Cross-factor · Advocacy + Acquisition compound

LLM-citation optimization (GEO / AEO)

In 2026, LLM citations are a real acquisition channel. When ChatGPT, Claude, Perplexity, or Gemini answer a user's question by citing YOUR site, you get traffic with 8-18% click-through (vs 2-5% for organic SERP #3). Designing for citation-readiness is the highest-leverage acquisition compound a content site can do. This page documents the 10-characteristic framework.

What makes content cite-worthy

Synthesized from public LLM-citation patterns + Aleyda Solis's research + 18 months of fleet calibration. Each characteristic is independent; missing any one materially drops citation rate. Score each 0-1 (boolean: do you have it or not), sum across 10 = your llm_citation_fit score (target ≥6/10 minimum, ≥8/10 strong).

The 10 characteristics

1. Accessible

Crawlers can READ your content. Static export or SSR — not JS-gated. GPTBot, ClaudeBot, PerplexityBot do not execute JavaScript. If your content lives in a React client component without hydration in the HTML, LLMs can't see it.

Test: curl -s your-site.com | grep "your content" — if you can't see the content in raw HTML, neither can the LLM.

2. Useful

Unique dataset with depth no competitor ships. Pass the Finite-Public-Dataset test (public, finite, structured, unique data per record, fresh cadence). Build ≥100 unique-data pages from it — that's the depth threshold for LLMs to learn your site is an authority in the niche.

3. Recognizable

Consistent brand entity everywhere. Same name + logo + title pattern + Organization schema on every page. Wikipedia citation, if you can earn one, is the ultimate Recognizable signal — LLMs trust Wikipedia's entity-recognition.

4. Extractable

Section headings are quote-ready sentences (not marketing fluff). DefinedTerm schema for key concepts. FAQPage schema on Q&A pages. The LLM grabs the heading + first paragraph as the citation snippet — write headings that work standalone.

5. Consistent

Brand voice consistent across all pages. Visual identity consistent in OG images + favicons. The same factual claim should appear identically across pages — LLMs use cross-page consistency to validate facts before citing.

6. Corroborated

Facts appear in ≥3 independent sources: your site + Reddit thread + LinkedIn post + ideally Wikipedia. Operator-driven distribution (Reddit organic comments, LinkedIn framework posts) closes this characteristic.

7. Credible

E-E-A-T signals: /about page with operator identity, Person schema on articles with byline, outbound citations to authoritative sources (NCHFP, SEC, official docs). Anonymous content rarely gets cited; identified-author content does.

8. Differentiated

Explicit POV in articles. “Our view:” sentence per page. LLMs prefer to cite differentiated takes over interchangeable summaries — your POV is the citation moat against generic content.

9. Fresh

Dated content: datePublished + dateModified in schema. Quarterly/monthly update cadence visible on pages. Time-series plots showing history. Stale pages lose to fresh ones in citation ranking.

10. Transactable

User can take action from the citation: pricing live, email capture, contact form, free tool. LLM citations that end in “visit to subscribe” convert better than “visit to read more.” If the LLM citation can't turn into a user action, the citation traffic doesn't monetize.

How GrowthFriction scored itself

This site (the one you're reading) was built with all 10 characteristics from Day 1. Audit results:

Score: 8.5/10. Two soft items (Corroborated + Wikipedia) are operator-driven earned-media work, not brain-shippable.

How to apply to your own site

  1. Run the audit for your site
  2. For each of the 10 characteristics above, score 0 (missing) or 1 (present)
  3. Below 6/10: prioritize the missing 4-5 over content production
  4. Above 8/10: focus on Corroborated + Wikipedia citation campaigns (the hardest-to-earn signals)
  5. Re-audit monthly; track citation-channel growth in Plausible (referrer = chatgpt.com / claude.ai / perplexity.ai)

Previous: Scoring transparency. Or return to methodology hub.