Documentation
Assessment Areas
Every Page Scan attempts four independent assessments. Each area produces its own score, findings, and recommendations from bounded evidence.
The four areas
SEO
25%Evaluates whether the page clearly and consistently communicates its subject, audience, and intended search task.
- Search intent, purpose, audience, and task satisfaction
- Semantic alignment across title, description, H1, introduction, and body
- Visible, social, and structured-representation fidelity
- Topical coverage, contextual links, and meaningful media
Page Experience
25%Measures browser delivery, performance, automated accessibility evidence, and technical experience on Mobile and Desktop.
- Mobile 70% and Desktop 30%, shown separately
- Performance 50%, Accessibility 30%, and Best Practices 20% per device
- URL field data, then origin field data, then labelled Lighthouse lab fallback
- LCP, INP, and CLS ratings and continuous contributions without double counting
AI Content
25%Evaluates whether an AI reader can identify, extract, summarize, ground, qualify, and act on the page's answer without inventing trust claims.
- Primary information need and evidence-backed three-bullet summary
- Q&A with explicit abstention and claim/evidence support
- Completeness, qualifications, actionability, and integrity
- Observable credibility, attribution, and maintenance signals
AI Discoverability
25%Evaluates observable technical eligibility and machine-readable discovery for supported AI search and answer products.
- Per-provider discovery access and indexing/snippet controls
- Canonical, applicable hreflang, metadata, Schema.org, and Open Graph
- Main-content access, provenance, and bounded citation resolvability
- A small two-point origin-level llms.txt orientation signal
Page Optimization Score
A Complete Result combines the four scores using the published weights above and rounds the result to the nearest whole number. If any area is unavailable, FutureContent returns a Partial Result: it keeps the available area scores, never reweights them, and does not present an overall Page Optimization Score.
One owner per underlying defect
A versioned Ownership Registry assigns every scored check to one Area and one Defect Family. The same evidence can inform more than one explanation, but it cannot multiply a score penalty. Several page surfaces or semantic artifacts that expose one problem become one consolidated finding.
Lighthouse is Page Experience evidence. Its aggregate SEO category is non-scoring context, and an individual audit explains its owned component without adding a deduction on top of that component score.
Ranges and comparisons
When some applicable checks are unavailable, the Area shows a range: the lower bound applies their maximum possible deductions and the upper bound assumes they pass. Known evidence is never reweighted. Same-version ranges show improvement only when the new range lies entirely above the earlier range; overlap is inconclusive. Different Evaluation Versions are not comparable.
Evidence, not a promise
Scores describe the evidence available to the versioned automated framework. They do not guarantee rankings, traffic, accessibility compliance, inclusion in an AI answer, or a business outcome.
Start with the four-category Page Scan checks overview.
The detailed rules and current checks are published in the Page Experience, SEO, AI Discoverability, and AI Content methodologies.