Technical SEO Audit: A 10-Step Checklist for 2026

A technical SEO audit finds the structural problems preventing your pages from being discovered, crawled, and ranked — regardless of how good the content is. This checklist covers the 10 areas that actually move the needle.

What a technical SEO audit checks (and what it doesn't)

Technical SEO is about infrastructure: can search engines find, access, and understand your pages? It is distinct from on-page SEO (content quality, keyword targeting) and off-page SEO (backlinks, authority).

A technical audit doesn't tell you whether your content is good. It tells you whether your content has a chance to be found.

The 10 areas below are ordered from foundational (if broken, everything else is moot) to refinement (gains are real but compounding on an already-functional base).

Step 1: Crawlability — can Googlebot reach your pages?

Before Google can index a page, it must be able to fetch it. Common crawlability failures:

• robots.txt Disallow rules blocking important pages — especially common after CMS migrations where staging-environment rules get deployed to production
• JavaScript-only navigation — <button onclick="..."> links that crawlers can't follow; key internal links that only exist after JS runs
• HTTP auth on the wrong environment — a dev subdomain that was accidentally published under the main domain
• Bot-blocking at the CDN or WAF level — Cloudflare challenge pages, aggressive rate limiting, IP blocks that hit Googlebot

Run the audit on your live domain, not localhost. Verify that all key sections are reachable from the homepage via crawlable <a href="..."> links.

Step 2: Indexability — are the right pages indexed?

Crawled ≠ indexed. Google crawls many pages but only indexes a subset. Check:

• noindex tags in the wrong place — common post-launch mistake when a staging noindex survives deployment
• Pages in the sitemap that have noindex — Google flags these as a signal of sloppy configuration
• Thin content pages being indexed — tag archives, search result pages, near-duplicate parameter URLs
• Canonical pointing away from the page — pages where rel=canonical points to a different URL won't be indexed at their own URL

Check Google Search Console → Index Coverage for the "Excluded" list. Any page there with the reason "Excluded by noindex tag" that you want indexed is an immediate fix.

Step 3: robots.txt — what you're blocking

A correct robots.txt is mostly empty — it blocks only what it intends to block.

Common mistakes:

# BAD: blocks everything, then re-allows (Google reads first Disallow)
User-agent: *
Disallow: /
Allow: /blog/

# GOOD: block only what should be blocked
User-agent: *
Disallow: /admin/
Disallow: /api/private/

Also check AI crawler access. By default, most sites block GPTBot, ClaudeBot, and PerplexityBot inadvertently via overly broad Disallow: / rules targeting specific bots. If you want AI citation traffic, check which AI bots can crawl your site.

Step 4: Canonical tags — duplicate content under control

Canonical tags tell Google which URL is the definitive version of a page. Errors compound quickly:

• Self-referencing canonicals that point to the wrong URL — https://example.com/page/ canonicalising to http://example.com/page/ (missing HTTPS, missing or extra trailing slash)
• Canonicals pointing to non-200 pages — a canonical target that returns a 301 forces Google to follow the chain; a 404 canonical is ignored
• Pagination pages without proper canonical handling — each paginated page should either self-canonicalise or canonicalise to page 1
• CDN or caching serving different canonicals — what your CMS outputs vs. what arrives at the browser may differ

The most common canonical mistake: HTTPS site with HTTP in the canonical, or www. vs. non-www. mismatch. Run an audit to catch them across all crawled pages automatically.

Step 5: Core Web Vitals — real user performance data

Google uses three Core Web Vitals as ranking signals:

| Metric | What it measures | Good threshold | |--------|-----------------|----------------| | LCP (Largest Contentful Paint) | Load time of the largest visible element | <2.5 seconds | | INP (Interaction to Next Paint) | Responsiveness to user input | <200 milliseconds | | CLS (Cumulative Layout Shift) | Visual stability during load | <0.1 |

Important: Google uses field data (real Chrome user measurements, collected via CrUX), not lab scores. A page can score 95 on PageSpeed Insights in lab conditions but have poor field data — and it's the field data that affects ranking.

The fastest fixes for each:

• LCP: Preload the hero image, remove render-blocking scripts above it, use next-gen image formats (WebP/AVIF)
• INP: Remove or defer long JavaScript tasks that block the main thread
• CLS: Add width and height attributes to all images; avoid inserting content above existing content after load

Step 6: Structured data — schema for rich results and AI

Valid Schema.org JSON-LD unlocks rich results (star ratings, FAQ dropdowns, breadcrumbs) and improves AI citation eligibility. Required on every site:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Your Brand",
  "url": "https://yoursite.com",
  "logo": "https://yoursite.com/logo.png"
}

Priority schema by page type:

• Homepage: Organization + WebSite (with SearchAction for sitelinks search box)
• Blog posts: Article with datePublished, dateModified, author, image
• Product/service pages: Product with Offer, or Service
• FAQ sections: FAQPage with mainEntity array
• How-to content: HowTo with step array

Use the schema markup generator to build valid JSON-LD for any of these types, then validate with Google's Rich Results Test.

