Entity Identification and Relationship Mapping

Lesson Overview

Having built up an understanding of entities and knowledge graphs in earlier lessons, here we return to execution and focus specifically on the two most critical practical actions: “how to inventory” and “how to draw the relationship graph.” The entity inventory answers “what objects do we have,” while relationship mapping answers “what relationships exist among these objects.” Together, the two make up the first half of modeling.

Whether it is an entity list or a relationship graph, the goal is to let search engines and AI reliably recognize and connect objects around “people, places, things,” rather than stopping at string matching (Per: Google, Schema.org).

Core Concepts

Step 1: A Systematic Entity Inventory

Build an entity inventory table around a real brand. We recommend listing 50–200 core entities, and labeling each entity item by item:

• Type (brand / organization / person / product / category / use-case problem entity)
• Official name, aliases, ambiguity risk
• Whether it already has a dedicated page
• Whether it is already covered by structured data
• Whether it already exists on third-party reference pages

The value of this table is that it exposes, all at once, “entities that should exist but do not,” “entities that have pages but no structured data,” and “entities with no off-site evidence at all,” laying out clear gaps for the work that follows.

The inventory should also include attribute identification—an entity is not “just a name,” but “a name + attributes + relationships.” Every entity should record its name, aliases, official URL, logo, description, parent organization, applicable categories, comparison dimensions, geography, price tier, feature characteristics, and other attributes. The more complete the attributes, the easier it is for the machine to disambiguate.

Step 2: Prioritizing Entities

The entities you inventory need not be treated equally. Rank them by the four dimensions of “business value × search demand × AI mention potential × maturity of on-site resources,” and make the high-value, easy-to-execute entities solid first.

Step 3: Connecting Entities into a Relationship Graph

After the inventory, connect the entities into a network. A knowledge graph is fundamentally “entities + attributes + relationships,” where nodes are entities and edges are relationships. We recommend first building the minimal viable loop of a brand knowledge graph:

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

Step 4: Defining Main Entities and Reusable IDs

When drawing the graph, answer “which is the main entity” for each key page, and assign stable entity IDs that can be reused across pages to core entities such as the brand, organization, and founder. In Schema.org, @id is used to reference the same entity across pages, and mainEntityOfPage is used to indicate which entity the page primarily describes (Per: Schema.org). The “entity → page” mappings in the inventory table should stay consistent with the main-entity definitions in the relationship graph.

Three Modeling Emphases for the Relationship Graph

Choose a modeling emphasis based on site type: organization-centric suits corporate websites, product-centric suits SaaS / e-commerce / DTC, and topic-centric suits media / content sites / B2B education sites. With a different emphasis, the core nodes of the relationship graph differ as well.

Exercise

Using a real brand, complete two things: first, produce a master entity list that includes type, ambiguity, page coverage, and structured-data coverage status; second, based on that list, draw a relationship graph containing at least 1 organization entity, 1 brand entity, 5 product or service entities, 10 topic entities, 5 person / case / comparison entities, and 30 relationship edges, and mark the reusable @id for each core entity.

Deliverables

• Master entity list
• Entity naming convention table
• Entity ambiguity and gap list
• Entity relationship graph and entity relationship dictionary