Why a standard, and why open?

AI answers are probabilistic: the same question asked twice can name different brands. Without a shared measurement discipline, every GEO claim is unfalsifiable marketing. We publish this framework openly — definitions, formulas, prompt-set design and scoring — so clients, competitors and researchers can apply it, challenge it, and improve it. An open standard is also a bet: if the industry measures AI visibility our way, this encyclopedia becomes the reference every engine cites. That is GEO, demonstrated on ourselves.

The four dimensions

01

Visibility

Appearance rate: across a graded prompt set for your category, how often does each engine name your brand? Scored 0–100 per engine.

Read the Visibility standard →

02

Citation Rate

Credit quality: when engines answer about you or your category, which URLs do they cite, and what authority tier are those sources?

Read the Citation Rate standard →

03

Semantic Coverage

Question-space completeness: what share of your category's full buyer-question matrix can engines answer with you in it?

Read the Semantic Coverage standard →

04

Brand Cognition

Accuracy and sentiment: is what AI says about you correct, current and favorable — and consistent across engines?

Read the Brand Cognition standard →

How the composite works

Each dimension is measured per engine against the same prompt set, then aggregated with published weights (Visibility 35%, Citation Rate 25%, Semantic Coverage 25%, Brand Cognition 15%). The full prompt-set design rules, grading rubric and cadence recommendations are in the measurement methodology chapter. The GEO Checker runs a lightweight version of this framework free.

Frequently asked questions

Are these standards proprietary?
The definitions, formulas and methodology are published openly on this site for anyone to use. Our commercial product is the measurement infrastructure and execution — the Checker, monitoring cadence, and the fix lists — not secrecy about the framework.
Why these four dimensions and not others?
They map one-to-one onto the four stages of the engine pipeline: Visibility measures retrieval+selection, Citation Rate measures the citation stage, Semantic Coverage measures grounding breadth, and Brand Cognition measures synthesis accuracy. A dimension that doesn't map to a pipeline stage can't be systematically improved.
How often should the four dimensions be measured?
Monthly at minimum for active categories; weekly during campaigns or reputation events. AI answers drift as models refresh and source graphs change — a quarterly cadence misses most movements.