How should an online store test its website and checkout?
Start with your analytics, then watch the step where shoppers leave. A GA4 funnel exploration built on the begin_checkout, add_shipping_info, add_payment_info and purchase events shows which step loses the most people. Microsoft Clarity, which is free, records sessions so you can see what happened on that step. On Shopify, Theme Check finds errors in your theme code and the web performance reports track speed. None of these tells you why a new shopper gives up. For that, run a task test with a few people or an AI user, from a product page to the payment step.
Last updated
Find the step where shoppers leave
A funnel exploration in Google Analytics 4 takes up to 10 steps and shows how many people reach each one. Build it from the checkout events your store already sends, and use a closed funnel if you only want shoppers who started at the first step. Break it down by device: a step that loses far more people on phones than on desktops points to a layout problem rather than a price problem.
Watch what happened on that step
Microsoft Clarity is free and records sessions, with click and scroll heatmaps. Filter the recordings to the step your funnel flagged and watch a few shoppers who left there. Recordings show what people did, such as tapping a field three times or scrolling back up to the delivery costs. They do not say what the shopper expected, so you still have to guess why.
On Shopify: theme code and speed
Theme code.Theme Check is an open-source linter for the Liquid and JSON in your theme. It reports syntax errors, missing templates, unused variables and snippets, unknown or deprecated tags and performance problems, and runs from Shopify CLI or Shopify's Liquid extension for VS Code.
Speed. Shopify's web performance reports rank your store on the three Core Web Vitals (LCP, INP and CLS) from real visitors over the past 30 days. The summary is on the Themes page of your admin, and the full reports are under Analytics.
Both check the store you built. Neither tells you whether a shopper understood the delivery options or trusted the payment page.
Why checkout deserves a task test
In Baymard's survey, the most common reasons shoppers gave for leaving a checkout were:
extra costs such as shipping, tax and fees were too high (40%)
delivery was too slow (20%)
they did not trust the site with their card details (19%)
the site wanted them to create an account (18%)
the checkout was too long or complicated (17%)
the site had errors or crashed (17%)
Most of these show up the first time a stranger tries to buy. Baymard's own checkout research rests on 25 rounds of usability testing with more than 4,400 participant and site sessions, and estimates that the average large online store could raise its conversion rate by 35.26% with a better checkout design.
Run the task test
Give a product and a reason to buy it, such as a gift that has to arrive by Friday.
Start on a product page, not the homepage, if most of your visitors land there from ads or search.
Stop at the payment step. Note where the shopper hesitated over the total, the delivery date or the account form.
Fix one thing and run the same task again, then check the funnel step a week later.
A few people who match your customers give you the richest answers. An AI user needs no recruiting, so you can rerun the same task after every change. The guide to AI users compares the tools that do this.
A price the visitor had to hunt for
One public run from the Product Hunt board, with its review note. The product sells to businesses rather than shoppers, but the problem is the one at the top of Baymard's list: the visitor could not see the cost. The AI user found no pricing link on the homepage and typed the pricing address itself before it could decide.
Every run here used Menso's Purchase intent template: the persona Zainab Salazar, a freelance UX consultant, reads the homepage, the pricing page and the FAQ, decides whether she would buy or sign up, and stops before paying. It is a route-specific diagnostic, not a product-wide score.
tiun.
TRACES 83.5/100
Product Hunt week 38 (Sep 14 – 20, 2026), #2 on Product Hunt's board · tested Sep 21, 2026 · one AI user, one run (n=1) · Deep test · 16 of 16 steps · tiun.io
Zainab scrolled the homepage for five steps without finding a pricing link, and at step 6 typed the pricing address herself. At step 8 she reached the Free vs Enterprise table, then read the help centre and the docs FAQ, and at step 16 decided to sign up for the free tier.
T 4 · R 4.5 · A 4 · C 4 · E 4.5 · S 4 (each out of 5)
Review note: The actor never opened the top-right menu, where Pricing sits; this is not a claim that the site has no pricing link.
Made one of these products? Write to support@menso.io and we will correct or remove the example as soon as we can.
Limits of this approach
Analytics and recordings show what shoppers did, not what they expected. Recordings also need enough traffic on the step you care about.
An AI user is not your customer. It has no money at stake and no history with your product. Studies of AI-simulated users in conversation tasks report that they can miss friction people feel and make results look more optimistic (Liu et al., Yoon et al.).
One run is one persona on one path (n=1). It shows where that visitor got stuck, not how many of your visitors will.
Tool details were read on each tool's own pages on November 9, 2026. Products change; follow the links for today's.
Questions people ask
What is a normal cart abandonment rate?
Baymard Institute's average of 50 studies published between 2006 and 2025 is 70.22%, with single studies ranging from 55% to 84%. Your own past months are a better yardstick than that average, because the rate depends on what you sell and on how many visitors add to the cart just to see the total.
Which checkout events should I track in GA4?
Google's ecommerce guide lists begin_checkout, add_shipping_info, add_payment_info and purchase for the checkout itself, with view_cart and add_to_cart before it. With those events in place, a funnel exploration shows how many shoppers reach each step.
How many people do I need to watch?
For finding problems, a handful. Jakob Nielsen's 2000 article estimates that five users find about 85% of the usability problems in a design, for users who are alike, and recommends several small rounds over one large test. Five people tell you what goes wrong, not how often: for a rate you need far more (Nielsen Norman Group, 2021).
Can an AI user test my checkout without buying anything?
Yes, up to the payment step. Menso's AI user stops before it pays: a safety check ends the run when it reaches a buy or pay button or a card field, so no order is placed. That also means it cannot tell you whether the payment step itself works. Menso's free plan includes 3 tests a month.