Share-of-Answer (SoA) Methodology

Generative Engine Optimization (GEO) requires objective, repeatable metrics. Citepath measures how frequently and accurately major AI models recommend your brand when users ask commercial and high-intent queries.

1. Core Metrics & Weighting

Our overall Share-of-Answer score (0-100%) synthesizes two main dimensions:

  • Direct Citations (60% weight): The model includes an explicit hyperlink or source attribution back to your domain in its response grounding.
  • Brand Mentions (40% weight): The model names your brand or product in the generated text, even if an explicit hyperlink is omitted.

2. Supported LLM Engines

Scans test standardized buyer prompts across four leading AI answer engines:

ChatGPT (OpenAI)

Evaluated using web-search enabled GPT-4o models.

Claude (Anthropic)

Evaluated on Sonnet models with live web-grounding.

Gemini (Google)

Evaluated using Google Gemini Flash/Pro grounding.

Perplexity AI

Evaluated via live Perplexity Sonar search index.

3. Technical Readiness Audits

Beyond LLM responses, Citepath checks machine-readiness technical factors:

  • Robots.txt access: Verification that AI web crawlers (GPTBot, ClaudeBot, PerplexityBot) are not blocked.
  • JSON-LD Structured Data: Presence of Schema.org Organization, Product, or SoftwareApplication metadata.
  • llms.txt Availability: Check for machine-readable site summaries at /llms.txt.
  • Sitemap & Feed access: Discoverable site maps and RSS/Atom feeds for rapid indexing.

4. Objective Sourcing & Integrity

Citepath presents measured response data without unsourced claims. Competitor benchmark statistics are derived directly from real-time LLM prompt responses gathered during scan executions.

Explore live sample data

See sample reports generated using this methodology.

View Sample Reports →