Skip to content
CRO12 min read

Why Customers Abandon Checkout: Causes and a Checkout Audit Checklist

Why shoppers abandon checkout, plus a checkout UX audit checklist that pairs each check with how to test it, the evidence to collect and the likely fix.

01

Why do customers abandon checkout, and how do you find out which reasons apply to you?

Most checkout abandonment comes from four things: a total that turns out higher than expected, effort (accounts, long forms, awkward steps on a phone), doubt (about the site, delivery or returns) and outright failure (errors and declined payments). Industry surveys tell you which causes are common. Only a checkout audit tells you which ones are costing your store orders. An audit walks every step from cart to confirmation, notes possible friction, and then confirms or rules out each suspicion with evidence before anyone spends money on fixes.

This guide covers both halves: what the research says about why shoppers leave, and a checklist for auditing your own checkout, with a column for the evidence that proves each problem. It is written for ecommerce managers and founders who want to diagnose before they redesign. For how to design each checkout step once you know what is wrong, see the checkout UX design guide. If shoppers are leaving before they reach checkout at all, start with why shoppers add to cart but don't buy.

02

Key takeaways

  • Baymard's 70.22% average abandonment rate includes people who were only browsing. Measure your own baseline by step instead.
  • Unexpected extra costs are the most common stated reason, ahead of slow delivery, distrust and forced accounts.
  • An audit is only as good as its evidence. Pair every suspected problem with data that could prove it wrong.
  • Separate findings into fix now, test first and research more. Not every finding needs an A/B test.
  • Place real test orders on phones, with each payment method you offer. Many failures only appear there.
03

What the research says

Baymard Institute keeps a running average of published cart abandonment studies. Across 50 studies it stands at 70.22%, on a page last updated on 22 September 2025; no newer figure had been published when this guide was revised. That number includes a large group who were never going to buy that day: in the same research, 42% of US shoppers who abandoned said they were just browsing or not ready to buy.

Once those browsers are set aside, Baymard's US respondents gave these reasons for leaving during checkout (they could choose more than one):

Stated reason for leavingRespondents
Delivery, tax or other extra costs were too high40%
Delivery would take too long20%
Did not trust the site with card details19%
Had to create an account18%
Checkout felt too long or complicated17%
The site showed errors or crashed17%
Returns policy was not good enough13%
Could not see or work out the total cost in advance12%
Card was declined10%
Preferred payment method was missing9%
04

How to read those numbers

Treat the list as a set of hypotheses, not a diagnosis. The survey covers US shoppers; your customers, products and markets may weigh the causes differently. A store with free delivery will not lose many people to delivery costs, while a store selling bulky goods to several countries might lose a great many.

Baymard also estimates that a large ecommerce site could lift its conversion rate by 35.26% through better checkout design. That is a modelled estimate for large sites, based on usability issues Baymard has documented as fixable. It is not a figure any individual store should expect.

The useful question is narrower: at which step do your shoppers drop, on which devices and in which markets, and what evidence explains it? That is what the audit below is for.

05

The causes, grouped by what the shopper experiences

Cost surprises. The shopper had a price in mind, and checkout raised it. Delivery charges, taxes, duties or fees that only appear late are the classic trigger. The total should be predictable from the product page and the cart, not revealed at the payment step.

Effort. Every account prompt, extra field and confusing step gives the shopper a reason to stop. Baymard's benchmark of large sites counted 11.3 checkout form fields on average, against the 8 most sites actually require, and it found that the number of fields matters more than the number of steps. Typing is harder on a phone, so effort problems tend to be worse on mobile.

Doubt. Shoppers hesitate when checkout looks unlike the rest of the store, when the payment options are unfamiliar, when delivery dates are vague, or when returns terms are hard to find. Unexplained requests add to it: Baymard found 39% of benchmarked sites asked for a phone number without saying why, and 14% of shoppers it surveyed said they would never give one.

Failure. Some shoppers want to buy and simply can't: a validation rule rejects a real address, a discount code breaks the page, a card is declined with no explanation, or a wallet button does nothing inside an in-app browser. These are often the cheapest problems to fix and the easiest to miss, because nobody on the team checks out the way customers do.

06

How to run a checkout UX audit

The method below combines a heuristic walkthrough with data, so every finding ends up with evidence attached. Work through it in order; later steps depend on earlier ones.

