Content Playbook

The writing-side companion to Pillar 3. No technical fix earns citations on its own — authority, structure, and freshness are what surface a brand in answer engines.

Per-page structure (mandatory on every new post)

1. One topic per page. Two topics = two posts.
2. One strong H1. Matches title-tag intent.
3. Quick-answer paragraph. 40–80 words, first paragraph, direct answer.
4. One-paragraph summary per major section.
5. Semantic heading hierarchy. H2 → H3, never skip levels.
6. Define terms on first use.
7. Inverted pyramid. Answer first, context second.

Block palette

Block	Use for
Definition	”X is…” — canonical entity statement
Key facts	Bulleted, date-stamped
Pros / cons	Side-by-side tradeoffs
Steps	Numbered, verb-first how-to
Comparison	Table, us vs alternative
Pricing summary	Plans, prices, inclusions
Eligibility	”You qualify if…”
FAQ	Q&A pairs — highest citation density
Troubleshooting	Problem → diagnosis → fix
Changelog	What’s new, when — freshness signal

E-E-A-T

• Bylines on every article — auto-wired to JSON-LD Person schema.
• Author bio block — credentials, LinkedIn, sameAs + knowsAbout.
• About Us + Team pages — Organization + Person schema.
• Primary source citations — government, academic, registry data. Not blog-citing-blog.
• Methodology shown on benchmarks and comparisons.

Deprecated content

• CMS field: replacement_url.
• When set: noindex, follow, visible banner, excluded from llms-full.txt, Link: rel="successor-version".
• Monthly scan flags content > 12 months untouched.

Suppress from llms-full.txt

Tag archives, paginated archives, search results, author archives, attachment pages, noindex pages, pages under 200 words.

Open questions for v1.2 — content anti-patterns

v1.1 does not currently call out content anti-patterns explicitly. Expert research surfaces four worth tracking for the next revision:

1. Scaled bot content — thousands of automated pages targeting fan-out variations. Falls under Google’s scaled-content abuse spam policy (_sources/google-ai-optimization-guide.md.txt).
2. Inauthentic mentions at scale — buying aged Reddit accounts to seed fake recommendations. On Lily Ray’s anti-pattern list (_sources/expert-lily-ray.md).
3. Hidden-text instructions for LLMs — “ignore previous instructions” or similar prompt-injection bait in invisible HTML. Brand-risk regardless of citation impact.
4. Writing for bots, not humans — content that scores well on AI-friendliness audits but fails the reader-satisfaction test.

Lily Ray’s framing: “Many such GEO grifters were using this opportunity to simply repackage core SEO approaches using a different name.” Strong fundamentals win the RAG layer; anti-patterns get penalised.