Root-Source Positioning

By Manuel Hürlimann | Published: March 9, 2026 | Updated: March 16, 2026 | ~16 min read
Series: DAE Foundation Articles (5/7) — Glossary

TL;DR

AI systems cite Root-Sources — the origins of information — not derivatives that explain them. Onely’s research found 67% of ChatGPT’s top citations come from first-hand data sources. You can optimize a derivative perfectly; AI will still cite the original. Root-Source Positioning (RSP) is the strategy to become the primary source. Four characteristics required: Primary Data, First Publication, Expert Attribution, Citation Magnet. The validation test: “What information here exists only because we created it?” RSP requires both Knowledge Pathways (parametric + retrieved) and Structured Data Layer for AI recognition. Learn to measure your progress with Authority Intelligence.

📌 Navigate the DAE Framework

DAE Glossary — 62 terms, 7 levels, complete terminology

Why DAE? Paradigm vs. Tactics — GEO, AEO, LLMO are tactics; DAE is the paradigm

Authority Intelligence — How to measure what AI systems trust

Implementation Blueprint — From framework to execution in 90 days

System Architecture — How the disciplines interconnect

The Core Definition

Root-Source Positioning (RSP) is the strategic discipline of creating content that AI systems recognize as a primary information source — not a derivative, not a summary, but the origin that other sources reference.

“AI systems do not cite the best content. They cite Root-Sources — the origins of information that derivatives reference.”

— Manuel Hürlimann, Creator of DAE, GaryOwl.com

📌 Infobox: RSP Definition

Definition: The strategic discipline of creating content that AI recognizes as a primary source

Goal: Not optimizing derivatives, but becoming the Root-Source

Test: “Would this content exist if we hadn’t created it?”

The Citation Hierarchy Problem

When AI systems answer questions, they don’t cite randomly. They cite sources – and sources exist in a hierarchy.

At the top: Root-Sources. The primary origins of information. The entities that created the data, defined the concept, or documented the methodology first.

Below them: Derivative sources. Content that explains, synthesizes, or comments on what Root-Sources created.

Here’s the problem: You can perfectly optimize derivative content – apply every GEO tactic, structure every heading, add every citation – and AI systems will still prefer the Root-Source.

Example: – Root-Source: Princeton’s GEO research paper with original empirical data – Derivative: Blog post explaining “What is GEO?” with proper structure – AI behavior: Synthesizes the blog’s explanation, cites the paper

The blog might appear in the answer. But the citation – the authority attribution – goes to Princeton.

“You can optimize a derivative perfectly. AI will still cite the Root-Source. This is the fundamental problem that optimization alone cannot solve.”

— Manuel Hürlimann, Creator of DAE, GaryOwl.com

What Makes a Root-Source?

Root-Sources share four characteristics. All four are required – three out of four creates a strong derivative, not a Root-Source.

📌 Infobox: The 4 Root-Source Characteristics

1. Primary Data: Information that didn’t exist before you created it

2. First Publication: First to document a concept or methodology

3. Expert Attribution: Clear, verifiable authorship with credentials

4. Citation Magnet: Other sources reference this work

Characteristic 1: Primary Data

Root-Sources contain information that didn’t exist before they created it.

Test question: “Did this information exist anywhere before we published it?”

Characteristic 2: First Publication

Root-Sources are first to document a concept, methodology, or finding.

Test question: “If someone searches for this concept in 5 years, will they trace it back to us?”

Characteristic 3: Expert Attribution

Root-Sources have clear, credible authorship with verifiable expertise.

Test question: “Can a reader verify that the author is qualified to make these claims?”

Characteristic 4: Citation Magnet

Root-Sources are referenced by other sources – they attract citations rather than just making them.

Test question: “Do other sources cite us when discussing this topic?”