Step 7: Internal links — crawl depth and orphan pages

Internal links are how PageRank flows through your site and how crawlers discover new pages.

Check for:

• Orphan pages — pages with zero inbound internal links. If no other page links to it, crawlers may never find it, and it accumulates no internal PageRank.
• Pages deeper than 4 clicks — Google crawls frequently-linked pages more often. A page reachable only through a 6-click chain from the homepage gets crawled infrequently.
• Broken internal links (404s) — link equity goes to a dead end; user experience suffers.
• Links using JavaScript events instead of href — crawlers follow <a href="..."> links, not onclick handlers.

A quick architectural check: every important page (pillar content, high-priority landing pages) should be linked from at least 3 other pages and be reachable in ≤3 clicks from the homepage.

Step 8: Site architecture — URL structure and click depth

Good URL structure is:

• Flat and descriptive: /blog/technical-seo-audit beats /blog/2026/07/12/article-47389
• Consistent: pick a canonical format (/page/ vs. /page) and stick to it everywhere
• Free of URL parameters for content — /products?category=shoes is less crawlable and creates duplicate content risk vs. /products/shoes

For content sites: group related content under consistent URL prefixes (/blog/, /glossary/, /tools/) so crawlers can understand content relationships from structure alone.

Step 9: AI visibility — letting the new crawlers in

AI systems (ChatGPT, Claude, Perplexity, Gemini) have their own crawlers and their own way of consuming content. Five signals determine whether they can access and cite your site:

1. robots.txt: Are GPTBot, ClaudeBot, PerplexityBot, and other AI crawlers allowed?
2. llms.txt: Is there a machine-readable summary of your site at /llms.txt? (Generate one free)
3. FAQPage/HowTo schema: Do your key pages have structured Q&A markup?
4. Question headings: Are ≥20% of H2/H3 headings written as questions? ("What is X?", "How do I Y?")
5. Content chunkability: Are sections self-contained and under ~400 words average? LLMs retrieve passages, not whole pages.

AI visibility is not about gaming systems — it's about making your content easy for any machine to parse. The same patterns that make AI cite you also help with featured snippets and passage ranking.

Step 10: Meta tags — titles, descriptions, and OG tags

The most common meta tag audit finds:

• Duplicate titles across pages — every page must have a unique title (≤60 characters, keyword near the front)
• Missing or duplicate meta descriptions — descriptions should be 120–155 characters and written as ad copy (they drive CTR even though they're not a direct ranking signal)
• Missing og:image — pages shared on social or referenced in AI responses with no OG image get a generic grey box
• Title tags too long — anything over ~580px wide (~60 characters on average) gets truncated in desktop SERPs

Use the SERP preview tool to check how your title and description render in Google's interface before publishing.

The fastest way to run a technical SEO audit

This checklist is the manual version. The DeepSEOAnalysis free audit runs all 80+ checks automatically across your whole site and outputs a prioritised report — critical issues first, with specific fixes for each.

No signup. No email gate. No upsell required to see results.

---

Frequently asked questions

How often should I run a technical SEO audit?

Run a full audit whenever you: (1) migrate to a new CMS or hosting platform, (2) restructure your URL hierarchy, (3) add or remove large sections of content, or (4) notice a significant ranking drop that doesn't correlate with algorithm updates. Quarterly audits catch incremental drift — CMS plugins adding noindex tags, caches serving stale canonicals, new pages orphaned by missing internal links.

What's the difference between a technical SEO audit and an SEO audit?

A technical SEO audit focuses on infrastructure: crawlability, indexability, canonicals, Core Web Vitals, structured data, and site architecture. A full SEO audit also covers on-page factors (content quality, keyword targeting, heading structure) and off-page factors (backlink profile, domain authority). Think of technical as the foundation — if it's broken, on-page and off-page work can't fully compensate.

Does fixing technical SEO issues cause immediate ranking improvements?

It depends on the severity. Fixing a noindex on an important page can cause ranking recovery within days once Googlebot re-crawls. Fixing CLS across a site can improve Core Web Vitals scores (field data updates monthly). Fixing orphan pages improves crawl discovery but may take weeks for new internal PageRank to accumulate. Critical issues have the fastest impact; refinements accumulate over months.

What technical SEO issues hurt rankings the most?

In order of severity: (1) Crawl blocking — robots.txt or auth blocking Googlebot from key pages. (2) Noindex on pages that should be indexed. (3) Broken canonicals causing duplicate content. (4) Poor Core Web Vitals field data (LCP >4s is "Poor" and will be penalised in ranking). (5) Missing structured data (no rich results, reduced AI visibility). (6) Orphan pages with no internal links (won't be discovered or accumulate PageRank).