Conversion Rate Optimization Audit: A Step-by-Step Guide

What a conversion rate optimization audit is

A conversion rate optimization audit, usually called a CRO audit, is a structured review of how visitors move through your site toward a conversion — a purchase, a signup, a form submission, a booked call — and where they stop. Conversion rate optimization is the practice of increasing the share of visitors who complete that action. The audit is the diagnostic step before any change is made.

A CRO audit has two halves. The quantitative half uses analytics to find where the drop-off happens: which page, which funnel step, which device, which traffic source. The qualitative half uses recordings, heatmaps and feedback to find why it happens: confusing copy, a form that asks too much, a button nobody sees, a page that loads slowly on a phone. Either half alone misleads you. Analytics without recordings tells you where people leave but not why. Recordings without analytics tell you what a few visitors did but not whether it matters at scale.

Before you start: define the conversion

An audit without a defined conversion is a wander through analytics. Pick one primary conversion before you open any tool, and write it down in one sentence: "A visitor completes the checkout and reaches the order confirmation page" or "A visitor submits the contact form and reaches the thank-you page". Everything in the audit is judged against that sentence.

Also set the scope. Audit one funnel, not the whole site. The funnel is the sequence of pages between arrival and conversion: homepage to product page to cart to checkout to confirmation, or landing page to pricing page to signup form to welcome page. If your site has several funnels, audit the one that carries the most revenue or the most traffic first.

Step 1: Check that the data is trustworthy

Open Google Analytics 4 and confirm three things before you believe any number in it. First, that the conversion event is actually firing: look at the event count for the last 30 days and confirm it roughly matches what your orders or form submissions say. Second, that the tracking tag loads on every page in the funnel, including the thank-you page. Third, that internal traffic is filtered out, or at least identified, so your own team's visits do not inflate the numbers.

This step feels like delay, and skipping it is the most common audit mistake. A funnel that shows a 90% drop at checkout may mean the checkout is broken, or it may mean the thank-you page never fires its event. Fix the measurement first, then read the funnel. The Google Analytics audit post covers this check in detail.

Step 2: Map the funnel and find the biggest leak

In GA4, build the funnel for your primary conversion over the last 90 days, and split it by device and by traffic source. You are looking for the single step with the largest absolute drop — not the largest percentage drop, but the step where the most people leave in raw numbers. A step that loses 80% of 100 visitors matters less than a step that loses 30% of 10,000 visitors.

Write down the funnel with numbers at each step. For example: 10,000 visitors reach the product page, 2,500 add to cart, 1,000 reach checkout, 400 complete the purchase. The biggest leak is product page to cart: 7,500 people leave there. That is where the audit focuses first.

Also check the bounce rate in Google Analytics for the entry pages, and compare conversion by channel. If paid traffic converts at half the rate of organic traffic, the problem may be the landing page the ads point to, not the funnel itself.

Step 3: Watch real sessions in Microsoft Clarity

Microsoft Clarity is free, with unlimited session recordings and heatmaps, no traffic limits and no paid tier. Install it, wait until you have at least 100 recordings of the funnel, then filter the recordings to the step where the leak happens and watch 20 to 30 of them. Look for rage clicks (rapid repeated clicks on something that does not respond), dead clicks (clicks on elements that are not interactive), excessive scrolling, and sessions that stop mid-form.

Step 4: List every friction point, then rank it

Go through the funnel page by page and list everything that could stop a visitor. Common friction points include: a form with more fields than the offer justifies; a shipping cost that appears only at checkout; a mobile layout where the button sits under a sticky banner; copy that describes features instead of the outcome; a page that takes more than three seconds to load on a phone; and trust signals missing at the exact moment a visitor is asked to pay.

Rank the list by two factors: how many visitors the fix would touch, and how confident you are that it is the cause. A fix to the product page touches 10,000 visitors; a fix to the confirmation page touches 400. A form field you watched ten visitors struggle with is a more confident fix than a headline you merely suspect is weak. Put the high-reach, high-confidence items at the top, and do not start with a redesign. Most audit wins come from copy, form fields and speed.

