The SEO content audit your dashboard isn't telling you to run.

1. Indexing & crawl health

Articles that never indexed cannot rank. This is the first thing to fix because every other optimisation depends on it.

• Every published article URL is submitted to Google Search Console.
• No article published over 14 days ago shows "Not indexed" in GSC.
• robots.txt allows crawling of /blog/ (or your content path).
• No accidental noindex meta tags on published articles.
• XML sitemap includes all published articles and is submitted to GSC.
• Sitemap lastmod dates update when articles change.
• canonical tag on each article points to itself, not a duplicate.
• Page Experience signals (CLS, LCP, INP) are in the "Good" range in GSC.

2. On-page fundamentals

The structural basics that determine whether Google can parse and rank your article. Most "AI content does not rank" complaints trace back to one of these.

Why this matters →

• Title tag is under 60 characters and contains the primary keyword.
• Meta description is 150-160 characters and contains the primary keyword.
• Exactly one H1 per article, containing the primary keyword.
• H2 sections each answer a sub-question someone would search after the main keyword.
• No skipped heading levels (no H2 directly to H4).
• URL slug is short, descriptive, uses hyphens, contains the primary keyword.
• OpenGraph image, title, description are populated for social sharing.
• Article schema (JSON-LD) is present and validates in Google's Rich Results Test.
• FAQ schema present if the article has a Q&A section.
• BreadcrumbList schema present and matches visible breadcrumb.
• Internal links from at least 2 existing posts point to this article.
• Article links out to 2-3 related articles with descriptive anchor text.

3. Search intent alignment

An article that targets the wrong intent will never rank, no matter how good the writing. The SERP rewards format match, not keyword density.

Why this matters →

• Searched the target keyword in incognito and noted the dominant SERP format (listicle, guide, comparison, video).
• Article format matches the dominant SERP format. A "best CRM" search returns comparisons; your article is a comparison.
• Heading structure matches the average of top-ranking pages, not a generic template.
• Article addresses the questions in the "People also ask" box for the target keyword.
• Word count is within 25% of the average of top 3 ranking pages. Not 3x longer for the sake of length.
• Article includes the specific entities (tools, frameworks, brands) that top-ranking pages cover.
• Search intent classified before generation (informational, navigational, commercial, transactional), not after.
• Article includes at least one concrete example or data point not found in the top 3 ranking pages.

4. Per-article performance

Aggregate traffic numbers hide what matters. The decisions that grow organic traffic happen at the per-URL level.

Why this matters →

• Every published article is tracked individually in Google Search Console (not just aggregate site metrics).
• Position, clicks, impressions, and CTR are logged at the 30-day, 60-day, and 90-day mark per article.
• You know the top 10 articles by clicks over the last 90 days. Without opening GSC.
• You know which articles brought zero clicks in the last 30 days.
• Estimated ad value (clicks × CPC of target keyword) calculated for top articles.
• GA4 conversion tracking attributes leads/signups back to the originating article.
• You can answer "which article drove the most revenue last month" in under 5 minutes.
• Per-article CTR vs SERP position is reviewed — if CTR is far below the position average, the title or meta description needs work.
• Articles stuck on page 2 (positions 11-20) are flagged for refresh, not abandoned.
• Articles that never reached page 1 in 90 days are reviewed for intent mismatch.
• Per-article tracking is automated. You are not manually copying GSC numbers into a spreadsheet every month.
• Quarterly review identifies which content formats produced the highest cost-per-ranking-article.

5. Decay detection

You lose 20-40% of organic traffic per year to articles that quietly stop ranking. The cure is catching it within weeks, not quarters.

Why this matters →

• You receive alerts when any article drops 5+ positions month-over-month.
• You receive alerts when any article loses 30%+ of clicks month-over-month.
• You watch for impressions holding steady while clicks fall (title or meta losing relevance).
• You watch for impressions falling while position holds (the SERP changed format and your article no longer matches).
• Decay signals are monitored per URL, not just in aggregate dashboards.
• Top 10 articles by traffic are reviewed monthly for early decay signals.
• Articles older than 12 months are reviewed at least quarterly even if they appear stable.
• Detection runs automatically. You are not relying on noticing problems in aggregate reports.
• Detection thresholds are calibrated. Too tight and you ignore the alerts; too loose and you miss real decay.
• Alerts route to a person or system that will actually act on them within 30 days.

6. Refresh & maintenance

Refreshing a decaying article restores 60-90% of its original traffic. New articles have to earn traffic from scratch. The math favours maintenance.

Why this matters →

• Decay-alerted articles are refreshed within 30 days of the alert.
• Refresh preserves the URL, slug, and metadata. No accidental redirects.
• Refresh preserves all inbound internal links (no URL change means no link breakage).
• Refresh re-analyses the current SERP for the target keyword, not just edits the old article.
• Refreshed articles get their published date updated (Google uses this as a freshness signal).
• Refreshed URLs are re-submitted to GSC for crawling.
• Refresh budget allocated: roughly 1 hour per week per 50 articles published.
• Refresh is prioritized over new content when more than 5 articles are flagged as decaying.
• Version history of each article is preserved so refresh can be rolled back if it underperforms.
• Refresh performance is measured at 30 days post-refresh. If traffic does not recover, the article is retired or rewritten from scratch.