← Blog

Checkout abandonment analysis: find where you lose sales in 2 hours

For e-commerce and operations managers with a campaign coming: a two-hour checkout diagnosis with GA4, Clarity and a speed test, so you know what to fix first.

Get a summary withChatGPT Claude Google AI Mode
Abstract funnel with light dots flowing through and some leaking out the sides, showing checkout abandonment analysis
On this page

Your campaign starts in a few weeks. The checkout leaks, and you are not sure whom to pay or what to ask them to fix. This guide is for e-commerce and operations managers in exactly that spot.

A checkout abandonment analysis shows which step loses shoppers, on which device, and what they were doing when they left. You can run one yourself in about two hours with GA4, Microsoft Clarity and a speed test. All three are free.

By the end you hold a ranked list of causes, a way to prove each fix and a brief any developer can price. If the leak is one payment bug or an unclear shipping message, you will see that too, and you will not pay for weeks of work you do not need.

Key takeaways

  • A two-hour checkout diagnosis with GA4, Clarity and a speed test costs you nothing in tools and gives you a ranked list of causes.
  • Six GA4 events, from add_to_cart to purchase, map your checkout, and a funnel exploration shows which step loses the most shoppers.
  • Splitting that funnel by device, payment method and country turns one average into one named leak.
  • Clarity recordings show what shoppers did at that step, so you fix the real cause and not a guess.
  • Fix bugs straight away, test choices with an A/B test, and hand a developer an evidence pack they can price from.

Checkout abandonment analysis in two hours

The short answer: read three free sources of evidence in a fixed order. GA4 tells you where shoppers leave. Clarity shows what they were doing just before. A speed test tells you whether slow loading is part of it. Together they turn a vague worry about the checkout into a short list of named causes.

What you need before you start

Gather these before the clock starts. Each one saves you a stop later.

  • A GA4 property that already receives ecommerce events, or someone who can add them (the next section lists which).
  • A Microsoft Clarity project on your store. It records only from the day you install it, so install it today if it is missing.
  • Admin access to your store, to compare GA4 purchases with real orders.
  • A phone and a laptop, to place one test order on each.
  • A document where you write each finding as one sentence, with its date range.

If Clarity is new for you, give it three to five days of traffic before the recordings step. The GA4 part works on history, so you can start it now.

The plan, hour by hour

The two hours split into four blocks. Each ends with something you write down, so the analysis never stays only in your head.

  1. 0:00 to 0:30

    GA4 events and funnel

    Check the six events, build the funnel exploration and run the four data checks.

  2. 0:30 to 1:00

    Device and payment split

    Break the funnel down by device, payment type and country. Name the step that leaks.

  3. 1:00 to 1:40

    Clarity recordings

    Watch recordings of that step on that device. Note rage clicks, dead clicks and errors with timestamps.

  4. 1:40 to 2:00

    Speed test and written list

    Test the speed of the leaking step, then write the causes in fix-first order.

Keep the order. GA4 points at the step, Clarity explains it, and the speed test checks the one cause you cannot see in a recording.

Funnel, magnifying lens and hourglass on a desk, representing a two-hour checkout diagnosis in three steps
Three sources of evidence, read in a fixed order

Set up your GA4 checkout funnel

A funnel report in GA4 needs the checkout events to arrive in order. Google documents a recommended set of ecommerce events, and using those exact names lets the reports read your store. The GA4 ecommerce measurement guide (opens in a new tab) lists them with their parameters.

The six events that map your checkout

Six events cover the path from cart to paid order. Each one answers a plain question:

  • add_to_cart: did the shopper put something in the cart?
  • view_cart: did they open the cart page or drawer?
  • begin_checkout: did they start checkout?
  • add_shipping_info: did they complete the delivery step?
  • add_payment_info: did they choose and submit a payment method?
  • purchase: did the order complete?

view_item (a product page view) is a useful seventh event at the top of the funnel. This analysis starts at the cart, where the money is closest.

How the events arrive depends on your platform. On Shopify, the Google and YouTube channel app or a custom pixel in customer events can send them, and what you can add inside checkout depends on your plan. On WooCommerce, a GA4 plugin or Google Tag Manager sends them from the cart and checkout pages, and a developer can add missing ones with hooks (the points in the code where WooCommerce lets you add behaviour). Before you build anything, open GA4 DebugView and place one test order to see which events fire.

The diagram shows what you are looking for: visitors enter at the top, some leave at every step, and one gap stands out.

