count is the more useful lever.

• EUR-Lex — Directive 2011/83/EU on consumer rights ↗

  Article 6(1)(e) requires the total price inclusive of taxes plus all additional freight, delivery or postal charges before the consumer is bound; Article 8(2) requires the order button to read "order with obligation to pay" or an unambiguous equivalent, failing which the consumer is not bound; Article 9 sets the 14-day withdrawal period; Article 10 extends it to 12 months where the trader did not inform the consumer of it; Article 22 prohibits pre-ticked boxes for additional payments.

• Eurostat — More than a third of online shoppers face issues ↗

  Published 11 March 2026 on 2025 reference data: 35.4% of EU online shoppers encountered a problem, led by slower delivery than expected at 19.9%, a website that was difficult to use or worked unsatisfactorily at 11.5%, and incorrect or damaged goods at 10.4%.

• web.dev — Web Vitals ↗

  The thresholds behind cause 6: LCP under 2.5s, INP under 200ms, CLS under 0.1, measured at the 75th percentile of real visits — the interaction metric being the one that matters most on a payment form.