Immutable methodology · seo-rubric-v1

FutureContent SEO

SEO asks whether the page clearly and consistently communicates what it is about, who it is for, and which search intent it satisfies. It is a semantic content assessment—not a technical crawl or delivery audit.

Pinned identities

Evaluation Version
page-scan-2026-08-18-v16
Manifest
assessment-e79ae98ad2b433fc
Rubric
seo-rubric-v1
Specification
page-scan-v1-specification-v1
Ownership Registry
page-scan-v1-ownership-v1
Provider Registry
page-scan-v1-provider-registry-v2
Weights
page-scan-157-check-scoring-v1
Prompt / schema
page-understanding-prompt-v2 / page-understanding-schema-v2

The stable SEO methodology page points here while this is the active version.

Ownership boundary

SEO evaluates meaning and agreement across the title, meta description, H1, introduction, body, Open Graph representation, structured data, links, and media. It does not score presence, syntax, rendering, delivery, canonical mechanics, crawl eligibility, heading hierarchy, keyword density, character-count folklore, or minimum word counts.

One underlying problem has one scored owner and one Defect Family. For example, a mismatch across title, H1, introduction, body, Open Graph, and structured data becomes one consolidated search-alignment finding—not a separate penalty for every surface. Evidence can be reused, but score effects cannot be multiplied.

Scored defect families

The deterministic rubric owns applicability, severity, and arithmetic. Model output supplies bounded evidence and classifications; it never assigns the score. The configured maximum deductions below were validated by the approved fifty-page V1 benchmark.

seo.description_fidelity
7 points maximum
seo.entity_terminology_consistency
12 points maximum
seo.intent_coverage
18 points maximum
seo.internal_link_usefulness
5 points maximum
seo.media_relevance
2 points maximum
seo.search_intent_clarity
15 points maximum
seo.search_representation_alignment
20 points maximum
seo.social_representation_fidelity
5 points maximum
seo.structured_representation_fidelity
8 points maximum
seo.title_fidelity
8 points maximum

Original contribution and information gain remain visible diagnostics and do not affect this version's score.

SEO defect families (10)

  1. Page purpose, audience, and search intent

    Shared page-understanding artifacts with title, introduction, body, and task evidence.

  2. Search-presentation alignment and fidelity

    Title, description, H1, introduction, dominant body topic, Open Graph, and structured representations.

  3. Title specificity and fidelity

    Title meaning compared with audience, purpose, language, body, and peer boilerplate signals.

  4. Meta-description promise and fulfillment

    Description promise, intended outcome, introduction, and landing-page fulfillment.

  5. Heading and main-content intent coverage

    Inferred search task, headings, main content, fundamental questions, and task-satisfaction artifact.

  6. Topic, terminology, and entity consistency

    Metadata, headings, body, Open Graph, structured data, links, and media labels.

  7. Structured-representation fidelity

    Visible primary entity and valid exposed structured entity values.

  8. Social-representation fidelity

    Open Graph text and image meaning compared with visible page content.

  9. Contextual internal-link usefulness

    Internal-link context, anchor meaning, destination topic, and search-task relationship.

  10. Media contribution to understanding

    Visible media context, labels, captions, and relationship to the search task.

Applicability and evidence

Semantic scoring supports articles/news, documentation, product/service pages, landing pages, and policies/legal pages. If page type, purpose, or applicability is ambiguous, dependent checks are unavailable rather than failed. Evaluator uncertainty is never a page defect.

Evidence is limited to the application-owned normalized page document. Retained excerpts are capped at 240 characters, three references per item, and 12,000 aggregate excerpt characters per scan after deduplication. Copied excerpts expire after 90 days or earlier with the scan or Workspace; source field and location remain explainable.

Scores, ranges, and comparisons

SEO contributes 25% to a complete Page Optimization Score. An unavailable check preserves its maximum possible deduction in a non-reweighted Assessment Score Range. An unavailable Assessment Area withholds the overall score instead of displaying zero or reweighting the other areas.

Scans are comparable only when their exact Evaluation Version matches. A new score range is an improvement only when it lies entirely above the earlier range; overlapping ranges are inconclusive. Different evaluation generations are not comparable.

Limitations and no guarantees

This automated, single-run assessment does not guarantee Search or AI inclusion, citation, recommendation, ranking, traffic, trustworthiness, or any business outcome. It also does not prove that an AI agent can complete login, navigation, submission, recovery, or another end-to-end workflow.