Content Depth: Why Thin Pages Are Invisible to AI

AI engines need enough content on a page to determine whether it's worth citing. A 200-word product page that says "We offer live chat. Contact us for a demo" gives AI nothing to work with. There's no fact to extract, no comparison to reference, no depth to cite.

Content Depth carries 7% of your total AEO Site Rank. That makes it the third heaviest individual criterion, behind only Topic Coherence (14%) and Original Data (10%). The three together form the Content Substance tier, which accounts for roughly 55% of your entire score.

The scorer evaluates three dimensions: - Average word count across your key pages (homepage + blog sample) - Heading structure - how many H2 and H3 headings provide navigable structure - Deep-to-thin ratio - the proportion of pages with 1,500+ words versus pages under 500 words

A site where most content pages run 2,000+ words with clear heading hierarchy scores 9-10. A site dominated by 300-word stub pages scores 1-2. Most sites fall somewhere in between, with a mix of deep pillar content and thin product pages.

There's a cross-criterion interaction too. Content Depth receives an additional penalty when Topic Coherence is weak. Deep content on scattered, unrelated topics doesn't build authority - it just creates more noise. The scorer accounts for this by accepting a topicCoherenceScore parameter and applying a cap when coherence is low.

Word count is necessary but not sufficient. A 3,000-word wall of unstructured text scores worse than 2,000 words organized under clear headings. Here's what the scorer actually measures:

Word count thresholds: - Under 500 words: "Thin" content. Unlikely to contain enough substance for AI to cite. These pages drag down your site average. - 500-1,200 words: Moderate. Sufficient for simple topics but not competitive for complex queries. - 1,200-2,000 words: Solid. This is where most citation-worthy content lives. - 2,000+ words: Deep. These pages have multiple citation candidates and tend to be the ones AI actually references.

Heading structure scoring: - 0-2 headings on a long page: Poor structure. AI can't navigate the content efficiently. - 3-5 headings: Adequate. Each heading creates a potential extraction point. - 6+ headings: Strong. Clear topical segments that AI can independently evaluate and cite.

The deep/thin ratio: If 60% of your pages are under 500 words, your Content Depth score will be low regardless of how deep your best content is. The scorer evaluates the whole sample, not just the outliers. One magnificent 5,000-word pillar article can't compensate for 20 thin stub pages.

The practical implication: every page on your site doesn't need to be 3,000 words. But the pages that represent your core expertise - product pages, pillar articles, comprehensive guides - need to be substantive. And you need more deep pages than thin ones.

The worst way to improve this score is to pad your existing content with fluff. "In today's fast-paced digital world" repeated across 20 pages doesn't add depth - it adds noise. Here's how to add substance:

1. Expand with specific data Take your existing claims and back them with numbers. "Our platform is fast" becomes "Average response time: 1.2 seconds. 99.97% uptime over the past 12 months. Load tested to 10,000 concurrent connections." Same topic, triple the depth, infinitely more citable.

2. Add comparison sections AI engines love structured comparisons. A "How We Compare" section with a feature matrix adds 300-500 words of highly citable content. Use HTML tables, not prose - they're more extractable.

3. Include implementation details Turn "easy to install" into a step-by-step guide with code snippets, configuration examples, and common troubleshooting. This adds depth that's genuinely useful and positions you as the authoritative source.

4. Build FAQ sections into key pages 5-8 frequently asked questions with detailed answers can add 500-1,000 words of structured, Q&A-formatted content. Each question-answer pair is an independent citation candidate.

5. Add heading structure to existing long content Some pages already have the word count but lack structure. Adding H2 headings every 300-400 words makes the same content more parseable by AI and improves both the heading structure score and the overall depth evaluation.

Start here: identify your 10 most important pages. Check word counts. Any under 1,000 words? Those are your highest-impact expansion targets.

Content Depth doesn't exist in isolation. The scorer accepts a topic coherence score and applies a cross-criterion cap when coherence is low. The logic: deep content on random topics isn't authority - it's just a lot of words.

If your Topic Coherence score is 4/10, your Content Depth score gets capped at 6 regardless of actual word counts. This prevents a scenario where a scattered blog full of long but unfocused articles gets rewarded for depth alone.

The practical takeaway: fix coherence first, then invest in depth. Writing a 5,000-word article on a topic unrelated to your core business hurts you twice - it lowers coherence and the depth it adds gets capped anyway.

The winning formula is focused depth: deep content on topics that reinforce your thematic clusters. A customer support company with five 2,500-word articles all interconnected around support automation will outperform one with twenty 3,000-word articles scattered across unrelated topics.