Step 5: Check the technical basics

Run these checks on the funnel pages, because no amount of copy testing fixes a broken page:

• Page speed on mobile. Test the product and checkout pages on a real phone connection, not only on office wifi. Slow pages lose visitors before any persuasion can work.
• Form errors. Submit the form with a mistake in every field and read the error messages. Vague errors ("invalid input") cost more conversions than long forms do.
• Broken elements. Check that the checkout button works in every browser your visitors use, and that no third-party script blocks the page.
• Tracking again. Confirm the conversion event still fires after any change you make during the audit.

Step 6: Test one change, then measure

Pick the top item on your ranked list and test it. An A/B test is the honest method: half your visitors see the current page, half see the changed page, and you compare conversion rates. VWO offers A/B, multivariate and split URL testing with a 30-day full-featured trial; its pricing is quote-based and metered by monthly tracked users, so ask for a figure before you commit. Contentsquare, which now includes Hotjar, offers heatmaps, replays, funnels and surveys on a Free plan, with Growth from $49 a month billed annually.

If you cannot run an A/B test, change one thing and compare the two weeks before with the two weeks after, keeping traffic sources as steady as you can. Never change five things at once and declare the audit a success — you will not know which change did the work, and you will not know what to repeat.

Decide the success rule before the test starts: the change wins if conversion rises by a stated amount with enough visitors to trust the result. A test with 50 visitors proves nothing, whatever the percentage says.

Worked example: a checkout audit on a small store

A small online store sells a $45 product. Over the last 90 days, GA4 shows 12,000 visitors reaching the product page, 3,000 adding to cart, 1,200 reaching checkout, and 480 completing a purchase. The conversion rate from product page to purchase is 4%, and the biggest leak is product page to cart: 9,000 visitors leave there.

Clarity recordings of the product page show two patterns. Visitors scroll to the reviews section and stop, and several click the shipping information text expecting a delivery estimate. The heatmap shows the "Add to cart" button gets clicks, but only from visitors who scroll past the first screen — the button sits below the fold on mobile.

The ranked fix list starts with three items: move a compact delivery estimate next to the price, move the "Add to cart" button above the fold on mobile, and shorten the product description so reviews appear sooner. The store tests the button move first, because it touches all 12,000 visitors. Two weeks later, add-to-cart rises from 25% to 31% of product page visitors. At the same checkout completion rate, that is 720 purchases instead of 480 over 90 days — 240 extra orders, or $10,800 in revenue, from moving one button. The audit then moves to the next item on the list.

CRO audit checklist

• [ ] One primary conversion defined in a single sentence.
• [ ] One funnel scoped, with GA4 numbers for the last 90 days.
• [ ] Conversion event verified against real orders or submissions.
• [ ] Funnel split by device and traffic source.
• [ ] Biggest leak identified in absolute numbers, not percentages.
• [ ] At least 20 session recordings watched at the leak step.
• [ ] Heatmaps checked for scroll depth and dead clicks.
• [ ] Friction list written and ranked by reach and confidence.
• [ ] Mobile speed, form errors and broken elements checked.
• [ ] One change tested, with the success rule written before the test.
• [ ] Result documented, including what did not work.

Common mistakes in a CRO audit

• Auditing the whole site at once. A site-wide audit produces a long list and no priorities. One funnel, one leak, one fix.
• Trusting percentages without volume. A 50% drop on 40 visitors is noise; a 10% drop on 10,000 visitors is a project.
• Reading recordings without analytics. The angriest session you watch may represent three visitors out of ten thousand.
• Testing five changes together. You learn nothing you can repeat.
• Stopping at the report. An audit that ends in a document changed nothing. The audit ends when the first fix is live and measured.

For the measurement side of this work, the best website analytics software list compares the tools that sit alongside GA4, and the marketing report metrics guide covers which numbers belong in the report you send after the audit. If you want help running the audit itself, Adziq's analytics services page explains how that engagement is scoped.

Common questions

Sources

• Google Analytics 4 (Google Analytics Help)
• Microsoft Clarity
• Contentsquare pricing (Hotjar)
• VWO

Adziq sells marketing services and does not accept payment for placement in its articles. See our editorial policy.