Where checkouts lose people, and how to find it in yours

Most people who start a checkout and don’t finish leave at one particular step, for one of a handful of reasons, and your analytics can show you which. The Baymard Institute’s average across 50 studies puts cart abandonment at 70.22%, but a lot of that is browsing and comparing that no checkout could win. The part worth fixing is the shopper who meant to buy and was put off: by a cost that appeared late, an account they were made to create, a form that fought them, or an error.

This is the method I use to find it. Size the leak, read the funnel step by step, match each drop to its likely cause, fix the usual suspects in order of cost, and then recover what you can within the UK rules on marketing email. If the checkout is failing outright rather than leaking, start with why a Shopify website stops working instead.

How much leaks, and how much of it you can win back

The headline figure gets quoted a lot, and it’s worth knowing what it does and doesn’t mean before you measure yourself against it.

Cart abandonment isn’t checkout abandonment

Baymard’s 70.22% measures baskets that don’t become orders. Many of those people never reached the checkout: they were saving items, checking delivery costs or comparing prices. Checkout abandonment, people who started paying and stopped, is a smaller and more fixable number, and it’s the one this piece is about.

An average isn’t a benchmark for you

A store selling expensive, considered purchases can expect more people to leave and come back later than one selling cheap repeat items, whatever the checkout is like. The comparison that tells you most is your own: this month against last, phone against desktop, one country against another.

Some of it is just browsing

Baymard asks shoppers why they left and excludes “just browsing” from its list of reasons, which is a useful reminder that a share of abandonment is people doing exactly what they meant to do. No checkout change wins those back. The reasons below are the ones that do respond to fixes.

Why people leave

Baymard’s survey of US shoppers who abandoned a purchase, excluding those who were only browsing, gives the reasons. They fall into five groups, and each group points to a different part of the checkout.

Money: costs and totals

Surprise costs are the biggest single reason by a distance: 40% left over extra costs, and 12% because they couldn’t see the total up front. The fix is rarely lower prices. It’s showing delivery cost, or a free-delivery threshold, before the checkout starts.

Trust and accounts

19% didn’t trust the site with their card, and 18% left because they had to create an account. Familiar payment methods and a guest checkout address both, and neither costs much.

Friction and errors

17% found the checkout too long or complicated, and another 17% hit errors or a crash. Baymard’s checkout research found the average US checkout shows 23.48 form elements where 12 to 14 are enough, and puts the possible conversion gain from better checkout design at 35.26% for a large site.

Delivery and payment

20% left because delivery was too slow, 10% because their card was declined and 9% because there weren’t enough ways to pay. These are partly business decisions, delivery promises and payment providers, not just design.

Find the leak in your own store

The reasons above are everyone’s. Your checkout loses people at a specific step, and your analytics can show which one.

GA4’s checkout journey report

Google Analytics 4 has a checkout journey report under Monetization. It shows how many people began checkout and how many completed each following step: adding shipping, adding payment and purchasing. Google describes it as a closed funnel, counting only people who started at the first step, which makes the drops between steps easy to compare.

Read the drop, then its cause

Google’s own example is the right way to think about it: a large drop between adding shipping and adding payment points to the shipping step, such as high delivery costs or a login required without a guest option. Each step’s drop maps to a group of reasons in the table above.

It needs the right events

The report only works if the store sends the begin_checkout, add_shipping_info and add_payment_info events. Some stores send only begin_checkout and purchase, which shows the leak exists but not where. If your report has empty steps, the tracking is the first thing to fix.

Split it by device and country

A checkout that works well on a laptop can lose half its phone users at the payment step. Compare the funnel for mobile and desktop, and for each country you sell to. The gap between them is often bigger than any gap against a benchmark.

Watch it happen

Numbers show where people leave; watching sessions shows why. Place a test order yourself on a phone, with a card that’s never been used on the site, as a guest, to an address outside your home town. Most checkout problems are obvious the first time you meet them as a customer.

Fix the usual suspects, cheapest first

Most checkout leaks come from a short list of causes, and several cost almost nothing to fix.

Show delivery costs before checkout

Put delivery prices, or a free-delivery threshold, on the product and basket pages. It won’t stop everyone who thinks delivery is too expensive, but it stops them discovering it at the last step, which is when it hurts most.

Offer a guest checkout

Let people buy without creating an account, and offer the account after the order is placed. It addresses the 18% who left because an account was required.

Cut the fields

Remove every field you don’t need to take payment and deliver the order. Company name, a second address line, a separate billing address by default and a phone number you never call all add effort. Baymard’s 12 to 14 form elements is a useful target.

Offer the payment methods your customers use

Card, the common wallets, and whatever buy-now-pay-later your customers expect in your market. A checkout that doesn’t show a familiar method loses the 9% who wanted another way to pay and some of the 19% who didn’t trust it.

Remove the errors, and the waiting

Test the checkout on real phones after every app or theme change, and keep it fast. A slow payment page or a script that breaks the address lookup can quietly lose orders for weeks. Why a Shopify store loads slowly covers the speed side.

Recover the ones who leave

Some people will always leave. A good share of them can be brought back, if you do it within the rules.

Abandoned checkout emails

Shopify can email customers automatically after they abandon a checkout, with a link back to their basket, and lets you choose who receives it and how long after abandonment it sends. Its abandoned checkouts page tracks each email and whether the checkout was recovered, and a report shows the sessions, orders and sales the emails brought back. Shopify notes this works for the online store and Buy Button channels.

UK email rules

Under PECR, marketing emails need consent, with a limited exception, the soft opt-in, for people who have bought or negotiated to buy a similar product from you. The ICO’s guidance requires a clear chance to opt out when you collect the address and in every message, and it doesn’t spell out whether an abandoned checkout counts as negotiating to buy. The safe approach is a clear marketing opt-in or opt-out at the email field in checkout. This isn’t legal advice; if your recovery emails are a big part of revenue, get it checked.

Don’t teach people to wait for a discount

Shopify can add a discount to recovery emails automatically. Used on every abandoned checkout, it teaches regular customers that leaving the checkout is the way to get money off. Save discounts for high-value baskets or first-time customers, and measure whether they add orders or just reduce margin on orders you’d have had anyway.

Measure recovery honestly

Some people recovered by an email would have come back anyway. Compare recovery rates with and without the email, or before and after it started, and count the margin, not just the orders.

A checkout review, in order

Put a number on it

The checkout cost calculator turns checkouts started, completion rate and order value into a monthly figure, so you can weigh any fix against what it could recover.

When to get help

If the funnel shows where people leave but not why, or the tracking isn’t there to show either, a health check on the checkout reads the path from basket to order, measures each step and sets out what to fix first.