Entity Confidence Score: Measuring Consistency, Coverage and Disambiguation
Entity Confidence is a SiteNexis-derived 0–100 diagnostic of entity consistency, coverage and disambiguation. It helps identify representation gaps without claiming to measure an AI provider's private trust probability.
Entity Confidence is not a proxy for brand awareness or domain authority. SiteNexis uses it as a derived diagnostic of how consistently an entity is represented in the inspected content. The canonical score combines three scored dimensions — consistency, coverage, and disambiguation — while detection is a useful inspection question rather than an independently weighted formula term. Each area maps to a distinct remediation path.
Dimension 1: Entity Detection Rate
Detection is the first inspection question: is the primary entity explicitly named in the relevant body content, rather than only in schema? A name confined to an About page may be poorly distributed. SiteNexis reviews this coverage as evidence for the entity model, but detection is not a separate weighted output in the canonical score.
Dimension 2: Entity Consistency Score
Consistency measures whether entity attributes agree across all three sources: schema markup, body text, and metadata (title, description, OG tags). The attributes checked: entity name (must match exactly), entity type (must use the same schema type), founding date or established year (must be identical or absent), geographic information (city, country, region — must be consistent), and primary description (core brand description must not contradict itself). One inconsistency in the name attribute is enough to significantly reduce the consistency score.
▲The most common consistency failure in audits: the schema Organisation name uses the full legal entity name ("Acme Corporation Ltd") while body text consistently uses the trading name ("Acme"). These are the same entity, but AI systems may model them as distinct entities. Pick one name and use it everywhere in schema markup.
Dimension 3: Entity Coverage Score
Coverage scores attribute depth in the inspected representation: name, type, description, dates, location, industry, products or services, leadership, and external validation where available. Sparse coverage is a SiteNexis finding that can make downstream analysis less informative; it is not proof of shallow provider knowledge or fewer recommendations.
Dimension 4: Disambiguation Score
Disambiguation measures how clearly your entity is distinguished from other entities with similar names, types, or descriptions. A company named "Nexis" has a disambiguation challenge because multiple entities share that name. Disambiguation is achieved through: unique identifiers (company registration numbers, DUNS numbers in schema), geographic specificity (operating in Bristol, UK — not just "UK"), explicit type clarification (SaaS platform vs. consultancy), and external disambiguation links (Wikipedia article that specifically refers to your entity, not a disambiguation page).
The Composite and Its Uses
The canonical SiteNexis calculation is Entity Confidence Score = (Consistency × 0.30) + (Coverage × 0.35) + (Disambiguation × 0.35). The result feeds SiteNexis AI Visibility composites; it is not a provider trust probability. Production paths may expose a flatter heuristic representation, so interpret the score with its audit context rather than treating it as an externally standardized metric.