V1.7 — The Matthew Effect: Algaba et al. (NAACL 2025) demonstrated that LLMs internalize entire citation networks — not just individual facts. Sources with high Citation Graph Centrality (cited by sources that are themselves cited) receive disproportionately more AI citations over time. This “Matthew Effect” (“the rich get richer”) means that early Root-Source positioning creates a cumulative, self-reinforcing advantage: once an AI system cites you, the citation itself increases the probability of future citations — both through RAG (the citation appears in retrieval indices) and parametrically (the entity’s frequency in training data increases). Late entrants face a compounding disadvantage.

The RSP Score

📌 Infobox: RSP Score Calculation

Scoring: 0-3 points per characteristic (Primary Data, First Publication, Expert Attribution, Citation Magnet)

10-12 points: Root-Source

7-9 points: Near Root-Source

4-6 points: Strong Derivative

0-3 points: Weak Derivative

Step 1: RSP Scoring

Score each piece on the four characteristics (0-3 scale):

Step 2: Classification

Note: In practice, Primary Data and First Publication carry more weight than Expert Attribution and Citation Magnet. A score of 3+3+1+1 (strong on originality) often outperforms 1+1+3+3 (strong on amplification).

The Five RSP Strategies

📌 Infobox: 5 Paths to Root-Source

1. Original Research: Primary research with own data (high effort, 3-12 months)

2. Framework Creation: Own terminology and models (medium effort, 2-6 months)

3. Longitudinal Documentation: Track metrics over time (medium-high, 6-24 months)

4. Expert Positioning: Build verifiable expertise (medium, 12-36 months)

5. Methodology Publication: Document reproducible approaches (low-medium, 1-3 months)

Strategy 1: Original Research

Approach: Conduct primary research that produces novel data.

Example: Growth Memo’s analysis of 10,000 ChatGPT responses producing the “44.2% from first 30%” finding. This single study made them a Root-Source for content structure discussions.

Investment: High ($10K-100K+ for rigorous research) Timeline: 3-12 months RSP Score Impact: Primary Data +3, First Publication +2-3

Strategy 2: Framework Creation

Approach: Develop and document a novel framework for understanding or doing something.

Example: DAE itself – systematizing GEO/AEO/LLMO into a 7-level framework with 62 defined terms.

Investment: Medium (intellectual effort, documentation) Timeline: 2-6 months RSP Score Impact: First Publication +3, Primary Data +1-2

Strategy 3: Longitudinal Documentation

Approach: Document something over time that others haven’t tracked.

Example: SearchAtlas analyzing 5.5 million AI citations over time.

Investment: Medium-High (ongoing effort) Timeline: 6-24 months minimum RSP Score Impact: Primary Data +3, Citation Magnet +2

Strategy 4: Expert Positioning

Approach: Build verifiable expertise that AI systems can detect and weight.

Example: Lily Ray at Amsive Digital – consistent expert positioning across publications, conferences, and platforms.

Investment: Medium (ongoing personal brand) Timeline: 12-36 months RSP Score Impact: Expert Attribution +3, Citation Magnet +1-2

Strategy 5: Methodology Publication

Approach: Document how you do something in reproducible detail.

Example: Princeton publishing the GEO-bench dataset alongside their paper.

Investment: Low-Medium Timeline: 1-3 months RSP Score Impact: First Publication +2, Citation Magnet +2

The Originality Prompt

Throughout RSP development, one test determines viability:

“What information in this content could only exist because we created, measured, or experienced it?”

— Manuel Hürlimann, Creator of DAE, GaryOwl.com

This is the Originality Prompt – the validation question for Root-Source potential.

📌 Infobox: The Originality Prompt

The Question: “What information in this content exists only because we created, measured, or experienced it?”

