Technical SEO checklist

Most technical SEO checklists are a flat list of forty items with no order, which is unhelpful — because if a page cannot be crawled, nothing further down matters yet.

These are grouped in the order to work through them, and each group names the free tool that answers it.

The checklist

Tick these here — progress saves in your browser, so a two-session audit does not lose the first session — or take the PDF if you would rather print it.

7 groups · 33 checks

Crawling and indexing 7

• robots.txt does not block anything you need crawled — check it, do not assume
• The pages you care about return 200, not 3xx or a soft 404
• Search Console Pages report shows indexed counts close to submitted counts
• No page you want indexed carries a noindex tag left over from staging
• XML sitemap lists only canonical, indexable, 200-status URLs
• Sitemap is submitted in Search Console and reports no errors
• No orphan pages: everything important is linked from somewhere crawlable

Canonicals and duplicates 6

• One hostname only — pick www or apex and 301 the other
• HTTP redirects to HTTPS in a single hop
• Every page has a self-referencing canonical
• Trailing slashes are consistent, so /page and /page/ are not two URLs
• Parameterised and filtered URLs do not create indexable duplicates
• Uppercase and lowercase paths do not both resolve

Redirects and migrations 4

• Every legacy URL redirects in ONE hop to a genuinely equivalent page
• No redirect chains, and no redirect loops
• Nothing is swept to the homepage — Google largely discards an irrelevant redirect
• Internal links point at the final URL, not through a redirect

On-page basics 5

• Every page has a unique title under roughly 60 characters
• Every page has a unique meta description
• Exactly one h1 per page, and headings do not skip levels
• Images have alt text that describes the image, not the keyword
• Internal links use descriptive anchor text rather than "click here"

Speed and Core Web Vitals 4

• LCP under 2.5s on mobile, measured on field data rather than lab data alone
• CLS under 0.1 — every image and embed has explicit width and height
• INP under 200ms; check what your third-party scripts cost you
• Fonts are self-hosted or preloaded, and no more than you need

Structured data and rendering 3

• Structured data describes what is visible on the page, and validates
• Content is present in the HTML rather than requiring JavaScript to appear
• Mobile and desktop serve the same content

Measurement 4

• Search Console verified for the exact property you are reporting on
• Analytics fires exactly once per page — check, this is wrong more often than not
• Conversions from non-paid search are tracked and tested end to end
• Bing Webmaster Tools verified as well. It takes ten minutes and it is free

The three that cost the most when skipped

Out of everything above, three problems do most of the damage, and none of them looks broken from the front end. That is precisely why they survive for months.

One hostname, one hop. Pick www or the apex domain and 301 the other; redirect HTTP to HTTPS in a single hop. Duplicate hostnames split every signal you have — links, rankings, analytics — and the site looks completely fine while it happens.

Legacy URLs redirected to a genuinely equivalent page. Sweeping them all to the homepage is the most common migration mistake and the most expensive. Google largely discards a redirect to an irrelevant page, so years of accumulated links are thrown away in an afternoon, and there is no error message.

Analytics firing exactly once. A double-fired pageview or conversion makes every downstream decision wrong — budget, channel, targets — and it is more common than any other item here. Check it in preview mode before publishing, every time.

Which tool answers which check

What you need to know	Use	Note
Indexing and coverage	Google Search Console — Pages report	Your own data. No tool estimate beats it.
Crawl the whole site	Screaming Frog or Sitebulb	Free up to 500 URLs in Screaming Frog.
Core Web Vitals, field data	Search Console — Core Web Vitals	Real users. Lab tools are for diagnosing, not judging.
Core Web Vitals, lab data	Lighthouse or PageSpeed Insights	Use it to find the cause once the field data shows a problem.
Structured data	Rich Results Test and the Schema validator	Validates syntax, not honesty. It must match visible content.
Redirects and status codes	curl -I, or the browser network tab	Check the hop count, not just the destination.

Two distinctions worth being clear about. Field data and lab data are not the same thing. Search Console’s Core Web Vitals report is real users on real connections and is what Google actually uses; Lighthouse is a single simulated load, useful for diagnosing a cause once the field data has shown you a problem. Judging a site on Lighthouse alone will send you chasing numbers no user experiences.

And the structured data validator checks syntax, not honesty. Markup can validate perfectly while describing content that is not on the page, which is a manual-action risk rather than a technical one.

Where this fits

This is the technical section of the wider digital marketing audit in depth. If you are auditing a whole setup rather than the site specifically, start there — it scores eight weighted sections and tells you whether technical work is even your biggest gap. Doing technical SEO first because it is the most tractable is a common and expensive instinct.

Once the checks pass, technical health becomes one line in a monthly SEO report rather than a project.