Checkout funnel events

  1. 01Add to cartShopper puts a product in the cart
  2. 02View cartCart page or drawer opens
  3. 03Begin checkoutShopper starts checkout
  4. 04Add shipping infoBiggest gap, look here first
  5. 05Add payment infoPayment method submitted
  6. 06PurchaseOrder completes

Funnel of six GA4 checkout events showing visitors dropping off at each step, with the biggest gap highlighted as the step to investigate first

Visitors thin out at every step, and the biggest gap is where to look first

Build the funnel in Explorations

Once the events arrive, the report takes about ten minutes to build:

  1. In GA4, open Explore and create a Funnel exploration.
  2. Add the six events as steps, in the order above.
  3. Set the date range to at least the last 28 days, longer if your traffic is low.
  4. Add Device category as a breakdown, so every step shows mobile, desktop and tablet.
  5. If a step looks odd, compare the closed funnel (shoppers must follow the order) with the open one (they may join at any step).

Read the report as a list of drop-offs, not as one conversion rate. The step with the largest gap between its count and the next is where the analysis goes first.

Check the data before you trust it

A funnel built on broken tracking points you at the wrong fix. Four checks take about 15 minutes:

  • Compare GA4 purchases with orders in your store admin for the same dates. A gap of a few percent is normal, because of consent banners and ad blockers. A large gap means tracking needs repair first.
  • Confirm no step has more users than the step before it.
  • Confirm purchase fires once per order and not twice.
  • Place one test order on mobile and watch DebugView confirm every event.

If a check fails, write it down. It is your first finding, and often the cheapest fix of the day.

Read checkout drop-off by device and payment

A single average hides the leak. This block turns a general worry about the checkout into a sentence such as: mobile shoppers leave at the shipping step.

Compare mobile and desktop step by step

With Device category as the breakdown, read each step for mobile and desktop side by side. A gap on every step usually means the whole experience is harder on a phone. A gap on one step points at a specific form or script.

Here is an illustrative example for a store with 1,000 checkout starts. The numbers are made up to show how to read the report.

StepMobile (600 starts)Desktop (400 starts)
begin_checkout600400
add_shipping_info330310
add_payment_info240270
purchase150230

In this example mobile loses 45 percent of its shoppers between starting checkout and the shipping step, while desktop loses about 22 percent. Mobile ends at 25 percent of starts completing a purchase, desktop at 57.5 percent. The first thing to inspect is the mobile shipping form, not the payment page.

Split by payment method and country

The add_payment_info event can carry a payment_type parameter. GA4 shows a custom parameter in reports only after you register it as a custom dimension in Admin, and it collects from that day on. Once it is there, add it as a breakdown. If one method converts far worse than the others, you may be looking at a payment error and not a design problem.

Country is a standard dimension, so add it as a second breakdown. One country that drops at the shipping step points to shipping cost or delivery options for that country. One that drops at payment often points to a missing local payment method. For stores selling in Spain and Latin America, this split matters more than the average.

What a normal abandonment rate looks like

Baymard Institute collects studies of cart abandonment and reports an average of around 70 percent. Check the current figure on its checkout usability research (opens in a new tab), because it changes and varies by store, price and traffic source. Use it as context and not as a target.

Your own numbers are the benchmark that counts. Compare mobile with your desktop, this month with last month, and each step with the one before it. The leak is the step where you lose clearly more than your own baseline.

Find friction in Clarity recordings

GA4 tells you where shoppers leave. Clarity shows what they did on that step. Microsoft Clarity (opens in a new tab) is a free tool that records anonymised sessions and builds heatmaps of where visitors click and scroll.

Filter recordings to the step that leaks

Open Recordings and filter by the device GA4 pointed at, by the URL of the leaking step (cart, shipping or payment) and by date. Then add the friction filters: rage clicks, dead clicks, quick backs and JavaScript errors. Filter names can change, so check the Clarity docs for the current list.

Watch ten to fifteen recordings and pause wherever the shopper stalls. Write the timestamp and one line on what happened. Then open the heatmap of the same page. Clicks that land on something not clickable show a button or label that looks active and is not.

On Shopify, what Clarity can record inside checkout depends on your plan and setup, so read its documentation first. Recordings of the cart and the pages before checkout still help, and your own test orders cover the rest.

Rage clicks and dead clicks, in plain words

A rage click is several quick clicks or taps in the same spot. The shopper expects something to happen and it does not. A dead click is a click that gets no response at all, often on an image, a label or a button that looks active. A quick back is a shopper returning to the previous page almost at once.