  • Check the measurement. Confirm that each checkout step is tracked and that the numbers match orders in your back office. An audit built on broken tracking will point at the wrong step. A tracking audit is the place to start if you are unsure.
  • Map the funnel by step. Record how many sessions reach cart, start checkout, complete each step and pay. Split by device, new versus returning customers, country and traffic source.
  • Find the biggest drop that you can influence. Compare steps and segments with your own history, not with industry averages.
  • Walk the flow yourself. Go through checkout as a first-time customer on a phone and on a desktop, and note every moment of hesitation, surprise or extra effort.
  • Place test orders. Cover each payment method, a guest and a signed-in customer, a discount code, an international address if you ship abroad, and at least one in-app browser such as Instagram's.
  • Watch recordings of real sessions. Filter for sessions that reached checkout and didn't finish. Look for rage clicks, dead clicks and quick backs, as defined in tools such as Microsoft Clarity.
  • Collect the customer's voice. Read recent support tickets and chat logs about checkout, delivery and payment, and if volume allows, run a one-question exit survey on checkout pages.
  • Score and sort the findings. Use the checklist below to record evidence for each finding, then sort them into fix now, test first or research more.
07

Checkout audit checklist

Use this as a working sheet. For each row, run the test, collect the evidence and only then decide on the fix. The fixes listed are starting points; the checkout design guide covers how to design each one properly.

CheckHow to testEvidence to collectLikely fix
Cart shows a realistic totalAdd items as a new visitor in each market you sell toDrop from cart to checkout start, by country and cart valueShow delivery cost or the free-delivery threshold in the cart; show tax the way your market expects
Guest checkout is the obvious pathStart checkout as a new visitor and note what you are asked firstDrop at the first checkout screen; clicks on sign-in versus guestMake guest the default and offer an account after purchase (design options)
Only necessary fields are requestedCount fields on each step; ask what each one is used forTime per step; field-level error counts; recordings of hesitationRemove unused fields, default billing to delivery address, explain any phone request
Address entry accepts real addressesEnter addresses from each market, including flats and rural formatsValidation errors by field and country; support tickets about addressesLoosen strict validation; add address lookup where it works well
Delivery options show dates and pricesReach the delivery step for different postcodes and cart sizesDrop at delivery step; delivery questions to supportShow expected arrival dates and costs, not just service names
Customers' preferred payment methods are offeredCompare offered methods with what customers in each market usePayment provider reports; requests for missing methods; drop at payment by countryAdd the missing wallets or local methods that the evidence supports
Payment failures are recoverableUse your provider's test cards to trigger a declineDecline and failure rates by method, device and countryExplain the failure plainly, keep the order and entered details, offer another method
Errors keep what the shopper typedSubmit each step with deliberate mistakesRecordings of repeated submits; error log entriesValidate inline, explain how to fix the error, never clear valid fields
Discount codes behave predictablyApply valid, expired and invalid codesCode error rate; drop after a failed codeClear error messages; don't make the code field the most prominent element
Order review shows everything before paymentRead the final step as a cautious first-time buyerExit survey answers about cost or delivery; back-navigation from reviewShow items, total, delivery date and returns summary together
Checkout looks like the same storeCompare branding, fonts and tone across store and checkoutTrust-related exit survey answers; new versus returning customer gapConsistent branding; recognisable payment options; visible help and policies
Checkout works on phones and in in-app browsersTest orders on iOS and Android, including social app browsersCompletion rate by device, browser and traffic sourceFix keyboards, tap targets and wallet behaviour (mobile checkout fixes)
Steps load quickly for real usersCheck field data for checkout-bound pages; time payment processingCore Web Vitals at the 75th percentile; slow payment confirmationsRemove heavy scripts; show progress during payment; prevent double submission
Confirmation reassures and sets expectationsComplete a test order and read the confirmation page and emailSupport tickets asking whether an order went throughState the order number, delivery estimate and next steps clearly

Pro tip

Write the evidence down before you decide the fix. A row with no evidence is a hunch, and hunches belong in the research-more pile.

08

Fix now, test first or research more

Not every finding deserves the same treatment. A simple rule of thumb is to weigh three things: how severe the problem is for those who hit it, how many shoppers reach that step, and how confident the evidence makes you. This is a judgement aid, not a formula with a correct answer.

BucketWhen it appliesWhat to do
Fix nowA clear defect with strong evidence: a broken field, a failing payment method, an address format that is rejectedFix it, then confirm the drop at that step improves against your baseline
Test firstA plausible improvement where the best answer is uncertain and the step has enough trafficRun a controlled test; see what you can safely test in checkout
Research moreA suspicion with weak or conflicting evidenceGather more recordings, survey answers or test orders before changing anything
09

Hypothetical example

Hypothetical example: a homeware store sees checkout completion fall on mobile while desktop holds steady. The funnel shows the biggest mobile drop at the delivery step. Recordings show shoppers scrolling the delivery options and leaving. Test orders reveal that, for some postcodes, the cheapest option says only "Standard" with no date, and that the free-delivery threshold is mentioned nowhere before this step.

