Methodology

What this resource does

GEO Primer is an English-first learning resource. We explain AI search, interpret source material, and offer bounded checks that support human editorial judgment. We do not observe proprietary ranking systems or predict whether a page will be cited.

How we describe evidence

• Published research: findings are tied to the study’s method, sample, version, and experimental setting. A research interpretation is not a replication.
• Platform documentation: a provider’s description of its own product, limited to the scope and date of that source. Eligibility is not guaranteed inclusion.
• Observation: a recorded answer, mention, or citation under stated conditions. A small observation set is not a platform-wide estimate or a causal test.
• Editorial recommendation: a practical judgment about clarity, evidence, or reviewability. We label these judgments rather than presenting them as measured ranking factors.

Research notes show evidence type, scope and sample, interpretation, limitations, original sources, and a review date. Publication and review dates describe our entry, not the source’s publication date.

Knowledge assessment boundaries

The GEO Knowledge Assessment uses 18 single-choice scenarios across foundations, discovery and access, content evidence, brand trust, measurement, and implementation risk. A correct-count result is a check of knowledge under that question set—not certification, a ranking, or a website audit. Questions and scoring carry a version; explanations and learning links accompany a completed result.

Content structure checks

The Content Structure Checker accepts English plain text or Markdown, up to 30,000 characters, and runs in the browser. It inspects heading order, paragraph length, lists or steps, and Markdown source links. Code fences are not treated as prose, and raw HTML is not executed.

Paragraph-length thresholds are transparent editorial heuristics, not validated GEO ranking or citation factors. A link’s presence does not establish credibility or whether it supports a claim. Results separate Observed structure from things to Review manually, including factual accuracy, source quality, reader intent, and accessibility in the actual published page.

Versions and changes

31 August 2026 — initial phase-one methodology. Defined the editorial evidence labels and the boundaries of the two free tools. Tool-specific versions and check definitions are shown with their working surfaces. Changes to a check or question set should be recorded with a new version; observations across versions should not be silently combined.

Corrections and review

We link to original sources and record review dates so readers can inspect the reasoning. If a claim appears inaccurate or out of date, prepare the page URL, the claim, a supporting source, and the proposed correction. See Contact for the current contact availability.

No guide, checklist, or tool promises citations, rankings, traffic, or sales.