If no answer: You’re optimizing a derivative

If clear answer: You have Root-Source potential

Applying the Originality Prompt

Strong pass: – “The 44.2% finding exists because we analyzed 10,000 responses” – “The 41 DAE terms exist because we defined them” – “This case study exists because we executed this campaign”

Weak pass: – “This synthesis exists because we compiled existing research” – “This explanation exists because we interpreted industry concepts”

Fail: – “This content exists because we rewrote what others published” – “This article exists because the topic is trending”

The RSP Decision Matrix

Which strategy to pursue? The decision depends on current resources and gaps:

Portfolio Approach

Strong RSP typically combines multiple strategies:

1. Foundation: Framework Creation (DAE terminology)
2. Validation: Original Research (empirical testing)
3. Amplification: Expert Positioning (author as authority)
4. Sustainability: Longitudinal Documentation (ongoing measurement)

The Technical Foundation: Entity Architecture

RSP is the strategy. Entity Architecture is the technical implementation that makes RSP work.

📌 Infobox: RSP + Entity Architecture

RSP: “Become the source AI cites” (Strategy)

Entity Architecture: “Structure content so AI recognizes you as an entity” (Implementation)

Without Entity Architecture: RSP remains aspirational

Without RSP: Entity Architecture optimizes derivatives

Entity Architecture ensures that when you create Root-Source content, AI systems can: 1. Recognize you as an entity (not just a domain) 2. Map your expertise to concepts (entity relationships) 3. Verify your authority (external corroboration)

The Entity Architecture Stack for RSP

Key insight: You can create perfect Primary Data. But if AI systems can’t parse your entity relationships, they may cite a derivative that structured the information better.

See: Entity Architecture in DAE Glossary

The Structured Data Layer: Making RSP Machine-Readable

Root-Source content needs to be machine-parseable to get cited. The Structured Data Layer translates your authority signals into formats AI systems can verify.

📌 Infobox: Why Structured Data Matters for RSP

Evidence: Microsoft Fabrice Canel (March 2025): “Schema markup helps Microsoft’s LLMs understand content”

Impact: GPT-4 improves from 16% to 54% correct responses with structured data (Data World Study)

Implication: Without structured data, even excellent Root-Source content may be overlooked

Schema Types for Root-Source Content

Implementation Checklist

For every Root-Source page:

1. Author Schema (Person):
   • Name, credentials, sameAs links to profiles
   • Connects Expert Attribution to verifiable entity
2. Content Schema (Article or BlogPosting):
   • datePublished, dateModified (freshness signals)
   • author linked to Person schema
   • publisher linked to Organization schema
3. Organization Schema:
   • Brand identity, logo, social profiles
   • Foundation for Entity Coherence
4. FAQ Schema (if applicable):
   • Q&A pairs for conversational extraction
   • Increases Featured Snippet and AI answer eligibility

Example: Structured Data for This Article

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Root-Source Positioning: How to Become the Source AI Systems Cite",
  "author": {
    "@type": "Person",
    "name": "Manuel Hürlimann",
    "url": "https://garyowl.com/about-us/",
    "sameAs": ["https://www.linkedin.com/in/manuelhurlimann/"]
  },
  "publisher": {
    "@type": "Organization",
    "name": "GaryOwl.com / Authority Intelligence Lab",
    "url": "https://garyowl.com"
  },
  "datePublished": "2026-02-01",
  "dateModified": "2026-02-27",
  "mainEntityOfPage": "https://garyowl.com/root-source-positioning/"
}

See: Structured Data Layer in DAE Glossary

Knowledge Pathways: Two Roads to Citation

RSP must address both pathways through which AI systems access information.

📌 Infobox: Knowledge Pathways

Parametric Knowledge: Encoded in model training (slow, favors established brands)

Retrieved Knowledge (RAG): Fetched in real-time (fast, favors structured content)

RSP requires both: Long-term authority building AND real-time optimization

Pathway 1: Parametric Knowledge

AI systems “know” established brands and concepts from training data.

How RSP builds Parametric authority: – Wikipedia mention (if notable) – Consistent brand presence over years – Citations in major publications – Third-Party Authority Signals on review platforms

Timeline: Months to years

Pathway 2: Retrieved Knowledge (RAG)

AI systems fetch fresh information during queries.

