Technical SEO audit checklist: 10 checks that decide whether Google can read your site

A site can look perfect and still be invisible to Google. Most of what decides whether a page ranks happens below the surface: in status codes, canonical tags, render paths and crawl rules nobody sees from the front end.

This is the checklist we work through in every technical SEO audit. Each check says what to look for, how to find it with free tools, and what a fix looks like.

1. Crawling and site structure

Google finds pages by following links. A page that no other page links to, or that sits seven clicks from the home page, is crawled rarely and trusted little.

What to check: crawl the site with a crawler such as Screaming Frog (free up to 500 URLs) and look at click depth and orphan pages. Then compare the crawl with Search Console’s Pages report: anything Google knows about that your crawl didn’t reach is orphaned.

The fix: link every page worth ranking from a relevant page within three clicks of the home page, and keep filter and sort combinations from multiplying into thousands of crawlable URLs.

2. Indexing and canonicals

A page has to be in Google’s index before it can rank. Pages fall out for dull reasons: a stray noindex left over from a staging site, a canonical tag pointing at another page, or Google judging two pages to be duplicates and keeping only one.

What to check: in Search Console, open Indexing → Pages and read every reason under Why pages aren’t indexed. Use the URL Inspection tool on your most important pages: it tells you which canonical Google chose, which can differ from the one you declared.

The fix: one URL per piece of content, a self-referencing canonical on each, and noindex only where you mean it.

3. Redirects and broken links

Every redirect costs a little crawl time, and chains of them lose more. Broken internal links send visitors and crawlers to dead ends, and a deleted page that still has links from other sites wastes everything those links were worth.

What to check: your crawler’s redirect chains and 4xx reports, and the Not found (404) list in Search Console.

The fix: point every internal link straight at its final URL, cut chains to a single hop, and redirect retired pages that still earn links to the closest live page, not to the home page.

4. Core Web Vitals

Google measures three things from real Chrome users: how fast the main content loads (Largest Contentful Paint), how quickly the page responds to a tap or click (Interaction to Next Paint) and how much the layout jumps while loading (Cumulative Layout Shift).

What to check: the Core Web Vitals report in Search Console groups slow URLs by template. PageSpeed Insights shows the field data for a single page and lab data to debug it.

The fix: usually a handful of template-level changes. Size images properly and give them dimensions, load the hero image early, defer scripts the first screen doesn’t need, and reserve space for anything that loads late, such as embeds and cookie banners.

5. Mobile and JavaScript rendering

Google indexes the mobile version of your site, rendered by a smartphone crawler. If content, links or structured data only appear after a script runs, or only on desktop, Google may not see them.

What to check: run your key templates through URL Inspection and look at the rendered HTML and screenshot. Compare it with what a visitor sees on a phone.

The fix: render important content and links on the server, and make sure the mobile layout carries the same content as the desktop one.

6. Structured data

Structured data tells search engines what a page is: a product with a price, a service, an article with an author and a date, a breadcrumb trail. It makes pages eligible for richer results, and gives AI systems facts they can trust.

What to check: the Enhancements reports in Search Console, and Google’s Rich Results Test for individual pages.

The fix: JSON-LD for your organization, your products or services, your articles and breadcrumbs, with every value matching what the page actually shows.

7. Sitemaps and robots.txt

An XML sitemap is the list of pages you want indexed. robots.txt says what crawlers may fetch. Both go wrong quietly: sitemaps full of redirected and noindex URLs, or a Disallow line that blocks a whole folder by accident.

What to check: the Sitemaps report in Search Console, and your robots.txt read line by line against the folders you actually use.

The fix: a sitemap that lists only canonical, indexable pages with honest last-modified dates, and a robots.txt that blocks only what must stay out of search.

8. AI crawler access

The assistants that answer with sources run their own crawlers. OpenAI uses GPTBot for training and OAI-SearchBot for ChatGPT search; Anthropic uses ClaudeBot; Perplexity uses PerplexityBot. Google’s Google-Extended token controls whether your content trains Gemini, but it does not take you out of AI Overviews, which follow ordinary Googlebot rules.

What to check: your robots.txt and your firewall or CDN settings. Some security products block AI crawlers by default.

The fix: decide on purpose which assistants may read you, and allow them. Add an llms.txt file: a short, plain map of your most useful pages.

9. Titles, descriptions and headings

Google rewrites titles it finds unhelpful, and cuts long ones by pixel width, not characters. A missing or duplicated description leaves Google to choose a snippet for you.

What to check: your crawler’s report of missing, duplicate and over-long titles and descriptions, and one H1 per page.

The fix: a unique title for every page that leads with what the page is about, a description written as the answer to the search, and headings that outline the page.

10. Languages and regions

A site in more than one language or for more than one country needs hreflang tags so Google shows the right version to the right searcher. They must be reciprocal: if the English page points to the Bangla one, the Bangla page must point back.

What to check: your crawler’s hreflang report, looking for missing return links and tags that point at redirected or noindex pages.

The fix: a complete set of reciprocal tags on every language version, including a self-reference, plus an x-default for everyone else.

How long does a technical SEO audit take?

Running the checks takes a few days. Fixing what they find takes longer, and depends on your platform. Ours takes three weeks for a flat $1,000: a week to audit, a week to fix, and a week to check the fixes in Search Console and walk you through them. If you would rather keep the checklist and do it yourself, our free Chrome extension runs 82 technical checks on any page you open, on your own machine.