mkt-seo-audit

Audit a page or site across keywords, on-page, content gaps, and technical health, then return a prioritized impact/effort action list.

SKILL.md

GitHub

When to use

Use this when you need to know why a page or site underperforms in search and what to fix first, covering the four layers that move rankings: keywords, on-page, content, and technical.

Not for: writing the optimized content itself (use mkt-draft-content), paid search/SEM, or claims that require live crawl or analytics data you do not have — mark those as unverifiable.

Method

  1. Set scope and intent. Confirm the URL(s), audience, and primary search intent (informational, commercial, navigational, transactional). Decision point: if intent is unknown, infer from the content and state it.
  2. Keywords. Identify the primary keyword, secondary and long-tail variants, and intent match. Flag cannibalization or missing target terms.
  3. On-page. Check title tag, meta description, H1 and heading hierarchy, keyword placement, internal links, image alt text, URL structure, and readability.
  4. Content gaps. Compare coverage against ranking pages; find missing subtopics, thin sections, outdated facts, and schema/FAQ opportunities.
  5. Technical. Review indexability (robots, canonical, sitemap), speed signals, mobile-friendliness, structured data, and broken links — at the level the available data allows. Decision point: mark what needs a live crawl to confirm.
  6. Prioritize. Score each finding by impact and effort, then order the fixes with quick wins flagged.

Example

Page targets "expense management software" but the H1 says "Welcome." On-page: no keyword in title, meta description missing, one H2. Content gap: ranking rivals all cover "integrations" and "pricing" — this page covers neither. Fix order: (1) rewrite title + H1 with keyword (high impact, low effort), (2) add meta description, (3) add integrations + pricing sections. Technical: canonical looks correct but needs a live crawl to confirm indexing.

Pitfalls

  • Chasing technical minutiae while the page targets the wrong keyword or intent.
  • Recommending fixes without impact/effort scoring, so everything looks equally urgent.
  • Asserting technical findings (index status, speed) that actually require a live crawl — label them.
  • Optimizing for a keyword whose intent does not match the page's purpose.

Output format

Scope + target keyword + search intent: <...>

Findings by layer:
- Keywords: <short list>
- On-page: <short list>
- Content gaps: <short list>
- Technical: <short list>

Issue table:
| Area | Issue | Impact (H/M/L) | Effort (H/M/L) | Fix |

Prioritized action plan: <top items first, quick wins flagged>
Unverifiable without live crawl / analytics: <...>

Reference

Audit workflow (run in this order)

  1. Crawl the site (Screaming Frog / Sitebulb) for status codes, titles, canonicals, redirects.
  2. Pull Google Search Console: top queries, pages, CTR, coverage/indexing errors.
  3. Run PageSpeed Insights on the 3 templates (home, a product/feature page, a blog post) — not all 20 pages; templates repeat.
  4. Check site:domain.com and branded queries for what's actually indexed.

P0 catastrophic checks (an audit that skips these can pass a broken site)

  • JS rendering: compare rendered vs raw HTML (GSC URL Inspection "view crawled page"). SaaS sites on React/Next/Webflow often serve content only after JS — Google may see an empty <div>. This invalidates every downstream finding.
  • Accidental noindex / robots block on production pages.
  • Staging/dev subdomain leak: site:staging. / site:dev. — non-prod environments getting indexed (add noindex + auth).

Core Web Vitals thresholds (good)

LCP < 2.5s · INP < 200ms · CLS < 0.1 · TTFB < 0.8s. Report each as good / needs-improvement / poor per template.

On-page patterns

  • Title: Primary Keyword | Benefit — Brand (≤ ~60 chars). Meta description ≤ ~155 chars with the keyword + a CTA.
  • One H1 per page, keyword-bearing. Anti-pattern: generic H1s like "Welcome" or "Simple plans for everyone" that carry no query.
  • Add OpenGraph + Twitter card tags (CTR in social/SERP snippets).
  • SaaS-specific: build comparison / "vs" / alternatives pages — highest commercial-intent pipeline lever.

Prioritization quadrant

Score each fix Impact (H/M/L) × Effort (H/M/L):

  • High-impact / low-effort → do now (quick wins).
  • High-impact / high-effort → plan (e.g. rendering fixes, content builds).
  • Low-impact → backlog. Backlinks are the longest-lead-time lever — start early even though results lag.