GeoTest — Methodology · GEO v3.0

How the GEO Score works

A 0–100 rating built from six weighted categories, calibrated against citation behavior observed across ChatGPT, Perplexity, Google AI Overview, Claude, and Gemini (2024–2026).

Six weighted categories · 100 points total

CategoryWeightWhat it measures
Content Quality25 ptsDepth, heading hierarchy, cornerstone pages of 1,500+ words.
Citation Worthiness20 ptsFactual density, quotable definitions, comparison tables, freshness.
Structured Data18 ptsOrganization, WebSite, Product, Article, Breadcrumb, FAQPage JSON-LD.
Authority Signals17 ptsAuthor bylines, About/Team pages, founding date, contact info.
FAQ & Q&A Format12 ptsVisible Q&A sections plus FAQPage schema with 5+ entries.
Crawler Accessibility8 ptsViewport meta, canonical, no noindex, server-rendered text.

Sites scoring 80+ are reliably cited by Perplexity and Google AI Overview for their core keywords. Sites under 50 are usually invisible to generative engines — regardless of their traditional SEO authority.

What happens during a scan

  1. Fetch — crawl & render. We fetch your page the way an AI crawler does — checking viewport, canonical, noindex, and server-rendered text.
  2. Parse — extract signals. Headings, factual density, definitions, tables, schema blocks, bylines, FAQ structure — everything the rubric weighs.
  3. Score — apply the rubric. Each category is scored against calibrated thresholds and live citation outcomes.
  4. Fix — rank the fixes. Issues are ranked by point impact and effort, with copy/paste schema and content rewrites ready to ship.

Pro Site Analysis goes further: it crawls your site from the homepage, analyzes up to 200 pages (larger sites are sampled), and runs live citation tests against Perplexity (sonar-pro).

The full technical write-up is published on our methodology page.