What Does a Technical SEO Audit Actually Check?

That last point is where most audits fail. A commercial crawler pointed at a mid sized site will return several hundred flagged items. Almost none of them explain why the page that generates revenue is sitting on page three. The value of an audit is not the finding, it is the triage.

Why Does the Order of the Layers Matter?

Because a finding in one layer can make an entire layer below it irrelevant.

If a section of your site is blocked in robots.txt, there is no point analysing its page speed. If a category page is canonicalised to another URL, its schema markup will never produce a rich result. If two of your own pages are competing for the same query, improving both of them makes the problem worse rather than better.

Working the layers in order stops you spending effort on symptoms while the cause sits untouched.

Layer One: Can Search Engines Reach Your Pages?

Crawlability is the foundation. If a page cannot be reached, nothing else about it matters.

• Robots.txt directives. What is disallowed, and whether anything commercially important is caught by a broad rule.
• XML sitemap accuracy. Whether it lists the pages you actually want indexed, and nothing that returns a redirect or a 404.
• Orphan pages. Pages with no internal links pointing at them. Common after migrations and category restructures.
• Redirect chains and loops. Every hop dilutes signal and slows crawling.
• Server response codes at scale. Soft 404s are the ones that hide, because they return a 200 while showing an empty page.
• Log file review. Where access allows, this shows what search engines are actually crawling rather than what you assume they crawl.

Layer Two: Is the Right Set of Pages Indexed?

Indexation is where the gap between intention and reality usually shows up. The question is simple: does what Google has indexed match what should be indexed?

• Canonical logic. Self referencing canonicals where expected, and no page quietly canonicalised to a different URL.
• Noindex placement. Frequently left on after a staging environment goes live. It is worth checking on every audit.
• Parameter handling. Sorting, filtering and tracking parameters generating near infinite URL variants.
• Duplicate and near duplicate clusters. Especially location pages built by find and replace, which Google treats as one page.
• Index bloat. Thin, filtered or paginated URLs consuming crawl capacity that should go to revenue pages.

Layer Three: Does the Architecture Support the Pages That Matter?

This is the layer that produces the largest gains and the one most audits skip, because it requires judgement rather than a tool.

• Internal link graph. Which pages receive the most internal links, and whether those are the pages that generate revenue.
• Click depth. How many clicks from the homepage to your commercially important pages. Four or more is a problem.
• Anchor text distribution. Whether internal anchors describe the target page or say "read more".
• Keyword to URL mapping. One page per intended query. This is where keyword cannibalization gets diagnosed, and it is usually the single most useful artefact of the whole audit.
• Hub and cluster structure. Whether supporting content actually links up to the commercial page it is meant to support.

Layer Four: Does the Site Perform for Real Users?

Performance matters, but it is routinely overstated. A page that loads slowly and answers the query will usually outrank a fast page that does not.

• Core Web Vitals from field data. Real user measurements, not a lab score from a single test run. Covered in full on which Core Web Vitals actually affect rankings.
• Render blocking resources. Scripts and stylesheets delaying first paint.
• Image delivery. Format, compression, dimensions and lazy loading.
• Third party script cost. Usually chat widgets, tag managers and analytics. Often the largest single drag.
• Mobile rendering parity. Whether the mobile version contains the same content and links as desktop. Indexing is mobile first, so anything missing on mobile is effectively missing.

Layer Five: Can Search Engines Understand What the Page Is?

Structured data does not directly improve rankings. It changes how your result is displayed, which changes how often it is clicked.

• Coverage. Which templates carry schema and which do not.
• Validity. Errors and warnings against current specifications.
• Eligibility gaps. Rich results your competitors are winning that you are not marked up for.
• Consistency with visible content. Marked up content must appear on the page. Mismatches risk a manual action.

Layer Six: What Is the Search Data Telling You?

The final layer connects the technical findings to commercial reality using Search Console.

• Striking distance pages. Positions four to fifteen with real impressions and weak click through. Rewriting those titles is the cheapest gain available on most sites.
• Query to page gaps. Queries you receive impressions for with no page built to answer them.
• Matched period comparison. Always 28 days against the preceding 28 days. Comparing a 30 day month to a 31 day month produces conclusions that are arithmetic rather than real.
• Segmentation. Commercial pages against informational pages. A falling site wide click through rate is often a change in traffic mix rather than a decline in performance.

How Do You Know Which Findings Actually Matter?

Score each one on two axes: estimated traffic impact, and implementation effort. Then fix in that order. A canonical error on a category cluster is high impact and low effort, so it goes first. Missing alt text on two hundred decorative images is low impact and high effort, so it may never be worth doing at all.

A practical test: if a finding cannot be connected to a page that makes money or a query that has impressions, it belongs in an appendix rather than the fix list.

This is the checklist we work through on every engagement, and the same one behind the free initial audit on our technical SEO audit page.