How Checkout Friction Costs You Revenue (And 5 A/B Tests to Fix It)
Quick answer
Checkout friction — unexpected costs, forced account creation, long forms, hidden payment options, and unclear progress — is the primary driver of cart abandonment, with the Baymard Institute estimating that over 70% of shoppers leave without completing a purchase. The five highest-leverage A/B tests to address it are: enabling guest checkout as the default path, surfacing shipping costs earlier in the funnel, reducing non-essential form fields, showing trusted payment logos prominently before checkout begins, and adding a step progress indicator to make the endpoint visible.
Key takeaways
- The Baymard Institute puts average cart abandonment above 70%, and their research identifies unexpected costs, forced account creation, and over-long forms as the most common preventable causes.
- Checkout friction most often appears as friction-by-surprise — costs, steps, or requirements the shopper did not anticipate until they were already mid-funnel and losing motivation.
- Each of the five tests below targets a specific, documented abandonment trigger, making them high-priority candidates for any ecommerce experimentation roadmap.
Every ecommerce team knows their checkout flow leaks revenue. What is harder to pin down is exactly where those leaks are, how large they are, and in what order to fix them.
That ambiguity is why cart abandonment has remained stubbornly high for more than a decade — not because merchants don't care, but because the problem is structural. Shoppers abandon for different reasons at different points in the flow, and the causes are not always obvious from session recordings or heatmaps alone.
The Baymard Institute has studied online checkout usability more systematically than almost anyone. Their large-scale research, which pools data across thousands of U.S. and European shoppers, consistently finds cart abandonment rates above 70% on average — with some sectors considerably higher.
More importantly, Baymard distinguishes between "just browsing" abandonment (which no UX change will fix) and abandonment caused by concrete checkout problems that can be addressed. Their estimates suggest that roughly 35% of those losses are recoverable through improved checkout design, which at any meaningful revenue scale is a significant opportunity.
What makes this research useful for experimentation is that Baymard's work identifies specific friction points rather than vague experiential dissatisfaction. Shoppers leave because of a surprise they didn't anticipate: a shipping cost that appears for the first time at the payment step, a mandatory account creation wall, a form that asks for information they'd rather not provide.
Each of these is a testable hypothesis, not just a design preference. The goal of A/B testing checkout pages is to replace assumptions about what reduces friction with measured evidence from your own traffic and customer base.
Before running any of the tests below, it is worth establishing a clean baseline. Instrument your checkout funnel with event tracking at each step — cart view, checkout entry, address entry, payment entry, order review, and confirmation.
Step-by-step drop-off data tells you which friction points are actually costing you the most on your specific site, which determines where to test first.
How checkout friction shows up in ecommerce
Checkout friction is not a single thing. It is a category of experience problems that share a common effect: they cause a shopper who intended to buy to change their mind, or at least to pause long enough that the moment passes.
Baymard's research identifies unexpected shipping costs as the single most-cited reason shoppers abandon checkout — cited by roughly half of all abandoners in their surveys. The second and third most common reasons are site-required account creation and an over-complicated or over-long checkout process.
These three causes alone account for the majority of preventable abandonment.
What these causes share is the element of surprise, or what UX researchers sometimes call "expectation violation." The shopper builds a mental model of the transaction — this item costs X, I'll pay Y, I'll give my address and card and be done in two minutes — and then the checkout process contradicts that model.
The shipping cost turns out to be $14.99. There is a mandatory registration form before they can even enter an address.
The checkout has six steps instead of two. Each of these violations creates a decision point: is this purchase still worth the extra friction?
For a meaningful fraction of shoppers, the answer becomes no.
The practical implication is that reducing checkout abandonment is less about making checkout "easier" in a generic sense and more about eliminating expectation violations. Shoppers are not inherently unwilling to complete multi-step checkouts — they do it every day.
What they are unwilling to do is discover, mid-flow, that the transaction they mentally agreed to is not the transaction they are actually completing. The tests below each address a specific class of expectation violation, framed as a concrete change you can validate through experimentation.
