An SEO audit should explain what is wrong, why it matters, and what to fix first. Most checklists give you 200 items and no prioritization, which is why they sit unopened. Below is a tighter framework: ten categories, ordered the way we actually run an audit.

1. Crawlability and indexation

If search engines cannot reach or trust your pages, nothing else in this checklist matters. Start here.

  • Important pages return HTTP 200 (not 301 chains, not 404)
  • XML sitemap is present, accurate, and submitted in Search Console
  • robots.txt does not accidentally block important paths
  • No unintentional noindex tags on priority pages
  • Canonical tags point at the right URLs and resolve to a 200
  • JavaScript-rendered content is also visible to the raw HTTP response

2. Site architecture and internal links

  • Priority pages are reachable within 3 clicks from the homepage
  • Each priority page has at least 5 internal links pointing at it
  • Internal anchor text describes the destination (not "click here")
  • Hub-and-spoke patterns for topic clusters where applicable
  • No orphan pages (URLs in the sitemap with zero internal links)
  • Pagination handled cleanly (rel canonical to self, distinct H1s)

3. Metadata and on-page SEO

  • Unique <title> for every page, 50-60 characters, includes primary keyword
  • Unique meta description, 140-160 characters, written for CTR not robots
  • One <h1> per page, aligned with the query intent
  • Headings nest logically (H1 > H2 > H3, no skipping)
  • OG tags set for social/share previews; image returns 200
  • Canonical link in head matches the actual URL

4. Content quality and search intent

  • Each priority page targets a clear primary query and matches its intent
  • Pages have meaningful body content (200+ words minimum; usually much more)
  • No thin or duplicate pages that are competing with stronger ones
  • Original first-party content (data, examples, screenshots), not generic definitions
  • Top-of-page answers the query directly (especially for AI Overviews)
  • Updated dates reflect actual updates, not cosmetic refreshes

5. Technical performance and mobile

  • LCP under 2.5 seconds on real-world devices
  • INP under 200ms
  • CLS under 0.1
  • Mobile rendering matches desktop (no missing content)
  • Images use modern formats (WebP/AVIF) and proper width/height
  • Critical CSS inlined; non-critical JS deferred
  • TTFB under 600ms from primary geographies

6. Structured data and schema

  • Organization schema on the homepage with logo, sameAs, name
  • BreadcrumbList on deeper pages
  • FAQPage on pages with genuine FAQ content (no faking)
  • Article schema on editorial content with author, dates, abstract
  • Product/Offer schema on commerce-relevant pages
  • Validated via Rich Results Test or Schema.org validator
  • Schema content matches visible page content

7. Competitors and keyword gaps

  • Top 3 competitors identified by query overlap, not industry directory
  • Queries where competitors rank and you do not, prioritized by intent
  • Content topics they cover that you do not
  • Comparison and alternatives queries — do you own your own page or does G2/Reddit?

8. Analytics and conversion tracking

  • GA4 (or alternative) installed and not double-counting
  • Search Console verified and connected to GA4
  • Conversion events fire for forms, demos, signups, calls
  • Goals tagged with monetary value where possible
  • Page-level conversion data visible (not just sitewide)

9. AI search readiness

  • llms.txt present at the root of the domain (guide)
  • Entity disambiguation: schema, llms.txt, About page, and homepage all agree
  • FAQ schema on pages that would naturally be cited
  • Short, quotable sentences in the first paragraph of priority pages
  • Unique numbers and original data in pillar content
  • AI crawlers (GPTBot, ClaudeBot, PerplexityBot) not blocked in robots.txt unless intentional
  • Baseline check: ask ChatGPT and Perplexity "what is [your brand]" — is the answer accurate?

10. Prioritization framework

A checklist is only useful if it produces a small action list. After running through the above, rank findings on two axes:

  • Impact. How much traffic, citation share, or conversion would this fix plausibly move? Be honest — most findings are low-impact.
  • Effort.Engineering hours, content hours, or external work? An "easy" fix that requires a CMS migration is not easy.

Pick the top three by impact-per-effort. Ship those. Re-audit in 90 days. Most companies overspend on the long tail and underspend on the three things that would actually move the line.