How Google Uses Entities
Google doesn't search for keywords — it searches for entities and their relationships. Understanding this shifts how you approach content creation entirely.
What you'll learn
- Understand what entities are and how they differ from keywords
- Learn how Google's Entity Annotation system processes documents
- Understand the EAV (Entity-Attribute-Value) model and why it matters
- Know how to identify which entities belong on a given page
- Learn why entity coverage determines whether Google considers a page complete
What Is an Entity?
An entity is a distinct, identifiable concept with a stable meaning — a person, place, organization, product, service, or abstract concept. Unlike keywords, which are just strings of characters, entities have a defined identity in Google's Knowledge Graph.
The difference matters enormously in practice:
| Keyword | Entity Google Understands | Expected Attributes |
|---|---|---|
| forex trading | Forex Trading (Concept) | broker, currency pair, leverage, spread, regulator, risk |
| car insurance UK | Auto Insurance (Service) | policy type, insurer, regulator, premium, excess, coverage |
| Tom Cruise movie | Tom Cruise (Person) | filmography, genre, studio, director, box office |
| buy sofa | Sofa (Product) | brand, material, dimensions, price, delivery, warranty |
When a user searches "forex trading," Google doesn't look for pages containing the phrase "forex trading" — it looks for pages that adequately cover the entity Forex Trading by discussing its relevant attributes, related entities, and sub-concepts.
The Knowledge Graph: Google's Entity Database
Google maintains the Knowledge Graph — a massive database of entities and their relationships. Each entity has a unique identifier (like a Wikipedia article's URL or a Wikidata Q-number). When Google processes a page, it runs entity extraction to identify which Knowledge Graph entities are present and how confidently they appear.
EAV: Entity-Attribute-Value — How Google Models Content
Google evaluates content using the EAV (Entity-Attribute-Value) model. Every piece of information on a page can be decomposed into three components:
- Entity — the subject (e.g., "Forex broker")
- Attribute — a property of that entity (e.g., "regulatory status")
- Value — the specific value of that attribute (e.g., "FCA authorized")
A high-quality page covers not just the central entity but also populates its key attributes with accurate values. A page that only mentions "forex broker" without discussing regulation, spreads, leverage limits, and account types is covering the entity at a surface level — Google's systems can detect this incompleteness.
Before writing any piece of content, list the 10–15 EAV triplets that a complete, expert page on this topic should contain. Ensure your content addresses each of them explicitly. This is what it means to write for entity coverage.
Every Project Has Its Own Entities
A critical mistake in entity research is transferring entity lists between different sites. Entities are determined by three factors: niche, business model, and audience.
Example: two sites both cover "cybersecurity." One is a consumer antivirus product; the other is an enterprise WAF (Web Application Firewall) vendor. Their entity sets are completely different:
- Consumer antivirus: virus, malware, real-time protection, free trial, subscription, ease of use, compatibility
- Enterprise WAF: DDoS mitigation, compliance, OWASP Top 10, enterprise deployment, SLA, SOC 2, integration
Mixing these entity sets produces content that serves neither audience well and that Google cannot clearly categorize.
How to Find the Right Entities
Entity research is an evidence-based process — you don't invent the entities, you discover them from data:
- Competitor top pages in Ahrefs — sort by number of keywords (not traffic). Pages ranking for the most queries are covering the most entity attributes.
- People Also Ask — each PAA question corresponds to an entity attribute users expect answered.
- Wikipedia categories and "See Also" — Wikipedia structures topics similarly to how Google models them.
- People Also Search For — the related searches shown after clicking back from a result reveal sibling entities.
- GSC queries with impressions — Google is already associating your site with these entities.
Key Takeaways
- Entities are identifiable concepts with stable meanings; Google matches them in documents using Knowledge Graph IDs
- The EAV (Entity-Attribute-Value) model describes how Google expects content to be structured around entities
- A page must cover the central entity's attributes with specific values — surface mentions are not enough
- Every project has unique entities determined by niche, business model, and target audience
- Entity discovery is data-driven: use competitor top pages, PAA, Wikipedia, and GSC query data