What to A/B test to reduce checkout abandonment
1. Make guest checkout the default path, not the exception
Account creation is one of the most studied friction points in ecommerce checkout research. Baymard Institute's findings are clear: when sites force shoppers to create an account before they can buy, a significant proportion of them leave — particularly first-time visitors who have no prior relationship with the brand and no intrinsic motivation to register.
The value exchange makes sense from the merchant's perspective (email capture, repeat purchase data, loyalty program enrollment), but from the shopper's perspective it is an uncompensated demand placed between them and the thing they want to buy.
The test here is not simply adding a guest checkout option — most sites have one. The question is how prominently that option is presented relative to account creation, and whether the default state of the checkout landing screen guides shoppers toward registering or toward completing a purchase.
Many checkout flows frame account creation as the primary action and guest checkout as the escape hatch. Reversing that hierarchy — making "Continue as guest" the prominent primary action and "Create an account" the secondary option — changes the experience without removing account creation as a possibility.
You can also test a softer version of this: presenting checkout as a single path and offering account creation as an opt-in at the order confirmation step, after the purchase is complete. At that point, the shopper is satisfied, trust has been established, and the value of saving their details for next time is much more tangible.
Baymard's research supports post-purchase account creation as a high-conversion prompt compared to pre-purchase registration walls.
The test: Control shows a standard checkout entry screen with account creation as the primary CTA and guest checkout as a secondary link. Variant makes guest checkout the primary CTA and removes or de-emphasizes the account creation prompt until post-purchase.
Your hypothesis is that routing first-time visitors through guest checkout by default reduces the drop-off rate at the checkout entry step and increases completed orders in a single session.
2. Surface shipping costs before checkout begins
Unexpected shipping costs are the number-one reason shoppers abandon checkout, according to Baymard Institute's large-scale abandonment research. The word "unexpected" is doing the most important work in that finding.
Shoppers are not necessarily unwilling to pay for shipping — they are unwilling to discover the cost at the payment step, after they have already spent time filling out an address form and become mentally committed to the purchase. The gap between expectation and reality at that moment is large enough to break the transaction.
The fix is not necessarily to offer free shipping (though that eliminates the friction entirely if it is economically viable). The fix is to surface the actual or estimated shipping cost as early as possible in the funnel — ideally on the product page or in the cart view, before the shopper has clicked into checkout at all.
A prominent shipping cost estimate at the cart stage sets accurate expectations. By the time the shopper reaches the payment step, the cost is already factored into their decision.
There is no surprise, no expectation violation, no pause.
Testing this requires a few design decisions: where in the funnel to surface the estimate, how to handle variable shipping costs (zip-code-based calculators vs. flat-rate messaging vs. a range), and whether to show it as a dedicated line item or as part of an order summary widget. Each of these variables can be tested independently, but the highest-leverage starting point is the simplest version: showing a shipping cost estimate or a "free shipping on orders over $X" indicator in the cart view rather than waiting until the payment step.
The test: Control displays shipping cost for the first time at the payment/review step of checkout. Variant adds a shipping cost estimate or free shipping threshold indicator to the cart page, visible before checkout begins.
Your hypothesis is that showing shipping cost information at the cart stage reduces abandonment at the payment step, because shoppers arrive at that step with expectations already set.
3. Remove non-essential form fields to shorten checkout
Form length is one of the most direct inputs to checkout completion rate. Every additional field is a unit of effort that the shopper must decide is worth the transaction.
Some fields are genuinely necessary — shipping address, email, payment details. Others are collected out of habit, for internal data purposes, or because they were in the original checkout template and no one has questioned them since.
Phone number, company name, and a separate billing address when the shopper has already indicated it is the same as shipping are the most common examples of fields that cost effort without adding value for the shopper.
Baymard's checkout usability research finds that the average checkout contains nearly twice as many form fields as are strictly necessary to complete a transaction. Their benchmark across major ecommerce sites shows that a well-optimized checkout can be completed in fewer than 15 form fields for a new customer, but many sites require 20 or more.
Each unnecessary field increases cognitive load and the time required to complete the process — two factors that compound, particularly on mobile where keyboard input is more effortful.
