What a technical SEO audit checks in 2026
Four areas decide whether a site can be found and read by anything, human or machine. This is what each one catches, and how to judge an audit someone hands you.
By Kuraib AliPublished Updated 6 min read
A technical SEO audit checks four things: whether search engines and AI crawlers can reach and read your pages, whether those pages load fast enough for real visitors, whether the site's structure tells Google which pages matter most, and whether structured data describes those pages accurately. Many audits stop at a checklist with forty items and no order. The order the four get fixed in matters more than the length of the list.
This post is for judging an audit, whoever runs it. What I deliver is on the technical SEO page.
The four areas, and what each one catches
Crawling and indexing
This is where many sites lose visibility before content quality is even the question.
- A robots.txt rule that blocks too much. A common cause is a staging site's blanket disallow copied into production.
- Canonical tags pointing at the wrong URL. On shops, filtered and sorted versions of one category page often all claim to be the canonical one, so Google may index a version you did not want.
- Redirect chains. Google follows a limited number of hops, and each one adds a request before the final page is reached. The guide to 301 redirects and chains covers what is lost and what is not.
- Orphaned pages. A page with no internal link pointing to it is hard for Google to find, however good it is.
- A sitemap that disagrees with the pages. A sitemap that lists URLs returning a 404 or carrying a noindex tag tells Google a page matters, then the page says the opposite. The two need to agree.
One check matters now that did not five years ago: which AI crawlers your robots.txt allows. OpenAI says sites opted out of OAI-SearchBot "will not be shown in ChatGPT search answers", while GPTBot is its training crawler. Perplexity separates PerplexityBot, its search crawler, from training in the same way. Whether to allow the search crawlers and block the training ones is a decision, and an audit should show you which one your file makes.
Core Web Vitals and speed
Google's thresholds for a good experience are Largest Contentful Paint (LCP) of 2.5 seconds or less, Interaction to Next Paint (INP) of 200 milliseconds or less and Cumulative Layout Shift (CLS) of 0.1 or less, at the 75th percentile of real visits.
Two kinds of measurement get mixed up. Field data comes from real Chrome users through the Chrome UX Report, and it is what Search Console's Core Web Vitals report shows. Lab data comes from a single simulated run in a tool such as Lighthouse. Lab data is useful for finding why a page is slow, but a page can pass a lab test and still fail in the field on real phones and networks. An audit should report both and say which is which.
The usual causes when a site fails:
- hero images that send a desktop-sized file to a phone
- render-blocking CSS or JavaScript loaded before anything visible
- web fonts that shift the layout when they finish loading
- third-party scripts (chat widgets, tags, ad pixels) that load ahead of the content
Google also says it shows the most relevant content even when the page experience is weaker, so speed rarely rescues a page that does not answer the search. It matters most where several pages answer equally well, and for the visitors who leave before a slow page appears.
Site structure and internal links
How a site is structured decides which pages gather authority and which pages compete with each other for the same search, which is called keyword cannibalisation. Two pages aimed at the same intent split the signal between them. The fix is usually not a third page. It is choosing the stronger page, moving the useful content from the other into it, and redirecting the weaker one.
What stops it coming back is a clear structure: one main page per topic, with focused pages beneath it that link back to it and to each other. This site's guides are built that way, in two groups (SEO and AI search). An audit checks whether a structure like that exists, or whether every page stands alone with one link from a blog index.
Structured data
Schema markup (Organization, Person, Article, BreadcrumbList and the types that fit the business) tells a machine what a page is, not only what it says. What matters as much as the markup is consistency: the same business name, author and key pages described the same way everywhere they appear.
Keep it in proportion. Google says structured data is not required for its AI features and there is no special schema.org markup to add for them. The FAQ rich result stopped appearing in Google Search in May 2026, so FAQ markup is no longer something to sell as a visibility fix. Neither is an llms.txt file: Google says Google Search does not use it.
How do you know an audit found something real?
Anyone can run a crawler and export every warning: a missing alt attribute, a few 302 redirects that resolve fine, a meta description a few characters over a suggested length. None of that moves a ranking on its own. A useful audit separates the few issues that hold visibility back, usually in crawling, indexing or speed, from the long tail of warnings that pad a report.
The test I use: if fixing an item would not change what Google or an answer engine can reach, read or rank, it goes to the bottom of the list. A page published three weeks ago that still appears nowhere usually has a crawling or structure problem (no internal links, a canonical pointing elsewhere, a sitemap that was never regenerated), not a missing schema field. Fixing the schema first would feel productive and change nothing.
The same holds for AI search. Google says a page must be indexed and eligible to show with a snippet to appear as a source in AI Overviews or AI Mode. A page that cannot be crawled cannot be cited, for the same reason it cannot rank. More on where the two overlap in what Google's 2026 guidance says about AEO and SEO, and on diagnosing a missing business in why your business is missing from AI answers.
Where to start
I run every audit myself, against these four areas, in the order the fixes depend on. The free SEO audit is the first step: two pages on your WhatsApp within three working days, with the three fixes I would make first.
Frequently asked questions
How long is a useful technical SEO audit?
As long as the findings that change what Google can reach, read or rank, in priority order. A forty-page export of every crawler warning is not more thorough; it hides the few items that matter.
Does a technical audit need to check AI crawlers?
Yes. Answer engines use their own search crawlers, such as OpenAI's OAI-SearchBot and PerplexityBot, separate from their training crawlers. An audit should check which ones your robots.txt allows, on purpose rather than by default.
Should structured data be fixed first?
Rarely. Markup helps Google understand a page and earns some rich results, but it cannot help a page that is not crawled or indexed. Crawling and indexing come first.