What a CRO audit should actually contain
Learn what a comprehensive CRO audit should contain, from UX and landing pages to analytics, user behavior, conversion barriers, and actionable recommendations.
A conversion audit that is worth paying for produces a ranked list of specific, testable problems with an estimated value attached to each. Not a list of best practices. If the document you receive could have been written without ever opening your analytics, you bought a template.
Layer one: the numbers
Conversion rate by device, by template, by traffic source and by new against returning. Then the funnel drop-offs between them.
Device is usually where the first real finding sits. Across 486 million sessions and 117 brands, DRIP found desktop converting 1.56 times higher than mobile while mobile carried 78% of traffic. Most of the audience is on the worse experience, and most audits mention this in a sentence rather than treating it as the headline it usually is.
Layer two: behaviour
Session recordings on the top five templates, heatmaps, and a documented walkthrough of the checkout on a real mid-range phone rather than a desktop emulator.
Heatmaps are useful and easy to misread, which is why the audit should say what it did with them rather than just include the pictures. At a personal care brand we worked with, operating under strict brand guidelines that ruled out most layout changes, heatmap data drove a navigation redesign. That is what the deliverable looks like: a specific change, justified by a specific observation, inside the constraints that actually exist.
Layer three: the checkout
This layer should be checked against documented causes rather than opinion. Baymard’s meta-analysis of fifty studies puts average cart abandonment at 70.22%, with 48% of abandonments caused by unexpected extra costs at checkout, 19% by mandatory account creation and 18% by a checkout that is too long.
Three known causes covering most of the recoverable loss. An audit that does not explicitly test your checkout against them is not finished.
Layer four: technical performance
Not a screenshot of a speed score. A measurement on real devices, tied to revenue.
The correlational evidence is consistent. Google and Deloitte’s analysis of over 30 million mobile sessions across 37 brands associated a 0.1 second improvement in mobile load time with an 8.4% rise in retail conversion and a 9.2% rise in average order value. Meanwhile only around 42% of mobile sites pass all three Core Web Vitals. Odds are your site is in the failing majority, and odds are nobody has priced what that costs.
Layer five: the offer
Most audits stop at the interface and never examine what is being sold. Bundle structure, shipping thresholds, subscription framing and post-purchase offers change order composition rather than order count, which is where the largest effects usually sit. An audit that reviews five templates and never mentions pricing or threshold logic has skipped the most valuable layer in the document.
What the deliverable should look like
|
Section |
What it must contain |
|
Findings |
Each one specific, evidenced, with the data source named |
|
Prioritisation |
Ranked by expected value, not by ease |
|
Test briefs |
The top five findings written as hypotheses ready to build |
|
Constraints |
What cannot be changed, and what that rules out |
|
Measurement |
Which metric each test is judged on before it launches |
The three failure modes
The best-practice list. Generic recommendations that apply to every store. If your competitor could use the same document unchanged, it has no value.
The redesign in disguise. An audit that concludes everything needs rebuilding is usually a proposal for a bigger project, not a diagnosis.
The unranked pile. Forty findings with no ordering. Prioritisation is the hard part and it is the part you are actually paying for.
What to ask for before you commission one
Ask what data access is needed and refuse to proceed until it is granted, because an audit without analytics access is guesswork. Ask how many findings will come with test briefs attached. Ask what happens if the audit concludes your funnel is basically fine, and listen carefully to the answer.
Ask for one anonymised example of a previous audit. Not the case study. The document.
Where to start if you cannot commission one yet
Do the free version. Open analytics, split conversion by device, then walk your own checkout on a phone with mobile data and a full cart. Most brands find at least one thing in twenty minutes that has been costing money for a year.
The full sequence we run, audit through research to roadmap, is set out at parahgroup.com.
What's Your Reaction?
