Collecting Relevant Entities
Entity collection is a disciplined research process — not a brainstorming session. Learn the 5-step method for finding the entities Google actually expects on your pages.
What you'll learn
- Understand why entity lists cannot be transferred between projects
- Learn the three questions to answer before collecting entities
- Master the 5 data sources for entity research
- Understand the difference between tematic, subtopic, and intent levels
- Build a prioritized entity list for any page or project
Why Entity Collection Is Project-Specific
A foundational mistake in entity research is treating entities as universal. "Best practices" for entity coverage in the finance niche don't apply to a software niche. Even within the same niche, a B2B SaaS company and a consumer product company target completely different entities.
Entities are determined by three factors that must be understood before any collection begins:
- Niche — the domain shapes which concepts are relevant. Trading and cooking share no entity space.
- Business model — an affiliate site and a SaaS product covering the same topic will emphasize fundamentally different attributes (pricing comparison vs. product features vs. ROI).
- Audience — B2B and B2C audiences look for different attributes of the same product. Enterprise buyers care about compliance, SLAs, and integration; consumers care about ease of use and price.
The Three Questions Before You Start
Answer these before opening any tool:
- What does the client sell, and to whom? — business model, price point, B2B vs. B2C, geography.
- Which competitors are genuinely comparable in business model? — not just who ranks next to you, but who has the same product and audience. An informational competitor has different entity needs than a commercial competitor.
- What queries already bring the client traffic or impressions in GSC? — Google already associates the site with these entities. Start here.
Three Levels to Keep Distinct
When collecting entities, maintain clear separation between three semantic levels:
| Level | Definition | Example (Trading niche) |
|---|---|---|
| Tematic | The broad domain or category | Trading |
| Subtopic | A logical section within the theme | CFD trading, Tax for traders, Technical analysis |
| Intent | The specific user motivation behind a query | Informational, Transactional, Educational |
A subtopic requires a separate page when the two topics have: different semantic focus, different key entities, and different user intent. Use the separation test: "If I made one page, would it fully satisfy both queries?" If not — they need separate pages with separate entity sets.
The 5 Sources for Entity Collection
Source 1: Competitor Top Pages in Ahrefs (Primary)
This is the most important source. Open Top Pages for 3-5 competitors with similar business models, sorted by keyword count (not traffic). Pages ranking for the most queries are covering the most entity attributes that Google validates for this topic.
Analyze: page titles, H2 structure, and what the pages explicitly discuss. Each section heading maps to an entity or attribute. This gives you Google's implicit expectation for entity coverage in your niche.
Source 2: People Also Ask
Every PAA question is an entity attribute that Google considers associated with the topic. Enter your target query and expand all PAA boxes. Group them by theme — each theme cluster represents an entity or attribute category.
Source 3: Google Autocomplete and YouTube Autocomplete
Type your base query in Google and note every autocomplete suggestion. These are statistically frequent user extensions — they reveal which attributes users most commonly associate with the topic. YouTube Autocomplete often reveals different subtopics than web search (more "how-to" and tutorial intent).
Source 4: Wikipedia
Find the Wikipedia article for your main topic. The categories section, See Also, and Further Reading reveal the canonical academic organization of the topic's knowledge domain. Wikipedia structures topics using the same entity-relationship model that Google's Knowledge Graph is based on.
Source 5: People Also Search For (PASF)
After clicking a SERP result and returning to search, Google shows "People Also Search For" suggestions. These are sibling entities — related concepts that users search for in the same session. They reveal which entities Google groups together when processing this query space.
After collecting entities from data sources, you can use an LLM to expand the list by asking it to suggest additional entities of the same types already found. However, every AI suggestion must be validated against real data — LLMs frequently suggest plausible but irrelevant entities that would increase siteRadius and dilute focus.
Key Takeaways
- Entity lists cannot be transferred between projects — niche, business model, and audience determine what entities belong
- Three questions must be answered before collection: what is sold to whom, who are true business-model competitors, what does GSC show
- Maintain separation between tematic, subtopic, and intent levels — mixing them produces unusable entity collections
- Competitor top pages (sorted by keyword count in Ahrefs) is the primary entity research source
- Every PAA question and autocomplete suggestion maps to an entity attribute Google expects on the page