What Is a Citability Score?

A Citability Score is a composite measure of how likely an AI model is to select a specific page as a citation source when generating answers. It's not an official metric from any AI platform — it's a practical framework for evaluating content across the dimensions that our research shows most strongly predict citation selection.

In traditional SEO, Domain Authority provided a proxy for ranking potential. Citability Score does something analogous for AI search: it gives you a structured way to diagnose why some of your pages are being cited frequently and others never, and what specific changes will improve citation rates.

The Five Dimensions of Citability

Our analysis identified five measurable dimensions that collectively predict citation probability. These dimensions are weighted based on their observed correlation with citation frequency across different AI platforms:

• Factual Density (25% weight): The concentration of specific, verifiable data points per 500 words of content
• Structural Clarity (25% weight): The clarity and organization of headings, sections, and content architecture
• Authority Signals (20% weight): EEAT indicators including author credentials, schema implementation, and entity recognition
• Topical Completeness (20% weight): The percentage of common questions on the topic that the page answers
• Technical Accessibility (10% weight): The success rate of AI crawler access, including rendering and speed

The remaining 10% is accounted for by platform-specific factors (Bing indexation for ChatGPT, recency weighting for Perplexity, Google entity recognition for Gemini) that vary by target platform.

Dimension 1: Factual Density

Factual density measures how many specific, verifiable claims — statistics, percentages, named entities, dates, measurements — appear per 500 words of content. AI models strongly prefer content that makes specific claims over content that makes general assertions.

Scoring guide for factual density:

• High (20+ score): 8+ specific data points per 500 words. Example: percentages, survey results, named studies, specific tool names, measurable thresholds.
• Medium (12–19 score): 4–7 specific data points per 500 words. Some specific claims, but significant sections of generalization.
• Low (0–11 score): Fewer than 4 specific data points per 500 words. Primarily general statements and unsupported assertions.

Improving factual density: Replace vague claims with specific ones. "Many websites see improvement" becomes "74% of websites that implemented FAQ schema saw AI citation improvements within 8 weeks." Cite external studies when you don't have proprietary data. Every section should contain at least one specific, citable data point.

Dimension 2: Structural Clarity

Structural clarity assesses how easily an AI model can identify what your content is about, where specific information lives, and how pieces of information relate to each other. Highly structured content enables reliable extraction; poorly structured content produces unreliable chunks that are often misattributed or skipped.

Scoring criteria for structural clarity:

• Heading hierarchy: Proper H1→H2→H3 progression without skipped levels. Each H2 should represent a distinct topic section. H3s should be subtopics of the preceding H2.
• Section answer-lead: Does each H2 section open with a direct answer in the first 80 words? Answer-led sections score significantly higher than sections that build to their answer.
• List and table usage: Are enumerable items presented as lists rather than embedded in prose? Are comparisons shown in tables rather than described in paragraphs?
• Section length: Are sections between 200–400 words — the optimal extraction chunk size? Very short sections (under 100 words) lack depth; very long sections (over 600 words) exceed standard extraction windows.

Dimension 3: Authority Signals

Authority signals are the elements AI models use to assess whether content comes from a source with genuine expertise and accountability. This is the EEAT dimension of citability scoring.

Key authority signals and their weight in citation decisions:

• Named, credentialed author (high weight): Content attributed to a named author with verifiable credentials and a Person schema entity scores substantially higher than anonymous or team-attributed content.
• Schema implementation completeness (medium-high weight): Article schema with datePublished, dateModified, and linked author entity. FAQPage schema on Q&A content. Organization schema on brand pages.
• External citations (medium weight): Content that references and links to credible external sources is treated as more authoritative than content making equivalent claims without attribution.
• Original data or research (medium weight): Content presenting unique information that exists nowhere else on the web receives an authority premium — it's the only source for that specific information.
• Content freshness (medium weight): dateModified in schema that reflects recent updates. Content that was relevant a year ago and has been explicitly maintained and updated signals ongoing expertise commitment.

Dimension 4: Topical Completeness

Topical completeness measures what percentage of common questions on a topic your content answers. A page that answers 4 of the 10 most common questions about a subject will be cited less often than a page that answers 9 of 10 — even if the 4 answers are slightly better written.

Measuring and improving topical completeness:

• Compile the question set: Use Google's "People also ask" for your topic, competitor FAQ sections, and forum questions (Reddit, Quora) to build a list of 20–30 common questions.
• Map questions to your content: For each question, does your page provide a direct, complete answer? Mark as fully answered, partially answered, or unanswered.
• Calculate your coverage rate: (Fully answered / Total questions) × 100 = Topical Completeness score. Above 80% is strong; below 60% indicates significant citation opportunity gap.
• Fill the gaps: Add sections or expand existing sections to cover unanswered questions. Each additional answered question is a new query variant where your page can be cited.

Dimension 5: Technical Accessibility

Technical accessibility is a binary qualifier more than a gradient: either AI crawlers can successfully access, render, and extract your content, or they can't. Technical failures disqualify a page from citation regardless of how excellent its content is.

Technical accessibility checklist:

• Key content present in server-rendered HTML (not JavaScript-dependent rendering)
• Page load time under 2 seconds; TTFB under 500ms
• No robots.txt rules blocking AI crawler user agents
• Valid SSL certificate and HTTPS enforcement
• Page indexed in Bing (for ChatGPT citation eligibility)
• No crawl errors (4xx, 5xx) returned for the page URL

How to Improve Your Citability Score

Prioritize improvements by the weighted impact of each dimension and the ease of implementation:

• Quick wins (high impact, low effort): Add FAQPage schema to existing Q&A content. Add dateModified to Article schema. Move direct answers to the top of each section.
• Medium effort (high impact): Expand content to answer more questions in your topic area. Add specific data points to replace vague claims. Create named author pages with Person schema.
• Longer effort (essential for technical failures): Fix any JavaScript rendering issues. Submit to Bing Webmaster Tools. Resolve crawl errors identified in Google Search Console.

Run a citability assessment on your top 10 pages by organic traffic. Pages scoring above 70 are citation-ready. Pages scoring 40–70 need targeted improvements in their lowest-scoring dimensions. Pages below 40 need comprehensive restructuring before they'll earn consistent citations.