Report Methodology

Sample

Our reports are based on a curated sample of websites drawn from the Tranco Top 1M list, an academic ranking of the most popular websites globally that aggregates Cisco Umbrella, Cloudflare Radar, and Majestic rankings.

From this list we removed obvious infrastructure domains (CDNs, DNS providers, cloud platforms, ad tech, tracking domains, etc.) and took the top 1.000 remaining end-user websites as our baseline sample.

Scoring

Each website is analyzed using GEO-Score's automated audit engine, which evaluates 22 metrics across 4 pillars:

Content Quality (7 metrics): readability, comprehensiveness, transition words, LSI keywords, answer completeness, factual density, citability
AI Readiness (6 metrics): AI optimization, structure, image alt text, platform specificity, content type matching, semantic clarity
Technical Basis (5 metrics): AI bot crawler access, schema validation, sitemap discoverability, page speed, content freshness
Authority & Trust (4 metrics): citations, E-E-A-T, knowledge graph presence, topical authority

Scores range 0-100. The overall GEO Score is a weighted average tuned per page type (homepage, blog post, product page, etc.).

Limitations & caveats

Sample is global but skewed toward English-language sites. The Tranco Top 1M heavily features .com/.org/.net domains.
JavaScript-rendered sites may score lower because the crawler renders the initial HTML. Sites with critical content loaded via JS get partial scoring.
Bot-blocking sites cannot be analyzed and are recorded as failed crawls. Their exclusion may bias the sample toward more crawl-friendly sites.
Some metrics are correlational not causal. A high GEO score correlates with AI-citation visibility but isn't the sole determinant.

Data freshness

Reports are regenerated on the first business day of each new period (month / quarter / half-year). Underlying GEO-Score audits use the latest snapshot at time of report generation; an audit is at most 30 days old when included in a periodic report.

Last updated: 2026-06-02 · Back to reports

Sample

Scoring

Each website is analyzed using GEO-Score's automated audit engine, which evaluates 22 metrics across 4 pillars:

Content Quality (7 metrics): readability, comprehensiveness, transition words, LSI keywords, answer completeness, factual density, citability

AI Readiness (6 metrics): AI optimization, structure, image alt text, platform specificity, content type matching, semantic clarity

Technical Basis (5 metrics): AI bot crawler access, schema validation, sitemap discoverability, page speed, content freshness

Authority & Trust (4 metrics): citations, E-E-A-T, knowledge graph presence, topical authority

Scores range 0-100. The overall GEO Score is a weighted average tuned per page type (homepage, blog post, product page, etc.).

Limitations & caveats

Sample is global but skewed toward English-language sites. The Tranco Top 1M heavily features .com/.org/.net domains.

JavaScript-rendered sites may score lower because the crawler renders the initial HTML. Sites with critical content loaded via JS get partial scoring.

Bot-blocking sites cannot be analyzed and are recorded as failed crawls. Their exclusion may bias the sample toward more crawl-friendly sites.

Some metrics are correlational not causal. A high GEO score correlates with AI-citation visibility but isn't the sole determinant.