The 8-Signal GEO Framework

GEO scoring is built on eight measurable signals that determine how AI systems perceive, trust, and cite a website. Each signal is independently assessable and contributes a weighted component to the composite GEO score.

The Eight Signals

Signal 1: Structured Data

Presence and quality of Schema.org JSON-LD markup. AI systems use structured data to understand content meaning without natural language inference. Valid, domain-appropriate schema types on every page are essential.

Signal 2: Crawlability

Whether AI crawlers can access and parse your content. This includes robots.txt directives that explicitly allow AI bots, server-rendered HTML (not client-side JavaScript), and reasonable response times under 500ms.

Signal 3: Bot Crawl Activity

Actual crawl traffic from AI systems — are GPTBot, ClaudeBot, PerplexityBot, and Google-Extended actively visiting your site? Crawl frequency and page coverage indicate how much AI systems value your content.

Signal 4: Content Authority

The depth, specificity, and expertise demonstrated in your content. AI systems cite sources that provide authoritative, original information — not thin content or aggregated summaries.

Signal 5: Citation Patterns

Whether your content currently appears in AI-generated responses. Tracking actual citations across ChatGPT, Claude, Gemini, and Perplexity provides ground-truth validation of GEO effectiveness.

Signal 6: Performance

Server response time and reliability. AI crawlers have timeout thresholds — if your pages are slow, they may not be fully indexed. Edge-served pages with sub-100ms TTFB score highest.

Signal 7: AI Surface Files

Presence and correctness of AI-specific discovery files: llms.txt (the AI equivalent of robots.txt), ai-content-index.json, MCP manifests, and properly configured sitemap.xml with accurate lastmod dates.

Signal 8: Protocol Support

HTTP/2 and HTTP/3 support, proper SSL configuration, canonical URL handling, and correct Content-Type headers. Protocol-level signals affect how efficiently crawlers can fetch and trust your content.

How GEO Scoring Works

Each of the eight signals is scored independently on a 0-100 scale. Scores are stored in the geo_score_dimensions table with individual weights that reflect each signal's relative importance to AI citation likelihood.

The composite GEO score is a weighted average of all eight dimension scores. The formula:

Composite = (score1 * weight1 + score2 * weight2 + ... + score8 * weight8) / (weight1 + weight2 + ... + weight8)

Signal statuses are tracked in the geo_signal_status table with PASS, FAIL, or PARTIAL states. Every optimization action is logged in the geo_ledger_entries table for full audit trail and impact measurement.

Daily automated audits recompute scores and generate trend data, enabling objective measurement of GEO infrastructure improvements over time.