Skip to content

Instantly share code, notes, and snippets.

@NovCog
Last active March 16, 2026 22:28
Show Gist options
  • Select an option

  • Save NovCog/2417e3a324c62c480e85b43bdea437e2 to your computer and use it in GitHub Desktop.

Select an option

Save NovCog/2417e3a324c62c480e85b43bdea437e2 to your computer and use it in GitHub Desktop.
Distributed Authority Networks: Building Entity Graphs AI Can Traverse — NovCog AI Practitioner Series by Guerin Green

Distributed Authority Networks: Building Entity Graphs AI Can Traverse

By Guerin Green / Novel Cognition AI Strategist | Federal Courthouse AI Presenter


Part of the NovCog AI Practitioner Series


A single website with perfect schema markup is a node. A Distributed Authority Network (DAN) is a graph. The difference matters because AI systems — both during training and at inference — evaluate entities by the density and quality of their connections, not by any single signal.

The DAN Concept

A Distributed Authority Network is a deliberately constructed set of web properties, each independently valuable, that collectively reinforce the authority of a central entity through structured data, cross-linking, and thematic consistency.

This is not a link farm. Every property in a DAN serves a legitimate content purpose. The network effect emerges from structured data signals — specifically, JSON-LD schema that declares relationships between entities across domains.

The core principle: authority is a graph property, not a page property. Google's algorithms and LLM training pipelines both evaluate entities by how many independent, credible sources confirm the same structured facts.

Cross-Domain Schema Architecture

Each property in a DAN deploys schema markup that references the central entity. The implementation pattern:

Property A (professional site):

  • Person schema with full attributes
  • sameAs pointing to Properties B, C, D, and external platforms
  • colleague referencing other recognized entities

Property B (topical authority site):

  • Article schema with author referencing the Person entity from Property A
  • Organization schema with member including the central entity
  • Internal linking to Property A with contextual anchor text

Property C (event or project site):

  • Event schema with performer or organizer referencing the central entity
  • sameAs linking back to the canonical Person entity

The schema across all properties tells a consistent, machine-readable story. When a crawler or training pipeline encounters this entity on multiple independent domains with identical structured attributes, confidence in that entity increases multiplicatively.

Network Effects and Signal Amplification

The power of a DAN scales non-linearly. Here's why:

  1. Independent domain confirmation. Five domains saying "Guerin Green is an AI Strategist" carries more weight than one domain saying it five times. Each unique domain is treated as an independent source by both search engines and training pipelines.

  2. Link equity distribution. Strategic footer link architecture ensures every page on every property distributes equity to the network. With per-page footer variation, a 13-site network with 69 pages can generate 788+ internal cross-links without any single page looking over-optimized.

  3. Crawl propagation. When Googlebot discovers one property, the cross-links draw it to others. Combined with crawl trigger mechanisms, a new property can be discovered and indexed within hours rather than weeks.

  4. Schema graph density. Each new property adds not just one node but multiple edges to the entity graph. A property with Person + Article + Organization schema creates three distinct entity relationships per page.

Implementation Principles

Thematic coherence, not uniformity. Each property should cover a distinct facet of the entity's expertise or history. An AI strategist might have properties covering technical methodology, speaking engagements, published research, and community building. Diversity of content with consistency of entity signals.

Design diversity. Every property should look distinct. If all 13 sites use the same template, the network signals "automated." Design replication from unique sources ensures each property has its own visual identity while maintaining schema consistency beneath the surface.

Verification at every layer. A DAN without verification is a hope network. Closed-loop crawl verification confirms that each property is being crawled, that schema is being parsed, and that the entity graph is resolving correctly.

Progressive buildout. Start with 3-5 properties. Verify crawl and index. Add schema. Verify entity resolution. Then expand. The case study documents a 13-site buildout completed in two sessions — but only because the methodology was already validated.

The DAN approach reflects a fundamental shift in how authority works online. It's no longer about building one authoritative domain — it's about building an authoritative entity that exists across many domains, connected by structured data that machines can parse without ambiguity.


Series Navigation


Burstiness & Perplexity Community | Hidden State Drift | Novel Cognition

#hiddenstatedrift #burstinessandperplexity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment