What a technical audit is (and is not)
A technical audit asks one question per page: can Google get this page, understand it and show it? It is not a content review or a backlink review. If a page is indexed, fast and canonical, a technical audit has nothing more to say about it, even if it ranks badly.
That keeps the audit short. Most sites under a few thousand pages can be audited in a day.
The tools you need
| Job | Free option | Paid option |
|---|---|---|
| Google's view of your site | Google Search Console (Domain property) | None needed |
| Crawl every URL | Screaming Frog (free up to 500 URLs) | Screaming Frog licence, Sitebulb, Ahrefs or Semrush site audit |
| Check one URL as Google sees it | URL Inspection in Search Console | None needed |
| Speed for real users | Core Web Vitals report, PageSpeed Insights | Real-user monitoring |
| Structured data | Rich Results Test | None needed |
| Raw server response | curl -I in a terminal | None needed |
Step 1: start in Search Console, not in a crawler
Search Console shows what Google actually did. A crawler only shows what it could do. Open these three places first:
- Settings > Crawl stats. Look at Host status. Any red mark for robots.txt fetch, DNS or server connectivity is a blocker. Then look at the "By response" breakdown: a jump in 5xx or 403 responses often means a firewall or CDN rule is stopping Googlebot.
- Settings > robots.txt. It shows the file Google last fetched, when, and any parse errors.
- Indexing > Pages. Write down the "Why pages aren't indexed" reasons that affect pages you care about. Ignore reasons that only hit junk URLs. Our page indexing fix guide covers each reason.
Step 2: robots.txt and sitemaps
- Read every
Disallowline. Each one should block something you can name: cart, internal search, admin. Never CSS, JavaScript or whole sections by accident. - Open each sitemap listed in Indexing > Sitemaps. Status should be "Success". Then crawl the sitemap URLs: every one should return 200, be indexable and be its own canonical. Redirects, 404s and noindex pages in a sitemap waste Google's time and blur your signals.
Step 3: crawl the site and pull six exports
Point your crawler at the homepage and let it follow links. Then export:
- Status codes. Every 4xx and 5xx URL that is linked internally.
- Redirects. Chains (more than one hop) and loops.
- Indexability. Pages with noindex in the meta tag or the
X-Robots-Tagheader. - Canonicals. Pages whose canonical points to another URL, to a redirect, or to a 404.
- Crawl depth and orphans. Important pages more than 3 clicks from the homepage, and sitemap URLs no page links to.
- Titles and H1s. Missing or duplicated across pages (a template problem, so fix it once).
Step 4: inspect your most important URLs
Pick your 10 to 20 money pages: the homepage, top categories or services, top products or posts. Run each through URL Inspection and compare:
- User-declared canonical vs Google-selected canonical. If they differ, Google picked another URL as the main one. Find out why (duplicate content, wrong internal links, a redirect).
- Page fetch and indexing allowed. Both should say yes.
- View tested page. Click "Test live URL", then open the rendered HTML and screenshot. If your main text or links are missing, you have a rendering problem. See our JavaScript SEO checklist.
Step 5: speed and mobile
Open Experience > Core Web Vitals, mobile first. Groups marked Poor come first, then Needs improvement. Google's "good" thresholds are LCP 2.5 seconds or less, INP 200 ms or less and CLS 0.1 or less at the 75th percentile (web.dev). Test one URL per template in PageSpeed Insights. Google retired the Mobile Usability report and the Mobile-Friendly Test in December 2023, so check mobile pages with Lighthouse and a real phone.
Step 6: structured data and international
- Run one URL per template through the Rich Results Test. Fix errors at the template, not page by page. Our structured data checklist lists which types still earn a result in 2026.
- If you have more than one language or country version, test the hreflang set. See the hreflang checklist.
Step 7: rank what you found
Sort every finding into three buckets:
| Bucket | What goes in it | Examples | When |
|---|---|---|---|
| Blocker | Stops crawling or indexing of pages that matter | Disallow: /, noindex on templates, 5xx, Googlebot blocked by a firewall | This week |
| Broad | Hurts a whole template or many URLs | Duplicate canonicals, redirect chains in navigation, slow product template | This month |
| Polish | One page or a small gain | A long title, one missing alt text, one orphan post | When you are next in there |
How to hand findings to a developer
Write each finding as one ticket with four parts: the URL (or template), the evidence (a screenshot or the curl -I output), the expected result in one line ("returns 404, not 200"), and how to check it is fixed. Group tickets by template. One fix to a product template beats 400 tickets for 400 products.
Worked example (fictional site). brightsmile-dental.com ran this audit after a redesign. Step 1 found nothing in Crawl stats, but the Page indexing report showed 38 service pages as "Alternate page with proper canonical tag". Step 4 explained it: the new template's canonical pointed every service page to /services/. One template fix, then a request to validate in Search Console. The rest of the audit found only polish items.
How often to audit
Run steps 1, 2 and 4 after every big release: new theme, CMS move, redesign, new CDN or security rule. Run the full audit every 3 to 6 months. Check Crawl stats, Page indexing and Core Web Vitals monthly.
How to use this with an AI assistant
Connect SEOmatic at https://seomatic.ai/mcp and ask: "Run check_site on mysite.com. Then sign me in with connect_seomatic, run site_audit, and inspect my 10 top pages with gsc_indexing (action: batch_inspect). Sort what you find into blocker, broad and polish." Reading your Search Console data is free after sign-in; any fix is staged for your approval.
More from the technical SEO checklist
- The checklistTechnical SEO Checklist (2026)
- Guiderobots.txt for AI Crawlers: Which to Allow and Which to Block
- GuideHow to Fix INP (Interaction to Next Paint)
- GuideJavaScript SEO Checklist for React, Next.js and Single-Page Apps
- GuideHreflang Checklist: Set It Up Right and Fix Common Errors
- GuideStructured Data Checklist 2026: What Still Earns a Rich Result


