GEO Content Auditor and weighted 5-axis score

How it works

The GEO Content Auditor analyzes a URL (or raw text) and evaluates its probability of being cited by ChatGPT, Perplexity or Google AI Overviews. Since July 2026, every URL audit also displays a weighted GEO score /100 broken down into 5 axes, based on thresholds observed in large-scale AI citation studies. Each axis is measured through concrete signals extracted from the page's HTML — no subjective judgment.

The 5 axes and their weights

• Citability (25%) — the core of the score: self-contained answer blocks of 134 to 167 words (the format most quoted by LLMs), a direct answer within the first 60 words of each H2 section, concrete data density, and freshness (content under 3 months old is cited ~3× more; beyond 12 months it loses citation eligibility).
• Structural readability (20%) — single H1, 3 to 15 H2s, bullet lists, short paragraphs: the hierarchy LLMs rely on to extract.
• Authority & brand (20%) — identified author (byline or Schema), outbound links to sources, Article schema, FAQ section.
• Technical accessibility (20%) — at least 300 words readable without JavaScript, canonical present, no noindex, meta description, JSON-LD.
• Multi-modality (15%) — images with descriptive alt attributes, embedded video, tables: AI answer engines increasingly value multi-format content.

Signals displayed under the score

Under the 5 gauges, a summary line shows the raw measurements: number of 134-167 word answer blocks detected, proportion of H2 sections answering within the first 60 words, content age (from article:published_time or Schema datePublished/dateModified), and the ratio of images with alt. Recommendations generated by these new signals (e.g. “no self-contained answer block”, “content is 14 months old”) are added to the usual list, sorted by priority.