The audit records two findings. Missing delivery dates are a clear gap with good evidence, so they go in fix now. Whether to show the free-delivery threshold in the cart, and how, is less certain and could affect order value, so it goes in test first. After the fix, the team compares the mobile delivery-step drop with the weeks before it, and plans the threshold test with the prioritisation worksheet.

10

Market and platform differences

Payment expectations and address formats vary by country, so audit each major market separately rather than assuming one checkout suits all. In India, check that the methods customers actually use are offered and tested. In the UAE, payment options and address conventions differ again; see checkout optimisation for UAE shoppers. For a wider view of international payment methods, currencies and addresses, see designing checkout for international customers.

On Shopify, what you can change in checkout depends on your plan and on checkout extensibility, so some fixes in the checklist may need an app, a configuration change or a Plus feature. The Shopify checkout audit applies this method to Shopify's checkout specifically.

11

When a checkout audit is not the answer

If most shoppers leave before starting checkout, the problem is upstream: product pages, cart, pricing or the offer itself. Fixing checkout will not help much. If traffic is very low, you may not have enough sessions to see reliable step-level patterns; lean more on test orders, recordings and conversations with customers, and treat conclusions as provisional. And if the real issue is that your prices or delivery costs are uncompetitive, no amount of interface work will close the gap. The audit will usually make that visible, which is useful in itself.

12

Common mistakes

  • Benchmarking against a global average instead of your own step-by-step history
  • Redesigning the whole checkout when one step or one payment method is the problem
  • Auditing only on a desktop, signed in as an admin
  • Acting on recordings or survey answers alone, without checking the funnel data
  • Sending every abandoner a discount, which teaches customers to wait
  • Treating trust as a badge problem rather than a consistency and clarity problem
13

Conclusion

The research tells you where to look: costs, effort, doubt and failure. The audit tells you which of those your own customers are running into, with evidence to back each finding. Check your tracking, find the biggest controllable drop, walk the flow, place real test orders, then sort what you find into fix now, test first and research more. If a slow checkout is part of the picture, estimate what slow checkout steps cost you before deciding how much to invest.

Want your checkout audited with evidence, not opinions?

ZSpace Labs is a remote-first technology studio that works with clients globally on CRO audits. If useful, we can talk through your checkout funnel, the evidence you already have and where an audit would add most.

Start a Project
14

Sources

Baymard Institute: cart abandonment rate statistics (last updated 22 September 2025); Baymard Institute: checkout flow average form fields; Baymard Institute: explain the phone number field; Baymard Institute: how to audit your checkout flow for hidden friction (13 May 2026). Baymard is a commercial UX research firm; its figures are industry research, mostly from US shoppers.

web.dev: Web Vitals; Microsoft Clarity: semantic metrics (vendor documentation).

Accessed 15 October 2026.

FAQ

Common questions.

Baymard asked US shoppers who had abandoned during checkout why they left, setting aside people who were only browsing. Extra costs came first at 40%, followed by slow delivery (20%), distrust of the site with card details (19%), a required account (18%), and a checkout that felt too long or complicated or that threw errors (17% each).

Baymard's average across 50 studies is 70.22% (page last updated 22 September 2025). It counts everyone who adds to cart and doesn't buy, including people saving items or comparing prices, so it is not a target. Track your own rate by checkout step and compare it with your own past performance.

A structured review of every checkout step, from cart to confirmation, that looks for friction and then confirms or rules out each suspected problem with evidence: funnel data, field and payment errors, session recordings, test orders and customer feedback.

The audit finds and proves the problems; design work fixes them. Run the audit first, then use a design guide such as our checkout UX design guide to decide how to change each step.

At minimum, a step-by-step checkout funnel you trust, split by device. Payment failure reports from your payment provider, error logs, session recordings and a sample of support tickets make the audit much stronger. If tracking is unreliable, fix that first.

No. Clear defects, such as a broken field or a payment method that fails, should simply be fixed. Test changes where the right answer is uncertain and traffic is high enough. Where neither applies, do more research before changing anything.

The method is the same, but what you can change depends on your plan and checkout extensibility. See the Shopify checkout audit for the platform-specific version.

After any significant change to checkout, payment providers, delivery options or tracking, and otherwise on a regular cycle that suits how often your store changes. Re-run test orders whenever a theme, app or payment setting changes.

Rarely on their own. Distrust is usually about the whole experience: a checkout that looks different from the store, unfamiliar payment options, hidden policies or no visible way to get help. Check those first.

Get in touch

Have a project in mind?

Whether you're building a new digital product, improving an existing website, or looking to automate part of your business — let's talk.