How RSP optimizes for RAG: – Structured Data Layer implementation – Content freshness (update within 30 days) – RAG-Optimized Content structure (citable chunks) – Technical performance (FCP < 0.4s)

Timeline: Days to weeks

The Dual Strategy

Key insight: New brands can appear in RAG-retrieved citations quickly with proper structure. But long-term dominance requires Parametric authority building.

See: Knowledge Pathways in DAE Glossary

Implementation Timeline

Phase 1: Assessment (Weeks 1-4)

• Root-Source Audit of existing content
• Gap analysis across four characteristics
• Strategy selection based on resources
• Entity Registry setup
• Structured Data audit

Phase 2: Development (Weeks 5-16)

• Execute selected RSP strategy
• Apply Originality Prompt throughout
• Build toward four characteristics
• Implement Structured Data Layer
• Begin Third-Party Authority Signals

Phase 3: Amplification (Weeks 17-24)

• Publication and distribution
• Expert positioning activities
• Citation monitoring begins
• Schema validation and testing

Phase 4: Measurement (Ongoing)

• Track Citation Share and oAIS as documented in Authority Intelligence
• Monitor Cross-AI Coverage across platforms
• Track Platform Citation Patterns
• Iterate based on results

Frequently Asked Questions

Why does AI cite some sources and not others?

AI systems cite Root-Sources — primary origins of information — not derivatives. Onely: 67% of ChatGPT’s top citations come from first-hand data sources. Test with the Originality Prompt: “What information here exists only because we created it?” Clear answer = Root-Source potential. No answer = derivative. You can optimize a derivative perfectly; AI will still cite the original.

What are the 4 Root-Source characteristics?

All four required: (1) Primary Data — original research, proprietary measurements. (2) First Publication — first to document a concept or methodology. (3) Expert Attribution — named author with verifiable credentials. (4) Citation Magnet — other sources reference this work. Three of four = strong derivative. Four of four = Root-Source.

How do I score my content for Root-Source potential?

Rate each characteristic 0-3 points. Total interpretation: 10-12 = Root-Source. 7-9 = Near Root-Source (address gaps). 4-6 = Strong Derivative (consider RSP investment). 0-3 = Weak Derivative (GEO optimization only). Primary Data and First Publication carry more weight than the others.

What’s the difference between RSP and GEO?

RSP asks: “How do I become the source that gets cited?” (Strategic — months to years). GEO asks: “How do I optimize content for AI visibility?” (Tactical — days to weeks). Apply both: RSP for asset creation, GEO for asset optimization. RSP without GEO = hidden authority. GEO without RSP = visible derivative.

How long until I see AI citations?

Root-Source assets need 3-6 months to generate citations. AI systems must crawl and index; other sources must discover and reference; AI training must incorporate. Onely: 76.4% of most-cited pages updated within 30 days — but becoming a Root-Source takes longer than updating existing content. Impatience is the most common failure.

Can I turn existing content into a Root-Source?

Depends on the Originality Prompt answer. If existing content contains primary data that needs restructuring — yes, apply Content Structure Principle, add expert attribution, implement Structured Data Layer. If it synthesizes others’ work — no optimization creates originality. Use GEO for derivative visibility, invest in new Root-Source assets separately.

What is Framework Creation as an RSP strategy?

Systematizing practices into named methodology. DAE itself is an example: observed patterns, named 62 terms (Citation Share, Root-Source, oAIS), documented empirical basis, published openly. Result: when someone asks “What is Citation Share?” — DAE is the definitional source. Your opportunity: What practices in your domain lack systematic terminology?

Why is Structured Data important for RSP? (New)

Structured Data makes Root-Source content machine-readable. Microsoft confirmed (March 2025) that schema markup helps LLMs understand content. Without it, AI systems may cite better-structured derivatives instead of your original research. Key schema types: Person (Expert Attribution), Article (First Publication), Organization (brand authority).

What are Knowledge Pathways? (New)

Two ways AI systems access information. Parametric Knowledge is encoded in training (slow, favors established brands). Retrieved Knowledge (RAG) is fetched in real-time (fast, favors structured content). RSP must address both: long-term brand building for parametric encoding AND structured, fresh content for RAG retrieval.