At checkout, the usual sources are a Pay button that waits for a script to load, a coupon field that clears the form, or a dropdown that will not open on mobile. We read the same metric on our own site, so we know a few such sessions are normal. One step where they cluster is the signal.

Magnifying glass over cursor footprints that stop at a closed door, showing where shoppers stall in Clarity recordings
Clarity shows where shoppers circle and stop

Run a speed test on the same step

Recordings show a stall but not what caused it. A speed test does. Run PageSpeed Insights on your cart URL with the mobile view, and use Lighthouse in Chrome DevTools for checkout pages that need a cart session. Read the three Core Web Vitals, the measures Google uses for page experience.

2.5s

LCP, how fast the main content appears

Google's good threshold, 75th percentile

200ms

INP, how fast the page answers a tap

Google's good threshold, 75th percentile

0.1

CLS, how much the layout jumps

Google's good threshold, 75th percentile

INP matters most here, because a Pay button that answers late produces exactly the dead clicks you saw. Write down the three values and the heaviest third-party scripts the report names. If you only have time to read one number, read INP on the payment step.

Checkout drop-off diagnosis: five usual causes

Most leaks trace back to one of five causes. Each leaves a recognisable sign in GA4 and in Clarity, so you can match your findings to the list without guessing.

Slow scripts delaying the checkout

The sign: a poor INP value, dead clicks on the Pay button and a mobile drop around payment. Every chat widget, review badge, tracking pixel and app adds script weight. The hidden costs of Shopify apps are a good read if your store has grown by adding them.

The fix is often smaller than a rebuild. Remove scripts the checkout does not need, load the rest later and retest. A developer does this in hours when the list of scripts is clear.

Shipping cost that appears too late

The sign: a large drop right after the checkout starts, on every device, with recordings showing shoppers pausing at the totals and going back to the cart. Baymard's research lists unexpected extra costs among the reasons shoppers give for leaving.

The fix is a message, not code: show a shipping estimate or a free-shipping threshold on the product page and in the cart. You can often write it yourself, with a small theme edit to place it.

Forced account creation

The sign: a drop at the first checkout step and recordings that stall on a login or register form. Baymard also lists forced account creation among the reasons shoppers leave.

The fix is guest checkout, with an offer to create an account after the order. Most platforms have a setting for it. It is also a good candidate for an A/B test, because some stores value accounts for repeat orders.

Payment errors at the last step

The sign: a drop between payment details and purchase, concentrated in one payment method or country, with JavaScript errors or repeated clicks on the pay button. This is a bug, so treat it as the first fix.

Test every method on every device, and read your payment provider's logs for declines and timeouts. How much of checkout you control depends on the platform, and the WooCommerce and Shopify comparison shows where the cause sits in the platform and where in your store.

Mobile form friction on small screens

The sign: a mobile-only gap at the shipping form, with recordings of shoppers typing, deleting and retyping. Usual culprits are the wrong keyboard for an email or phone field, no autofill, too many fields and tap targets that are too small.

The fix is the right input types, autofill attributes, fewer fields and an address lookup if your volume justifies one. Removing fields is your call. The theme changes are a developer's afternoon.

Symptom, fix and who does it

Use this table to sort your findings into a fix-first order. Bugs that stop payment come first, free messages second, structural work last.

SymptomFixWho does it
Mobile drop at payment, dead clicks on PayRemove or delay non-essential scripts, retest INPDeveloper
Large drop after checkout starts, all devicesShow shipping cost and a free-shipping threshold in the cartYou, plus a small theme edit
Pauses and exits on the login formTurn on guest checkoutYou, in platform settings
One payment method or country failsTest each method, read provider logs, fix the errorDeveloper with the payment provider
Mobile-only gap at the shipping formRight keyboards, autofill, fewer fieldsDeveloper (field removal: you)
GA4 purchases differ a lot from store ordersRepair the tracking eventsDeveloper or analytics specialist

Notice how many rows you can take yourself. If your list sits mostly in those rows, you may not need a developer before the campaign.

Test each fix with an A/B test

An A/B test shows two versions to real visitors and tells you which one converts better, so you keep what works. It is the right tool for some fixes and a waste of time for others.

Fix bugs, test choices

If something is broken, fix it. If you are choosing between two reasonable designs, test them.

Bug or choice

A payment button that does nothing is a bug: fix it today. A shipping message, guest checkout or the number of form fields is a choice: test it.

Testing a bug only delays the repair while shoppers keep leaving. Testing a choice protects you from shipping a change that feels right and converts worse.

