A practical CRO audit framework: where websites actually lose revenue, what to measure, and how to turn findings into a prioritised testing roadmap.
Most conversion problems are not mysteries. They are measurable, findable, and usually hiding in one of a handful of predictable places. A proper CRO audit is how you find them, and it is the difference between a testing programme built on evidence and one built on someone's opinion about button colours.
This is the framework we use when we audit a site for the first time. It works for ecommerce and B2B lead generation alike, and you can run a decent version of it with tools you already have.
Start With the Money, Not the Website
Before you open a single analytics report, write down how the website makes money. What counts as a conversion, what is each one worth, and which pages sit on the path to it? A B2B site might live or die on demo requests worth thousands each. An ecommerce store cares about add-to-basket rate, average order value and checkout completion.
This sounds obvious. It is also the step most audits skip, which is how teams end up celebrating a lift in newsletter signups while revenue stays flat. Every finding in the rest of the audit gets weighed against this commercial map.
Step 1: Check the Measurement Before You Trust It
An audit built on broken tracking finds broken conclusions. Before analysing anything, verify that your key events actually fire, that conversions are not double-counted, and that form submissions, purchases and calls are all captured. In our experience roughly half of first-time audits surface a measurement problem significant enough to change the conclusions.
If the foundations are shaky, fix them first. Our tracking and measurement work exists precisely because this step fails so often.
Step 2: Find Where the Funnel Actually Leaks
Build a simple funnel from landing to conversion and look at the drop-off between each step. You are looking for the step that underperforms its benchmark, not the step with the biggest absolute drop. Every funnel narrows. The question is whether yours narrows faster than it should at a specific point.
- Segment before you judge. A blended conversion rate hides everything useful. Split by device, traffic source and new versus returning visitors. A site that converts well on desktop and terribly on mobile does not have a conversion problem, it has a mobile problem.
- Compare high-intent pages against each other. If one service page converts at three times the rate of another with similar traffic quality, the weaker page is your cheapest experiment.
- Watch the exits. Pages with high exit rates in the middle of the funnel are telling you where confidence collapses.
Step 3: Watch Real People Struggle
Analytics tells you where people leave. Session recordings and heatmaps tell you why. Watch twenty recordings of visitors on your weakest funnel step and patterns emerge quickly: forms that punish small mistakes, delivery costs revealed too late, buttons that look clickable but are not, mobile layouts that bury the one thing people came for.
Pair this with the handful of questions that matter: what almost stopped you buying? What did you need to know that the page did not tell you? A one-question survey on a thank-you page routinely outperforms weeks of internal debate.
Step 4: Audit the Persuasion, Not Just the Plumbing
Once the mechanical issues are logged, read the page the way a sceptical buyer would. Does the headline say what you do and who it is for? Is the proof specific, with numbers and names, or is it a wall of adjectives? Is the next step obvious, and is the cost of taking it clear? Most conversion problems that survive the technical fixes are trust problems wearing a design costume.
Step 5: Turn Findings Into a Prioritised Roadmap
A good audit ends with a ranked list, not a document nobody reads. For each finding, estimate three things: the revenue impact if you are right, your confidence that you are right, and the effort to test or fix it. Rank by impact and confidence divided by effort.
Two rules make the roadmap honest. First, some findings do not need a test. A broken form or an error message on checkout is a fix, not a hypothesis. Ship it. Second, respect your traffic reality. Meaningful A/B tests need enough conversions to reach significance. Below several hundred conversions a month on the tested page, weight the programme towards research and audit fixes instead of split tests.
What a CRO Audit Should Deliver
Done properly, you finish with a verified measurement setup, a map of where revenue leaks and why, evidence from real user behaviour rather than opinion, and a testing roadmap ranked by commercial impact. That roadmap is where conversion work stops being a cost and starts compounding: each win lowers your acquisition cost across every channel you run.
This is the audit we run at the start of every conversion rate optimisation engagement, connected to the client's revenue model from day one. If you would rather see what it finds on your site than run it yourself, speak to us. You can also explore how we approach data and analytics as the foundation for decisions like these.