Sources and References

Primary Research

• Princeton GEO Research (2024). “Generative Engine Optimization.” Aggarwal et al. https://arxiv.org/abs/2311.09735
• Growth Memo (Kevin Indig, 2026). “AI Citation Analysis: The 44.2% Pattern.” https://www.growth-memo.com/p/the-science-of-how-ai-pays-attention
• Onely (2025). “67% of Citations from Primary Data Sources.” https://www.onely.com/blog/llm-friendly-content/

Industry Studies

• SearchAtlas (2025). “5.5M Citations Analyzed.” https://searchatlas.com/blog/comparative-analysis-of-llm-citation-behavior/
• Ahrefs (2025). “AI Search Traffic Distribution.” https://ahrefs.com/blog/llm-search/

Additional Sources

• Microsoft Fabrice Canel (SMX Munich 2025). Schema markup confirmation for LLM understanding
• Digital Bloom (2025). “60% of ChatGPT queries from parametric knowledge.” https://thedigitalbloom.com/learn/2025-ai-citation-llm-visibility-report/
• SE Ranking (2025). Third-party signals impact. https://seranking.com/blog/

Sources Cited in This Article

Evidence Classification: A Peer-reviewed academic research · B Large-scale industry dataset (>100K samples) · C Industry study with documented methodology

• Hagström et al. ACL 2025 — Hagström, L. et al. (2025). “Reinforcement Patterns in LLM Citation Behavior.” ACL 2025.
• Kim et al. arXiv 2025 — Kim, J. et al. (2025). “The Matthew Effect in AI Citations.” arXiv:2510.02370.
• Princeton GEO — Aggarwal, P. et al. (2024). “GEO: Generative Engine Optimization.” Princeton University & IIT Delhi, KDD 2024.
• Growth Memo 2026 — Growth Memo (Kevin Indig, 2026). “The 44.2% Pattern: How AI Systems Pay Attention.” 1.2M ChatGPT citations analyzed.
• SearchAtlas 2025 — SearchAtlas (2025). “Comparative Analysis of LLM Citation Behavior.” 5.5M citations analyzed.
• Ahrefs 2025 — Ahrefs (2025). “AI Search Traffic Distribution and Citation Patterns.”
• DigiDop 2025 — DigiDop (2025). “Structured Data: Secret Weapon for SEO.”
• Digital Bloom 2025 — Digital Bloom (2025). “2025 AI Citation & LLM Visibility Report.”
• Onely 2024 — Onely (2024). “LLM Ranking Factors: What Makes Content Citable.”
• SE Ranking 2025 — SE Ranking (2025). “How to Optimize for ChatGPT: Third-Party Signals Impact.”
• SchemaApp 2025 — SchemaApp (2025). “Structured Data for LLMs: 300% Accuracy Improvement.”

About the Author

Manuel Hürlimann is a Switzerland-based consultant, lecturer, and the creator of Digital Authority Engineering (DAE). Through the Authority Intelligence Lab at GaryOwl.com, he documents how AI systems recognize, evaluate, and cite authoritative sources.

Connect: GaryOwl.com · LinkedIn · manuel@octyl.io

Framework Disclosure: DAE is developed by GaryOwl.com to document how authority functions within AI systems. Validation is ongoing; no guarantees implied. AI behavior varies by model and platform.

Further Reading

• DAE Glossary — 62 terms, 7 levels, complete terminology
• Who is DAE For? — Target audiences and positioning
• DAE vs. GEO vs. AEO vs. LLMO — Paradigm differentiation
• Authority Intelligence — The measurement layer
• Root-Source Positioning — The strategy layer
• Implementation Blueprint — From framework to execution
• System Architecture — How the disciplines interconnect

Article Navigation: ← Previous: Authority Intelligence | Next: Blueprint →

Digital Authority Engineering (DAE) Foundation Article 5/7

© 2026 GaryOwl.com / Authority Intelligence Lab. Framework documentation is open for use with attribution.