The test here should be specific rather than a wholesale redesign. Identify the fields in your current checkout that are optional for fulfillment purposes and test removing them one at a time, or as a small group.
Phone number is often the easiest candidate — it is used by fewer shipping carriers than merchants assume and can be made optional or removed entirely without operational impact for most SKUs. Billing address auto-population (defaulting to "same as shipping" and collapsing the billing address form unless the shopper explicitly changes it) is another high-yield change that removes several fields from the typical flow.
The test: Control shows the full current checkout form including phone number, company name, and a separate billing address section. Variant removes the phone number field, removes company name, and auto-collapses billing address to "same as shipping" with a toggle to expand.
Your hypothesis is that reducing the form to essential fields shortens time-to-completion and increases the share of shoppers who reach the order confirmation step.
4. Show trusted payment logos prominently before checkout begins
Payment anxiety is an underrated source of checkout friction. Shoppers want to know, before they invest time in a checkout flow, that their preferred payment method is accepted and that the transaction will be secure.
If that information is absent or only appears at the payment step, some shoppers will hesitate — particularly those who prefer PayPal, Apple Pay, Google Pay, or buy-now-pay-later options like Klarna or Afterpay, which offer a layer of purchase protection or payment flexibility that a plain credit card form does not.
Research from the Baymard Institute on payment UX highlights that payment method trust signals function differently from general site trust signals (SSL badges, security certifications). Where general trust badges communicate "this site is legitimate," payment logos communicate "you can pay the way you prefer, and the transaction will be familiar."
The latter is more specific and more actionable for shoppers who have strong payment method preferences — which includes a growing share of mobile shoppers who use Apple Pay or Google Pay as their default and find entering card details manually to be a meaningful friction point.
The test is straightforward: add a row of payment method logos (Visa, Mastercard, PayPal, Apple Pay, Google Pay, and any BNPL providers you support) to the cart page or product page, above the checkout CTA, rather than only showing them at the payment step. You can also test positioning — in the cart summary, directly under the checkout button, or as a footer element on the product page.
The variant that performs best will likely depend on your specific customer mix and how many of your shoppers use alternative payment methods.
The test: Control shows payment method logos only at the payment entry step of checkout. Variant adds a compact row of accepted payment logos to the cart page, positioned near or below the primary checkout CTA.
Your hypothesis is that making accepted payment methods visible before checkout begins reduces drop-off among shoppers who prefer non-card payment methods and increases overall checkout entry rate.
5. Add a progress indicator to make the checkout endpoint visible
One of the subtler but well-documented sources of checkout anxiety is uncertainty about how long the process will take. A shopper who clicks "Checkout" and encounters a form with no indication of how many steps follow has to decide, without information, whether to commit to the process.
If the form looks long or unfamiliar, or if they have been burned by unexpectedly long checkouts on other sites, that uncertainty can trigger abandonment before they've given your flow a fair chance.
A step progress indicator — a simple numbered header like "Step 1 of 3: Shipping" — does two things. First, it sets a concrete expectation: the shopper knows the process has a defined endpoint, and they can see how far through it they are at any given moment.
Second, it creates a mild completion effect: once a shopper has completed Step 1, the cost of abandoning at Step 2 feels larger because they've already invested effort. This is a well-studied behavioral principle — partial completion increases follow-through — and it applies to multi-step checkout flows in the same way it applies to forms, surveys, and onboarding sequences.
The design of the progress indicator matters. Simple numbered text headers ("Step 2 of 3") are effective and require minimal development.
Visual step bars with labeled stages ("Cart → Shipping → Payment → Review") are slightly richer and make the overall structure more legible. Both outperform no indicator at all.
Testing the two variants against a no-indicator control will tell you whether the simpler implementation is sufficient for your audience, or whether the richer design produces a meaningful additional lift.
The test: Control is your current multi-step checkout with no step indicator. Variant adds a step counter or labeled step bar to the top of each checkout page, showing current step and total number of steps.
Your hypothesis is that making the checkout endpoint explicitly visible reduces anxiety-driven abandonment in the middle steps of checkout and increases completion rates from the shipping step onward.
