Skip to main content
Methodology

A traceable GEO methodology

The score uses 12 components, versioned weights, and source-quality labels. Some signals are direct, while others are proxy-based or declared.

Read the methodology View a sample result
12 components with source labels Diagnostics and supporting signals The 13 user intents
Visualization for the page: A traceable GEO methodology
Definition and protocol

12 components with source labels

The GEO Visibility Score covers K1–K12. For each component, the platform stores a direct, proxy, or declared source status and a reference to the material used in calculation.

Not all twelve values should be treated as directly measured. Source status is part of correct score interpretation.

Methodology rule

Diagnostics and supporting signals

Diagnostic views break components down into inputs and evidence stored for the run. They help locate weaker areas and open answers, audit data, or declared information.

Not every diagnostic is an independent measurement. Some explain the score through proxies, so a recommendation must identify its source status.

Methodology rule

The 13 user intents

Measurement questions are built around thirteen structured user intents, covering the journey from category discovery to decision. The fixed intent structure makes results comparable across brands, markets, and measurement cycles. The questions within each intent are tailored to the brand category, but the structure itself stays constant.

Methodology rule

The score formula

How components aggregate into the GEO Visibility Score is published: weights, normalisation, and layer-combination rules are explicit and versioned. The same input data always yields the same score — the formula is deterministic. Changing the formula means a new methodology version, recorded in the changelog.

Methodology rule

Measurement confidence

Every score carries a Measurement Confidence value — a measure of how complete and reliable the underlying measurement was. Confidence accounts for scope coverage, answer repeatability, and data recency, among other factors. We consider a score without confidence information incomplete — which is why confidence is part of every report.

Methodology rule

Protocol and collector limits

A run stores the question set, engine, language, country, pass number, full answer, and disclosed citations.

Before interpreting a result, review measurement mode, provider statuses, and answer completeness. Missing technical data should not be treated as confirmed brand absence.

Methodology rule

Recommendations and their basis

Recommendations use the score, technical audit, answers, and a template library. The selector considers component, priority, cost, and constraints before assigning a horizon.

Not every recommendation has direct causal evidence. For proxy signals, its role is to identify a hypothesis for implementation and later review.

Methodology rule

Methodology limitations

The methodology has explicitly described boundaries: AI answers are non-deterministic, models change, and a visibility score is not a revenue forecast. We publish the full list of limitations and the rules for interpreting score changes. Honest measurement boundaries are a precondition of its usefulness.

Methodology rule

Versions stored with the result

The result stores the methodology version, sub-index version, weight profile, weight checksum, question-set version, and collector versions.

The public page is not yet a complete log of every implementation change. Run metadata and source-control history remain the basis for technical audit.

Continue exploring

Related GeoFolks areas

Compare plans
Methodology Measurement confidence

Measurement confidence combines coverage, source fidelity, repeatability, freshness, and validation. These are technical indicators, not guarantees that an answer is true.
Methodology Providers and measurement channels

GeoFolks uses configured API integrations. A model name, consumer application, and API channel are not the same environment.
Methodology Measurement quality control

Automated checks run for every report. Human review is confirmed only by a recorded QA status and reviewer.
Next step

Trace the complete method

The score uses 12 components, versioned weights, and source-quality labels. Some signals are direct, while others are proxy-based or declared.

Read the methodology Discuss the scope