Skip to content

feat: add SEO foundation dimension (scoring model v1.4)#19

Merged
ryanportfolio merged 1 commit into
mainfrom
claude/aeo-vs-geo-6af996
Jul 20, 2026
Merged

feat: add SEO foundation dimension (scoring model v1.4)#19
ryanportfolio merged 1 commit into
mainfrom
claude/aeo-vs-geo-6af996

Conversation

@ryanportfolio

Copy link
Copy Markdown
Owner

What

Adds a ninth weighted dimension, SEO foundation (weight: low), and bumps the scoring model to v1.4.0.

Deliberately narrow: presence/length of title, description and canonical are already scored under topical focus & metadata. This dimension covers the cross-cutting failures single-page checks cannot see:

  • Canonical self-consistency: a canonical pointing away from the audited page means every consumer that honors it attributes the content to a different URL (www/slash/scheme variants are tolerated).
  • Title and meta-description uniqueness across the fetched pages: duplicated metadata makes distinct pages indistinguishable to a retriever.

Also folds an unscored heading-hierarchy advisory (skipped levels, e.g. H2 to H4) into answer-readiness.

Deliberately excluded

Core Web Vitals, backlinks, mobile rendering, hreflang: classic SEO with near-zero GEO signal, and out of mission for a citation-readiness audit.

Scoring model

  • Version 1.3.0 to 1.4.0 (scores comparable only within a version).
  • Golden pins recalibrated: the three pinned overalls are unchanged (99/33/25); each context gains a seoFoundation score (100/30/30).
  • The good-article fixture's canonical was corrected to be self-referencing (it previously pointed at a foreign host, which the new check rightly flags).

Docs

README and site copy (index/about/audit app shell) updated from eight to nine dimensions.

Verification

  • npx tsc --noEmit: clean
  • npx vitest run: 161/161 pass

🤖 Generated with Claude Code

New ninth weighted dimension, seoFoundation (weight low): canonical
self-consistency on the audited page and title/meta-description
uniqueness across the fetched pages. Presence/length of title,
description and canonical stay under topical focus; this covers the
cross-cutting failures those single-page checks cannot see.

Also folds an unscored heading-hierarchy advisory (skipped levels)
into answer-readiness.

Scoring model bumped to 1.4.0; golden pins recalibrated (overall
scores for the three pinned contexts are unchanged: 99/33/25).
Fixture canonical corrected to be self-referencing. README and site
copy updated from eight to nine dimensions.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
@ryanportfolio
ryanportfolio merged commit d64c137 into main Jul 20, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant