tmj-90 Audit SEO systematically
Three layers in impact order: technical (crawl/index/speed) → on-page (titles/headings/links) → content (intent match / E-E-A-T / thin pages). Fix the highest layer first.
Core Web Vitals pass/fail thresholds (75th percentile real-user data)
| Metric | Good | Needs work | Poor |
|---|---|---|---|
| LCP | ≤ 2.5s | 2.5–4.0s | > 4.0s |
| INP | ≤ 200ms | 200–500ms | > 500ms |
| CLS | ≤ 0.1 | 0.1–0.25 | > 0.25 |
Technical layer (check first)
- Crawlability —
robots.txtnot blocking important paths; XML sitemap present and submitted to Search Console; no orphan pages (no internal links). - Indexation —
noindexnot set on pages that should rank; canonical tags point to the correct URL; HTTPS on all pages. - Speed — LCP from real-user data (Search Console → Core Web Vitals report); identify LCP element; check image optimisation + CDN + render-blocking resources.
- Mobile — Google uses mobile-first indexing; test with Mobile-Friendly Test.
On-page layer
- Title tags — ≤ 60 chars; primary keyword near the front; unique per page; describes content accurately.
- Meta descriptions — ≤ 160 chars; includes a CTA-like phrase; unique per page.
- H1 — one per page; matches or closely mirrors the title tag; primary keyword present.
- Heading hierarchy — H1 → H2 → H3 without gaps; no keyword stuffing.
- Internal linking — key pages linked from multiple internal pages; anchor text descriptive (not "click here").
Content layer
- Search intent match — is the page format what the searcher expects? (Informational → guide/article; transactional → product/pricing page; navigational → homepage).
- E-E-A-T signals — author bio with credentials; publication date + last-updated date; primary-source citations; specific examples and data.
- Thin/duplicate content — pages with < 300 words of unique content; pages with duplicated sections across the site.
- Keyword targeting — primary keyword in title + H1 + first 100 words + URL slug; LSI terms used naturally in body.
Steps
- Scope the audit. Full site or specific pages? Technical + on-page, or one focus? Access to Search Console and analytics? Establish baseline: current organic traffic trend, target keywords, known issues.
- Technical audit. Crawl with Screaming Frog or Search Console; flag crawl errors, indexation issues, redirect chains, broken internal links.
- Core Web Vitals. Pull from Search Console → Core Web Vitals report (real-user data). Identify pages failing LCP/CLS; prioritise by traffic volume.
- On-page audit. Check title tags, H1s, meta descriptions for top 20 pages by traffic. Flag duplicates, over-length, keyword mismatches.
- Content audit. Identify thin pages (< 300 words unique); spot intent mismatches; E-E-A-T gaps.
- Prioritised recommendations. Order by: impact × effort. Technical blocking issues first; then on-page quick wins; then content investment.
- Verify. After implementing: recheck Search Console for crawl/index errors; re-run Lighthouse for CWV; confirm fixes appear in next crawl.
Review checklist
- No pages blocked in robots.txt that should rank.
- Sitemap submitted and valid — no 4xx/5xx URLs in sitemap.
- Title tags unique and ≤ 60 chars on all indexed pages.
- LCP ≤ 2.5s on mobile for highest-traffic pages.
- No thin-content pages (< 300 words unique) ranking for commercial queries.
- Internal links to key pages from multiple relevant pages with descriptive anchor text.
Rules
- Technical issues that block crawl/index are fixed before content work — no point optimising a page Google can't read.
- Never recommend keyword stuffing — intent match and E-E-A-T outperform density in modern search.
- Core Web Vitals from real-user data (Search Console) override Lighthouse scores — both matter but prioritise the real-user signal.
References
references/seo-audit-checklist.md — full 80-point checklist by layer.