GuidesAudits

Technical SEO audit checklist: 30 checks, in the order to do them

A technical SEO audit checklist you can work through in an afternoon: crawlability, indexing, page basics, links, speed, security and structured data, with the limits that matter.

6 min readUpdated 2026-10-31By the Auditmatic team
On this page

A technical SEO audit finds the problems that stop search engines from finding, reading and trusting your pages. This checklist goes in the order that matters: if a page cannot be crawled, nothing else on it counts. You can do it by hand in an afternoon, or run the audit below and get the same list with the affected pages named.

1. Can search engines reach your pages?

  • robots.txt exists, is not blocking pages you want found, and points to your sitemap.
  • XML sitemap lists only pages that return 200 and are indexable. Submit it in Google Search Console and Bing Webmaster Tools.
  • Status codes: important pages return 200. A removed page returns 404 or 410, not a 200 that says "not found".
  • Redirects go straight to the final URL. A chain of three hops wastes crawl time and dilutes signals; fix chains to one hop.
  • HTTPS everywhere, with the http and www variants redirecting to one canonical address.

2. Are the right pages indexable?

  • No stray noindex on pages you want in search results. This is the most common launch-day mistake: a staging setting that reaches production.
  • Every page has one canonical URL, and it points to itself unless you mean otherwise.
  • Duplicate content is consolidated: the same page reachable at several URLs should redirect or canonicalise to one.
  • Thin pages with almost no text are expanded, merged or set to noindex.

3. Page basics

  • A unique title tag on every page, around 50 to 60 characters, that names the page's main topic.
  • A unique meta description, around 120 to 155 characters, so it is not cut off in results. It does not affect ranking, but it affects whether people click. How to measure the width.
  • One h1 per page, and headings in a sensible order.
  • The viewport meta tag is set, and text is readable on a phone without zooming.
  • Images have alt text that describes them, and are sized for the space they fill.
  • No broken internal links. A broken link on your own site is the easiest fix on this list.
  • Important pages are linked from other pages on your site. A page with no internal links is hard for crawlers to find. If a report says a page has no outgoing links, see when that matters and when to ignore it.
  • Link text says where the link goes. "Read the pricing" beats "click here".

5. Speed and experience

  • Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, and Cumulative Layout Shift under 0.1, measured on real devices.
  • Render-blocking CSS and JavaScript are kept small, and fonts use font-display: swap.
  • Responses are cacheable, and the server answers quickly. A slow first byte limits everything after it.

6. Security and trust

  • Security headers are set: Strict-Transport-Security, X-Content-Type-Options, a Referrer-Policy and a frame-ancestors policy.
  • Your contact details, privacy policy and terms are easy to find.

7. Structured data and sharing

  • JSON-LD matches the page: Organization and WebSite on the homepage, Article on posts, Product on products, FAQPage where you really have questions and answers. Fill the recommended fields, not only the required ones.
  • Open Graph and Twitter tags give every page a proper title, description and image when it is shared.

8. For AI answers

  • Say plainly what you do, for whom and what it costs, near the top of the page. AI engines quote pages that answer directly.
  • Offer an llms.txt file that points to your most useful pages. It is a young convention, but it costs minutes.

How to prioritise the 30 checks

Thirty checks sounds like a lot, but they do not carry equal weight. Sort every finding by two questions: does it stop a page from being found, and how many pages does it touch?

PriorityExamplesWhy
Fix todayStray noindex, robots.txt blocking pages, 5xx errors, broken redirectsPages cannot appear in results at all.
Fix this weekDuplicate URLs (www and http), missing titles, redirect chains, mixed contentThey waste crawling and split signals, and a template fix covers many pages.
Fix this monthShort or long descriptions, missing alt text, missing security headers, slow imagesReal but small gains, mostly for clicks and trust.
Ignore for nowUnused CSS or JavaScript notices, minor structured data suggestionsThe effort is high and the effect is hard to see.
ErrorPage is blocked by noindex3 pages
WarningMeta description too long12 pages
NoticeUnused JavaScript on the homepage1 page
Illustration of a findings list, not live data. The error is a blocker and goes first; the notice can wait.

The tools you need

You can run most of this audit with free tools. Google Search Console shows what Google has indexed and why a page was left out, and Bing Webmaster Tools does the same for Bing. curl shows the raw response a crawler receives, including status codes, redirects and headers. Lighthouse, built into Chrome, measures speed and accessibility. A site audit tool ties these together by crawling every page and listing the findings in one place, which is the part that is slow to do by hand.

How often to run it

Run the full checklist after any launch, migration or redesign, then audit on a schedule. Monthly is enough for most small sites. The next deploy can undo a fix without anyone noticing, so a scheduled audit pays for itself the first time it catches a stray noindex. See the SEO audit checklist for how to rank what you find, and the guides on pages with no outgoing links, long meta descriptions and long title tags for three findings you will meet often.

What we found when we audited our own site

We ran this audit on auditmatic.app the week we launched. It scored 97, and still found real work: a meta description that would be cut off in results (934 pixels against a limit of about 920), a structured data block missing one recommended field, four pages with descriptions that were too short, and a contact page with too little text to index. A good score does not mean there is nothing to fix. The list is where you start.

Fix it with your AI agent

If you build with Claude Code or another agent, it can run this audit for you and apply the fixes. See how to run an SEO audit with Claude Code, or the step-by-step strategy. The audit tool behind this checklist is described on the site audit page.

Frequently asked questions

How long does a technical SEO audit take?

By hand, an afternoon for a small site. With a crawler, the scan takes minutes and the time goes into fixing what it finds.

Do I need to be a developer?

Not to find the problems. Many fixes are settings in your CMS. For template changes you may need a developer, or an AI coding agent working from the audit's fix prompt.

What is a good audit score?

The score is only a summary. A site at 99 can still have a blocker if it affects one important page, so read the errors before the number.

Is technical SEO more important than content?

Technical problems decide whether your content can be found at all, so they come first. After that, content decides what you rank for.

Run it on your own site

Enter your address and you get a score, every finding in plain English with the pages affected, and a prompt you can paste into an AI coding agent to fix them.