What an SEO Audit Should Actually Find
A 90-page PDF of automated warnings is not an audit. A real one tells you what is wrong, what it is costing you, what to fix first, and what to ignore.
Most SEO audits are a crawler report with a cover page. The tool finds four thousand issues, most of them trivial, ranks them by its own severity scale, and the document gets delivered without anyone deciding what actually matters for your business.
You can tell because the recommendations would be identical for any site: add alt text, fix meta descriptions, reduce image sizes. All true. None of it explains why you are not ranking.
A real audit answers four questions: what is wrong, what is it costing you, what should be fixed first, and what should be ignored.
What it should cover
Crawling and indexing
The baseline. Can search engines reach your pages, and are they being stored?
This should reconcile three numbers: how many pages you have, how many are in your sitemap, and how many are indexed. Gaps between them are where problems hide. The audit should explain each gap - not just report that it exists.
It should also identify pages that are indexed but should not be: staging URLs, duplicates, parameter variants, thin pages diluting the site.
Rendering
Whether content is present in the HTML or assembled by JavaScript. If it is the latter, the audit should say what depends on rendering and what the risk is.
Site architecture
This is where the leverage usually is, and where automated tools have least to say.
The audit should map how pages relate: click depth for commercially important pages, orphan pages, whether internal linking reflects what matters, and whether the structure supports a coherent topical model. See how website architecture affects SEO.
Intent mapping
For each important page: what query should it rank for, what does it currently rank for, and does its format match what search engines are rewarding for that term?
This is the check that explains most “good content that does not rank” situations. A tool cannot do it because it requires judgement about the searcher. Covered in search intent explained.
Content quality and duplication
Which pages are too thin to rank, which are competing with each other, and which are close to working and need a push rather than a rewrite.
Consolidation recommendations should be specific: this page absorbs that one, this URL redirects there.
Technical hygiene
Canonicals, redirects, status codes, sitemap accuracy, robots directives, structured data validity, heading structure, metadata. The standard checklist - but filtered to what is actually causing a problem, not every warning the crawler emitted. The recurring ones are listed in common technical SEO problems.
Performance
Core Web Vitals measured per template, using field data where available, not a single homepage test.
Competitive context
What the sites currently ranking are doing that you are not. This is what turns an audit from a list of defects into a plan - because the standard is not perfection, it is being better than what currently ranks.
What the deliverable should look like
The format matters as much as the coverage.
Prioritised, not exhaustive. Ordered by expected impact against effort. A good audit tells you which three things to do first and which forty to ignore.
Specific. Not “improve internal linking” but “these eleven pages have no internal links pointing at them; here is where each should be linked from”.
Costed, at least roughly. You cannot prioritise without knowing what each item takes. An audit that recommends a full replatform without saying so is not being straight with you.
Explicit about expected effect. Some fixes are near-certain wins. Some are hygiene with no direct ranking effect. Some are speculative. A good audit distinguishes between them rather than presenting everything as urgent.
Honest about what it cannot fix. If the real problem is that your competitors have far more authority and you are not going to outrank them on the head term this year, the audit should say so and redirect you toward terms you can win.
Warning signs
It is mostly a tool export. If the structure matches a crawler’s default report categories, nobody did any thinking.
Everything is critical. Severity scales that mark four hundred issues as high priority are not prioritising.
No business context. An audit that never mentions what you sell or who buys it cannot tell you which pages matter.
Guaranteed outcomes. Nobody controls ranking systems. Guarantees are a signal about the seller, not the work.
No mention of intent. If the word does not appear, the audit is technical-only - useful, but incomplete.
Recommendations you could have got from a blog post. Generic advice means nobody looked at your specific situation.
What happens afterwards matters more
An audit is a diagnosis, not a treatment. The value is entirely in what gets implemented.
The most common failure is not a bad audit - it is a reasonable audit that lands in an inbox and never gets actioned, usually because nobody owns it or the recommendations were too vague to turn into tasks.
Before commissioning one, agree who will implement the findings. If it is your team, the deliverable needs to be specific enough to work from. If it is the agency doing the audit, ask for the implementation estimate alongside it.
When you do not need one
Worth saying plainly. If your site is six months old with fifteen pages and no traffic, you do not need an audit. You need more content targeting terms you can realistically rank for, and probably some links.
Audits earn their cost on established sites - where there is existing traffic to protect, accumulated technical debt, enough pages for structure to matter, and a gap between the effort invested and the results achieved.
If you are not sure which situation you are in, that is a shorter and cheaper conversation than a full audit.
Our audits end in a prioritised list with expected effect and rough cost per item - not a PDF of warnings. Send us your URL and we will tell you whether an audit is what you actually need. More on our approach to SEO.