What Is a
CRO Audit?
A 5-Step Guide

The first step is understanding how users actually move through your site and where conversion happens — or fails to happen. This means mapping the key commercial pages: homepage, category pages, product pages, basket, checkout, and confirmation. At each stage you're asking the same question: where are customers dropping off, and how many?

Start with your conversion funnel in GA4. Look at add-to-basket rates, checkout initiation rates, and checkout completion rates separately. A business with a 60% basket abandonment rate has a different problem to one with a 20% checkout drop-off — even if the overall conversion rate looks the same.

Not all traffic behaves the same way — and conversion rate varies significantly by source. A visitor arriving from a branded search term is far more likely to convert than someone clicking a broad social media ad. Aggregating all traffic into a single conversion rate hides the performance differences that matter most.

Segment your traffic by acquisition channel — organic, paid search, paid social, email, direct — and compare conversion rates, bounce rates, and average session depth across each. This tells you which channels are bringing high-intent traffic and which are inflating your session numbers while contributing little to revenue. It also tells you where the biggest opportunity lies: fixing the channel with the worst quality traffic, or improving the experience for the channel with the best.

Once you understand where users are dropping off and which traffic sources are underperforming, the next step is forming hypotheses and testing solutions. This is where CRO becomes an experimental discipline rather than a design exercise.

Good CRO testing starts with a clear hypothesis: "We believe changing X will improve Y because of Z evidence." Every test should have a defined success metric and a minimum runtime to achieve statistical significance. Common high-impact test areas include product page headlines and imagery, CTA copy and placement, delivery and returns messaging, trust signals near checkout, and form field reduction on lead or registration pages.

Small changes can have meaningful impact — but only if you test properly. Changing five things at once and seeing conversion go up tells you nothing useful about what worked.

The structure and design of a page is a conversion lever in its own right — separate from messaging or traffic quality. If a site is visually confusing, hard to navigate on mobile, or buries the key action below the fold, users will abandon regardless of how good the product or price is.

Common layout issues that damage conversion include: unclear visual hierarchy that doesn't guide the eye toward the CTA; competing elements that distract from the primary action; trust signals placed too late in the journey to reduce hesitation; and mobile experiences that were adapted from desktop rather than designed for thumb use from the start.

Reviewing pages through session recordings — watching real users interact with your site — is one of the fastest ways to identify layout problems that analytics alone won't surface. When you see a user repeatedly tapping something that isn't a link, or scrolling past a CTA multiple times, that's a layout problem, not a traffic problem.

A CRO audit is a starting point, not a destination. The businesses that consistently improve conversion over time are the ones that treat it as a continuous process — a rolling cycle of hypothesis, test, measurement, and learning — rather than a project that gets done once and filed away.

After each test, document what you learned — not just whether conversion went up, but why you think it happened and what it tells you about customer behaviour. These learnings compound. A test that fails still generates insight if you've run it correctly. Over 12 months of consistent testing, a business develops a genuine understanding of what its customers respond to that no competitor can easily replicate.

Set a regular cadence: new test every two to four weeks, results reviewed monthly, priorities adjusted quarterly based on cumulative learnings. The compounding effect of consistent CRO over a year is almost always larger than any single test would suggest.