Set up one clean test

Change one thing per test, so a result has one explanation. Name the measure before you start, for example completed purchases per visitor who reaches the shipping step. On stores we look after, we set the test up in werun.Experiments, so visitors are split evenly and the result is calculated for you.

As a common rule of thumb, run a test for at least two full weeks so weekdays and weekends both count. Do not stop it on day three because one version looks ahead. Early leads often fade.

If your store has too few checkout starts to settle a test before the campaign, ship the fix the evidence supports. Then watch the funnel closely afterwards. Honest advice beats a test that never reaches an answer.

Read the result

Open the same GA4 funnel for the same device and compare the step before and after the change. Compare equal stretches of days, and remember that campaign traffic behaves differently from normal traffic. Keep the winner and write down the loser, because a failed idea saves the next conversation. On stores we look after, the monthly report shows the result when the month closes.

What to hand to a developer

A developer can scope and price a fix only from evidence. The pack below turns your two hours into a brief, so the first reply is a quote and not a list of questions.

The evidence pack

Put these items in one shared folder, name each file with its date and put the single biggest leak first.

  • Funnel screenshot from GA4 with the date range visible
  • Mobile and desktop split of every step
  • Payment method and country split
  • Three Clarity recordings with the timestamp of the stall
  • Speed test report for the leaking step, with the heaviest scripts
  • Your five causes in fix-first order, with the evidence for each
  • Devices and browsers where you saw the problem
  • The success measure, such as mobile checkout-to-purchase rate

With this pack, a developer sees what you saw. The conversation moves from what might be wrong to what it takes to fix it.

Folder of evidence cards handed across a table, representing the checkout brief given to a developer
One folder gives a developer what you saw

What to ask back

Ask the same questions of anyone you brief:

  • Which cause would you fix first, and why?
  • What number proves the fix worked, and where will we read it?
  • Which devices and browsers do you test on?
  • Is the work tried on a copy of the store before it goes live?
  • What do you need from me, such as access or a staging store?
  • Is the price fixed, or hours, and what does the quote include?

If your store runs on Shopify, the Shopify team at werun.dev fixes theme and checkout issues. If it runs on WooCommerce, the WordPress team does the same for checkout on WordPress.

Where to start this week

The first step takes ten minutes: install Clarity if it is missing, so recordings start building today. Then block two hours for the analysis, ideally once Clarity has a few days of data. Book the GA4 funnel for the first hour, because it needs no waiting.

If your list ends in a payment bug or a shipping message, you may fix it before a developer is needed. If it ends in slow scripts or mobile forms, you now hold the evidence to ask for a clear quote. When the diagnosis is done, the post on who to hire before a sales campaign covers whom to ask and in what order.

If you would rather have a person look at it with you, tell us your goal. We reply in under 2 hours on business days, and we read your funnel before we quote anything.

Frequently asked questions

How do I find out why my checkout loses sales?

Build a GA4 funnel from add_to_cart to purchase, split it by device and payment method, then watch Clarity recordings of the step with the biggest drop. A speed test of that step shows whether slow scripts add to the loss.

What is a good checkout abandonment rate?

Baymard Institute reports an average cart abandonment rate of around 70 percent across studies, and it varies by store. Compare yourself first with your own previous months and with your own desktop and mobile gap.

How do I set up a checkout funnel in GA4?

Make sure add_to_cart, view_cart, begin_checkout, add_shipping_info, add_payment_info and purchase arrive in GA4. Then open Explore, create a Funnel exploration, add those events as steps and break the report down by device category.

Why do customers abandon checkout on mobile?

Common reasons are slow scripts that make the pay button unresponsive, long forms with the wrong keyboard, and payment methods that fail on small screens. Clarity recordings filtered to mobile show which one is yours.

What are rage clicks and dead clicks in Microsoft Clarity?

Rage clicks are repeated fast clicks or taps in one spot, a sign the shopper expects a response that does not come. Dead clicks are clicks that trigger nothing at all, often on elements that look clickable.

How long does a checkout abandonment analysis take?

About two hours when GA4 already receives the checkout events and Clarity has a few days of recordings. Adding missing events or waiting for recordings to build takes extra time before you start.

How do I reduce checkout abandonment before a sales campaign?

Fix payment bugs and slow scripts first, then show shipping costs earlier and offer guest checkout. Check the funnel again by device, and A/B test the choices if you have time before the launch.

Running into this on your own stack?

Tell us the platform and the problem. We'll tell you what it takes and who does it.

Start a conversation No commitment. Just a conversation.