Free A/B Testing Tool
Run your next A/B test the right way
Visual editor, 15 KB script, GA4-native — and free forever up to 100,000 monthly visitors. No developer required.
Running these tests
The five tests above are not a ranked list to run in order — they are a menu. The right starting point depends on where your checkout is losing the most traffic right now.
If your funnel data shows the biggest drop at checkout entry, guest checkout and payment logo visibility are your highest-priority tests. If you see large drop-off between address entry and payment entry, the shipping cost and form length tests are more relevant.
If abandonment is distributed across all middle steps, the progress indicator is a good cross-cutting experiment to run first.
Prioritization aside, there are a few practical principles that apply across all five. Run one test per major checkout component at a time.
Overlapping tests on the same page can produce interaction effects that make results difficult to interpret. Set your sample size before you start — underpowered tests that are called early in favor of the variant are one of the most common sources of false positives in conversion optimization.
And define your primary metric (checkout completion rate, revenue per visitor, or completed transactions) before you look at the data.
For a more systematic approach to deciding what to test, the process of identifying friction through qualitative and quantitative research is covered in depth in the Mida guide to CRO research. If you are building out a broader testing roadmap for your ecommerce site, the Mida ecommerce testing ideas library catalogs validated test hypotheses across the full funnel — product pages, cart, checkout, and post-purchase flows.
Ready to test these hypotheses on your own checkout? Try Mida free — now available even if you don't have an account.
FAQs
Q: How do I know which checkout friction test to run first?A: Start with funnel drop-off data. Look at where the largest single-step abandonment occurs in your checkout flow. If the biggest drop is at checkout entry, test guest checkout first. If it is at the payment step, shipping cost visibility is usually the higher-priority test. Let your own data set the priority order rather than following a generic sequence.
Q: Do I need a large amount of traffic to run these tests?A: Statistical significance requires a meaningful sample size, and the required sample depends on how large a difference you expect the variant to produce. For checkout tests, where effect sizes can be significant, many stores with moderate traffic can reach reliable results within two to four weeks. Use a sample size calculator before starting to confirm your site has enough volume to detect the effect size you are testing for.
Q: Can I run multiple checkout tests at the same time?A: It is technically possible through mutual exclusion (splitting traffic so each shopper only sees one experiment), but it is operationally complex and increases the risk of interaction effects. For most teams, running checkout tests sequentially — one at a time — produces cleaner results and is easier to communicate internally.
Q: What if a test that worked for a large ecommerce brand does not work for my store?A: That is exactly what A/B testing is for. Published research and industry case studies are useful for generating hypotheses, but they reflect other companies' customers, traffic sources, and product categories. Your test results reflect your shoppers. A test that does not move the needle on your site is still a valid result — it tells you that particular hypothesis was not the bottleneck in your specific context.
Q: Should I test on mobile and desktop separately?A: If your traffic splits meaningfully between mobile and desktop (and most ecommerce sites skew heavily toward mobile), segmenting results by device type will reveal whether a change helps both audiences equally or drives results primarily in one context. Form length reduction, in particular, tends to show stronger effects on mobile where keyboard input is more burdensome. Segment your analysis even if you run a single unified experiment.
Sources
- Baymard Institute. Cart Abandonment Rate Statistics. baymard.com/lists/cart-abandonment-rate. Baymard pools data from 49+ published studies and their own research; their average cart abandonment figure has held above 70% consistently.
- Baymard Institute. E-Commerce Checkout Usability (research report). Covers form field benchmarking, account creation friction, and payment UX across major ecommerce sites.
- Baymard Institute. Checkout Optimization (topic page). baymard.com/checkout-usability. Research-backed guidelines covering over 300 checkout UX parameters.
- Nielsen Norman Group. Indicators, Validations, and Notifications: Pick the Correct Communication Option. Background on progress indicators and user expectation management in multi-step flows.
- Klarna / PYMNTS. Consumer Payments Research (2023–2024). Data on alternative payment method adoption rates and shopper preference for BNPL and digital wallets at checkout.
