Building Brand Entities: Knowledge Graph Optimization

Lesson Overview

Entities are not a series of isolated pages, but a network of relationships. When there are no clear relationships among a brand, products, founder, categories, comparison pages, and case studies, it becomes hard for search engines and AI to reliably understand “who is who, what belongs to what, what is comparable to what, and which is the main entity.”

Google’s Knowledge Graph Search API itself organizes results through entity nodes and schema.org types; the Schema.org data model also explicitly supports connecting pages and entities using @id, sameAs, mainEntityOfPage, and similar properties (Per: Google, Schema.org). This lesson focuses on how to build a stable knowledge graph for a brand and use these key properties correctly.

Core Concepts

What Is a Knowledge Graph

A knowledge graph is not an “encyclopedia entry,” but a network of “entities + attributes + relationships,” where nodes are entities and edges are relationships. Building a brand knowledge graph essentially means clearly mapping out the objects related to the brand and the relationships among them, so that the machine can traverse from object to object along the edges.

The Minimal Viable Loop of a Brand Knowledge Graph

A company should establish at least the following relationships to form a minimal loop that machines can reliably understand:

• Organization → Brand
• Brand → Product / Service
• Organization → Founder / Team
• Product → Category
• Product → Use Case
• Product → Comparison Entity
• Article → mainEntity
• FAQ / Guide → related Entity

Main Entities and Subordinate Entities

Before modeling, you must first answer several questions: Does the homepage describe the organization or the brand? Is the main entity of a product page a Product or a Service? Does a comparison page have one main entity or several? If you do not think these through first, your Schema will easily become muddled later, with multiple objects competing for the “lead role” on a single page.

The Roles of @id, url, sameAs, and mainEntityOfPage

These four properties are the skeleton that connects pages and entities, and their definitions must be used precisely (Per: Schema.org):

• url: better suited to indicating the object’s official or authoritative page
• sameAs: used to point to well-known pages that indirectly yet unambiguously establish the object’s identity
• mainEntityOfPage: used to indicate which entity the page primarily describes
• @id: serves as a canonical URI for referencing an entity across pages in JSON-LD, and is the preferred way to reference an entity across pages

A minimal brand entity can be expressed like this:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "@id": "https://example.com/#organization",
  "name": "Example Brand",
  "url": "https://example.com/",
  "sameAs": [
    "https://www.linkedin.com/company/example",
    "https://www.wikidata.org/wiki/Q000000"
  ]
}

By reusing the same @id across pages, your product pages, article pages, and about page can all point back to the same stable brand object, rather than each creating a new object.

Three Methods for Modeling Entity Relationships

Different types of sites suit different modeling emphases. We recommend mastering three models:

• Organization-centric: suited to corporate websites
• Product-centric: suited to SaaS / e-commerce / DTC
• Topic-centric: suited to media / content sites / B2B education sites

How an Entity Graph Serves GEO

AI answers are not generated by scraping a single page word for word; they rely more on a synthesized understanding across multiple sources, pages, and entity relationships. As a result, “visibility” in GEO is often built on whether your brand and its related entities are clearly organized, repeatedly confirmed, and consistently present across pages. A stable knowledge graph is precisely the source of that consistency.

Exercise

Draw a v1 knowledge graph relationship diagram for a brand, containing at least: 1 organization entity, 1 brand entity, 5 product or service entities, 10 topic entities, 5 person / case / customer / comparison entities, and 30 relationship edges. You can complete this in Miro, FigJam, or Excel.

Deliverables

• Knowledge graph relationship diagram v1
• Entity relationship dictionary
• Main-entity and subordinate-entity definition document