Technical SEO audits have a length problem. A 200-point checklist looks thorough and buries the four things that would have made a difference under 196 that would not.
These eighteen are ordered by impact. The first six decide whether your pages exist as far as Google is concerned. Everything after that is refinement, and refinement on an unindexed page is worthless.
Work top down and stop when you run out of problems.
Indexation: Does Google Have Your Pages At All?
1. Compare pages published against pages indexed
In Search Console, open the Pages report. Compare the indexed count against how many pages you actually have.
A large gap is the most important finding in any audit and it makes everything below it irrelevant until resolved. Twenty-two published pages and fourteen indexed means eight pages are doing nothing.
2. Read the exclusion reasons, one by one
The Pages report groups non-indexed URLs by reason. Each one means something different and the fix differs accordingly.
| Reason | What it means | Typical fix |
|---|---|---|
| Crawled, currently not indexed | Google saw it and declined | Thin or duplicate content, weak internal linking |
| Discovered, currently not indexed | Google knows it exists, has not crawled | Crawl budget, or low perceived value |
| Duplicate without user-selected canonical | Two URLs, no canonical declared | Add a self-referential canonical |
| Alternate page with proper canonical | Working as intended | Usually nothing |
| Page with redirect | URL redirects | Fine, unless it is in your sitemap |
| Not found (404) | Gone | Redirect if it has a replacement, 410 if not |
| Server error (5xx) | Server failed | Fix urgently, this stops crawling |
5xx errors deserve immediate attention. Google reduces crawl rate on a site returning server errors, which delays everything else you do.
3. Check robots.txt is not blocking what you need
A single misplaced Disallow can remove a section of the site from search. Check it directly, and check that the sitemap is referenced.
Also decide deliberately about AI crawlers. GPTBot, ClaudeBot, PerplexityBot, and Google-Extended each honour robots.txt, and a default you never chose is still a choice. See how to rank in AI search.
4. Look for stray noindex tags
<meta name="robots" content="noindex"> left over from a staging environment is a genuinely common cause of a page that will not index no matter what else you do. Check the rendered HTML, not the template.
5. Verify the sitemap resolves cleanly
Every URL in the sitemap should return 200 and be the canonical version. Sitemaps commonly contain redirected URLs, 404s, and pages marked noindex, all of which waste crawl budget and signal carelessness.
The sitemap should list what you want indexed. Nothing else.
6. Confirm one canonical host
Your site should be reachable at exactly one hostname. If www and non-www both serve content without redirecting, Google sees two copies of everything and has to guess which is authoritative.
Test all four: http://, https://, http://www, https://www. Three should redirect to the fourth, in one hop.
URL Consistency
7. Pick a trailing slash convention and hold it
/about and /about/ are different URLs. One should redirect to the other, and every internal link, sitemap entry, canonical tag, and structured data reference should use the surviving form.
This sounds pedantic and it is not. Every link pointing at the redirecting version passes signal through a redirect instead of directly, and a canonical pointing at a redirect is a canonical pointing at nothing.
Check specifically: internal links in content, canonical tags, sitemap entries, and url, @id, and breadcrumb item fields in your structured data. That last group is missed almost universally.
8. Give every page a self-referential canonical
Every indexable page should declare a canonical pointing at itself, in absolute form, on the canonical host, in the correct slash form.
Missing canonicals are the single most common finding in real audits. Pages without one are left to Google's inference, and Google infers wrong often enough to matter.
9. Keep redirect chains to one hop
/old to /new to /newest loses signal at each step and wastes crawl budget. Point the original directly at the final destination.
This matters most after a site migration, where chains accumulate silently.
10. Handle legacy URLs deliberately
Old URLs from a previous site keep getting crawled for years. Each deserves a decision:
- A real equivalent exists: 301 to it
- No equivalent and never will: 410 Gone, not a 301 to the homepage
That second point matters. Redirecting an unrelated dead URL to the homepage is a soft 404 to Google. It keeps the URL alive in the index and dilutes the target. 410 tells Google to drop it permanently.
Rendering and Speed
11. Confirm content exists without JavaScript
Fetch a page with JavaScript disabled, or use the URL Inspection tool's rendered HTML view. If your main content is absent, you are relying on Google's rendering queue, which is slower and less reliable than server-rendered HTML.
12. Pass Core Web Vitals on mobile
Field data in Search Console, not a lab score on your laptop. Full treatment in Core Web Vitals.
13. Check the mobile experience genuinely
Indexing is mobile-first, so the mobile version is the version that counts. Look for content hidden on mobile, tap targets too close together, text too small, and horizontal scrolling.
Horizontal overflow is the most common defect, and tables are the usual cause. A three-column table is already wider than a phone. Every table needs its own horizontal scroll container so it does not drag the whole page sideways.
Structured Data
14. Validate it, and check the URLs inside it
Run pages through the Rich Results Test. Beyond validity, check that every url, @id, and breadcrumb item matches the canonical exactly, including the trailing slash.
Mismatched identifiers split the entity. The canonical says one thing, the schema says another, and Google reconciles them imperfectly.
15. Only mark up what is visible
FAQ markup must correspond to text a reader can see on the page. Marking up questions that are not rendered is a guidelines violation, not a clever shortcut.
It happens by accident when schema is maintained separately from content: the FAQ section gets redesigned or removed, and the markup stays. Worth an automated check in your build if you can manage it.
The Rest
16. Titles under 70 characters, descriptions under 160
Beyond those lengths, search results truncate and you lose the end of your own message. Check every page, not a sample. If your site appends a brand suffix, count it.
17. Fix internal orphans and dead ends
A page with no internal links pointing at it is hard for Google to find and reads as unimportant. Crawl your own site and look for pages reachable only from the sitemap.
Internal linking is also the cheapest ranking lever most sites ignore. Related content linked from within body copy passes real signal, and it is entirely under your control.
18. Check HTTPS is complete
Valid certificate, no mixed content, no internal links still pointing at http://. HSTS if you can support it.
Doing This Repeatedly
An audit found and fixed once comes back. Sites regress, especially the URL and canonical issues, because a single template change can undo all of it silently.
The fix is automation. Put the checks that matter into your build so a regression fails the build rather than shipping. Canonical present, on-domain, self-referential, correctly slashed. Sitemap URLs resolving. Titles and descriptions within budget. Structured data URLs matching canonicals. Tables inside scroll containers.
That converts a recurring manual audit into a problem that cannot recur. It is a few hours of scripting and it repays itself the first time it catches something.
Where to Start
If you have never audited the site, do checks 1 through 6 today. They take an afternoon and they typically find something that explains a lot.
Then, once pages are actually indexed, the constraint is usually content rather than technical, and ranking in AI search or local SEO will do more than any further technical work.
Want an Audit That Ends in Fixes?
Our digital marketing team runs technical audits and implements the results, including the build-time checks that stop problems returning.