Technical Readability Engineering: Schema.org Is P0, llms.txt Is Not a Silver Bullet

Readability is the precondition of source operation

Many brands treat GEO as "copywriting", ignoring a prerequisite: if the engine cannot read or parse your content, all prior effort is zero. The white paper calls this layer "technical readability engineering".

This article excerpts Chapter 2 (2.7–2.10), clarifying the priority of tech standards: what is mandatory (P0), what should be done (P1), what is only nice-to-have (P2).

Excerpt from the white paper

Structured data Schema.org (highest priority): key types MedicalWebPage, MedicalCondition, Drug, MedicalProcedure, Physician, FAQPage, HealthTopicContent. Required properties: author, reviewedBy, lastReviewed, medicalAudience, citation.

llms.txt, proposed by Jeremy Howard in 2024-09, is a community proposal, not a standard. Google explicitly does not support it (John Mueller compared it to the keywords meta tag); SE Ranking found zero material effect across 300k domains; Ahrefs says no major LLM adopts it. It can supplement a content map but is not a reliable optimization dependency.

SSR vs CSR: server-side rendering ensures crawlers get the body directly; client-side rendering (SPA) loads dynamically and is easily missed. Medical content pages must be SSR or prerendered, or they are invisible to generative engines.

Tech priority order: P0 — Schema.org medical structuring, SSR/prerender; P1 — robots.txt allow (GPTBot/Bytespider/PerplexityBot/ClaudeBot/Google-Extended), IndexNow push; P2 — llms.txt (nice-to-have, not core).

Deep interpretation: priority mismatch is most brands' pitfall

First, P0 must come first, and neither is optional. Schema.org tells the model "who to trust" — author/reviewedBy/lastReviewed/citation feed E-E-A-T signals; SSR tells the model "it can read". Common errors: missing reviewedBy, relative @id, non-ISO dates, empty sameAs, markup inconsistent with visible content (can trigger manual penalties). Validation chain: validator.schema.org → Rich Results Test → Search Console.

Second, the llms.txt "silver bullet" myth must be broken. Many vendors sell it as a GEO panacea, but Google does not support it, no major LLM adopts it, and 300k-domain tests show zero material effect. Its real place is a content-map supplement at P2. Budgeting on llms.txt while skipping Schema.org and SSR is a textbook priority mismatch.

Third, robots.txt allow is the overlooked P1. Many sites default-block GPTBot/ClaudeBot/PerplexityBot for security, actively excluding themselves from AI answers. Meanwhile IndexNow pushes updates immediately after guideline/consensus changes, preserving freshness — critical when GEO content half-life is only ~13 weeks (to be verified).

Fourth, CSR/SPA is a hard wound for medical pages. Many H5/mini-program landing pages render client-side; crawlers and models get no first-screen body. The white paper observes only ~23% of pure-JS and ~7% of PDF content is effectively parsed by AI (to be verified). Medical pages must be SSR or prerendered.

Nine Health's checklist: ①SSR/static for all medical pages; ②per-page MedicalWebPage + FAQPage Schema (with reviewedBy/lastReviewed); ③robots allow compliant AI UAs; ④IndexNow on guideline updates; ⑤llms.txt as supplement not main. Solve readability first, then source operation.

Get the full white paper

This article excerpts chapters from the Nine Health white paper "Compliance as Growth: GEO Compliance & Growth in Healthcare" (2026 edition). The full edition covers the GEO tech stack, the compliance foundation, source engineering, the E-E-A-T trust system, and the eight capability-building modules, with a compliance self-check list, a three-color risk matrix, and a phased roadmap.

To request the full white paper or a one-on-one brand GEO diagnostic, contact Nine Health: phone 18126259547 / WeCom (see "Contact Us" on the site) / email jianyi@nine-health.com. Our team responds within 24 hours.