A real CRO audit report: we ran ours on our own homepage
Most audit tools ask you to pay before you can see what you get. This is a real report, run on optimizemywebsite.ai itself. It scored 68 out of 100 and flagged our own headline as the highest-impact problem on the page. Nothing below has been tidied up.
The headline score
The report opens with one number out of 100, a count of what is working, and a count of what is not. Ours came back at 68 — labelled Room for Improvement — with six dimensions scoring above 70% and four flagged as opportunities.
A single number is the least useful part of any audit taken alone. It matters as a baseline to move, not as a grade. The value is in what sits underneath it.
| Dimension | Score | Priority |
|---|---|---|
| Value to the customer | 55 | High |
| Messaging and proposition | 60 | High |
| CTA placement | 65 | High |
| Element design | 65 | High |
| CTA wording | 70 | High |
| Design standards | 70 | Standard |
| Hierarchy and layout | 70 | Standard |
| User experience | 75 | High |
| Colours | 75 | High |
| Social proofing | 80 | Standard |
Why the priority flag matters more than the score
Note that Colours scored 75 and carries a high-priority flag, while Design standards scored 70 and does not. The flag is not a restatement of the number — it marks dimensions where a weak result reliably costs conversions on this page type.
Read the flagged rows first, in score order. In our report that means value to the customer at 55, then messaging at 60, then CTA placement and element design at 65. Working the lowest number regardless of flag would have sent us to the same place by luck rather than by reasoning, and on another page it would not.
What a single finding looks like
Each dimension expands into an explanation of what is being measured, an analysis of what the page actually does, and recommendations with a concrete rewrite. This is the report's verdict on our lowest-scoring dimension, unedited:
Analysis: "The value proposition is present but lacks clarity on specific benefits. Users may not fully understand what they gain from the service, which can hinder conversion."
Recommendation, marked High Impact: enhance the clarity of the value proposition by explicitly stating the benefits of using the service. Before — "Find what’s costing you sales." After — "Discover hidden revenue opportunities and boost your sales by optimizing your website." Affected section: Header and Hero.
That is the level findings arrive at. Not "improve your messaging", but the sentence currently on the page, a replacement, and where to make the change.
The three priority actions it gave us
Alongside the per-dimension detail, the report collapses everything into a short list of what to do first. Ours was three items:
- 1. Unclear value proposition and messaging — High Impact. Revise to clearly articulate the benefits and outcomes of the service. Sections: Header and Hero, How It Works.
- 2. Primary CTA visibility and effectiveness — High Impact. Enhance the visibility and wording of the primary CTA. Section: Header and Hero.
- 3. Inconsistent design elements — Medium Impact. Standardise font sizes and button styles for a cohesive experience. Sections: Header and Hero, Service Details.
Where we disagreed with our own report
Two of its suggestions we are not acting on, and saying so is more useful than pretending an audit is always right. It proposed replacing a section heading with a claim about a specific percentage increase in client sales, and adding client logos to the testimonial block.
Both would improve the page as scored. Both would also mean publishing figures and endorsements we cannot substantiate, which is a line worth holding regardless of what it costs in points. An audit optimises for the rubric; you are responsible for whether the change is true.
This is the practical version of the caveat below: findings are hypotheses, and some of them should be rejected.
What the report does not tell you
This is the section most audit tools leave out, and it is the one worth reading before you rely on the output.
- Whether a change will work. The report reads a page; it cannot run a controlled comparison. Every finding is a hypothesis with an unknown effect size until you test it.
- Where visitors actually drop out. That lives in your analytics, across sessions the report never sees.
- Which fix to do first in commercial terms. It does not know your traffic mix, so weight its findings by the pages that carry your volume.
- Anything behind a login. Onboarding, dashboards, upgrade prompts and post-purchase flows are invisible to an automated pass.
- Why a visitor hesitated. Motivation is not observable in markup — that is a user-research question.
How to read one in ten minutes
Start with the priority-flagged dimensions in score order, then cross-reference each against the traffic the page receives, so a moderate problem on a high-traffic page outranks a severe one on a page nobody reaches.
Pick one change, ship it, and re-run the same audit. Score movement against an identical rubric is the only part of this that tells you whether you were right — which is why a repeatable rubric matters more than any individual number inside it.
Internal links
Related audit paths
FAQ
Questions teams ask before they audit
Is this a real report?
Yes. It was run on optimizemywebsite.ai on 15 August 2026 and scored 68 out of 100. The scores, analysis text and before/after examples are unedited product output, including the findings that criticise our own homepage.
What do the scores actually mean?
Each dimension is scored out of 100 against the same criteria on every page. The absolute number matters less than two comparisons: the same page over time, and one page against another. A single score in isolation is the least useful way to read the report.
Why publish an audit that criticises your own site?
Because a sample report where everything scores well would tell you nothing about the tool. Ours flagged our headline as the highest-impact problem on the page, which is a more honest demonstration of what the output looks like than a curated example would be.
Can I see a report for my own site?
Yes — a free scan produces the same structure shown here for a page you choose. Running it is the fastest way to judge whether the output format is useful before committing to anything.
How is this different from a Lighthouse or SEO report?
A Lighthouse report measures performance and technical health. An SEO report measures whether a page can be found. This measures whether the people who already arrived are likely to act, which is a different question needing different criteria.
Should I act on every finding?
No. Weight findings by the traffic a page receives and by whether the dimension carries a priority flag — and reject any recommendation that would have you publish something you cannot stand behind. We rejected two of ours for exactly that reason.