Entity SEO: Why Named Things Beat Keywords Now

Two sentences, same topic:

"Email authentication helps improve your deliverability with major providers."

"SPF, DKIM, and DMARC are checked by Gmail, Yahoo, and Microsoft, and since 2024 all three are mandatory for senders above 5,000 messages a day."

The first one contains the keyword. The second one contains things — named, specific, connected to each other and to the wider world.

Guess which one gets quoted.

That's entity SEO in a nutshell, and it's become considerably more important now that a meaningful share of your audience never sees your page at all — they see a passage of it, lifted into an answer somewhere else.

The short version

An entity is a specific, identifiable thing — a person, organisation, product, place, concept, or event that exists independently of the words used to describe it.

Search engines stopped matching strings and started mapping entities years ago. Google's Knowledge Graph is exactly that: a database of things and the relationships between them.

Two practical consequences:

1. Covering a topic properly means covering its entities, not repeating its keyword
2. Named things are quotable in a way vague claims aren't — which is why cited content consistently shows higher proper-noun density than uncited content

Strings versus things

The old model: you wanted to rank for "email deliverability," so you used the phrase "email deliverability" a certain number of times in certain places.

The current model: Google understands that email deliverability is a concept, connected to SPF, DKIM, DMARC, Gmail, sender reputation, spam filters, bounce rates, and Postmaster Tools. A page demonstrating command of those relationships reads as authoritative on the topic. A page repeating the phrase reads as a page repeating a phrase.

This is why keyword density stopped being useful advice and why "write naturally" started being good advice — writing naturally about a subject you actually understand produces entity coverage automatically. You mention SPF because you can't explain the topic without it, not because you're targeting it.

The shift for AI search is that entities now do a second job. They don't just signal expertise to a ranking algorithm; they make individual passages retrievable and quotable.

Why entities matter more for citation

An answer engine assembling a response isn't looking for the best page. It's looking for passages it can use.

A passage full of named things is useful because it's checkable. "Gmail requires DMARC for senders above 5,000 messages a day" can be verified, attributed, and repeated. "Major providers have stricter requirements now" can't be repeated because it doesn't say anything.

Analyses of cited versus uncited content consistently find higher proper-noun density in the cited group. That finding gets reported as a trick — add more proper nouns! — but the mechanism is simpler than that. Specific content is quotable. Vague content isn't. Proper nouns are just the most visible marker of specificity.

There's also a fan-out effect worth understanding. When someone asks a question, engines typically search several related queries to assemble a complete answer. Entity coverage is what makes you retrievable across that whole fan of related queries rather than just the one you targeted.

Finding the entities your topic needs

No expensive tool required. Four methods, roughly in order of usefulness:

Read what already ranks — for nouns, not keywords. Open the top five results and list every named thing they mention: tools, standards, companies, people, metrics, specifications. The union of those lists is roughly the entity set your topic expects. Anything appearing in four of five results and missing from your page is a gap.

Ask an answer engine directly. Query ChatGPT or Perplexity with your topic and note which specific things it names in the response. You're seeing which entities the model already associates with the subject — which is exactly the association you want your page to sit inside.

Mine the "People also ask" and related searches. These surface the adjacent concepts real users connect to your topic.

Check Wikipedia's article on the concept. Wikipedia is heavily represented in AI citations and its internal linking is close to a hand-built entity map. Its "See also" section is often the fastest list of adjacent entities you'll find.

Covering entities without stuffing

The failure mode here is obvious and I want to name it before someone does it: making a list of entities and sprinkling them into paragraphs.

That produces exactly what it sounds like — content that names things without explaining them, which reads as padding to humans and doesn't help models either, because an entity mentioned without relationship carries no information.

What actually works is explaining relationships.

Weak: "Email authentication involves SPF, DKIM, DMARC, and BIMI."

Strong: "SPF lists which servers may send for your domain. DKIM signs each message cryptographically. DMARC ties both together and tells receivers what to do when they fail — and BIMI, which displays your logo, requires DMARC at enforcement before it will work at all."

Same entities. The second version establishes how they connect, which is the actual knowledge, and it produces several independently quotable sentences as a side effect.

The test: could a reader draw a diagram from this paragraph? If yes, you've described relationships. If they could only make a list, you've stuffed.

Being an entity yourself

The part most guides skip, and the one that matters most for a personal-brand site.

You want search engines and models to understand that you are a specific thing — a named person with a defined area of expertise — rather than an anonymous website producing text.

The mechanics:

Person schema, with sameAs. Link your profiles — LinkedIn, X, GitHub, publisher pages — so the entity that is you can be resolved across sources. This is the single most concrete step available.

A real About page that states who you are, what you do, and what you're known for, in plain checkable terms. Not a brand story. Facts a machine can read.

Consistent naming everywhere. The same name, the same job description, the same claims across your site and your profiles. Contradictions between your About page and your homepage don't just look sloppy — they make you harder to resolve as a single entity.

Author attribution on articles, linked to that Person entity rather than a generic admin account.

Verifiable credentials. Published books, named clients, speaking, quantified experience. Things that exist outside your own website and can be checked against it.

That last point is why the consistency one matters so much. If your About page says one thing and your homepage says another, you've given anyone verifying you a reason to trust neither.

How this connects to topical authority

Entity coverage and topical authority are the same idea at different scales.

One page covering a topic's entities properly demonstrates command of that topic. A cluster of pages covering an entire subject area's entities — with sensible internal links between them — demonstrates command of the subject.

Which is the real argument for pillar-and-cluster content structure. It isn't a publishing schedule. It's a way of mapping a domain of knowledge so that both the relationships and the depth are visible from the outside. Topic clusters and internal linking.

Frequently asked questions

What is an entity in SEO?
A specific, identifiable thing — a person, organisation, product, place, concept, or event — that exists independently of the words used to describe it. "Gmail" is an entity; "email provider" is a category. Search engines moved from matching keyword strings to mapping entities and their relationships years ago, which is what Google's Knowledge Graph is.

How is semantic SEO different from entity SEO?
They overlap heavily and the terms are often used interchangeably. Semantic SEO is the broader idea of optimising for meaning rather than keyword matching. Entity SEO is the practical core of it: identifying the specific things a topic involves and covering them and their relationships properly. If you're doing entity coverage well, you're doing semantic SEO.

Does the Knowledge Graph affect rankings?
Not as a direct ranking factor you can optimise for. Being a recognised entity in the Knowledge Graph tends to accompany the things that do drive rankings — established authority, consistent citations across the web, verifiable credentials. Person and Organization schema with accurate sameAs links help engines resolve who you are, which is worth doing regardless of whether it moves a ranking directly.

What to do next

Take your most important page and highlight every proper noun in it. Actual names — tools, standards, companies, people, specific numbers.

If there are fewer than a handful, that's your problem. Not because proper nouns are magic, but because their absence means you're writing in generalities, and generalities are exactly what a model can produce without you.

Then open the top three ranking pages for the same query and do the same exercise. The gap between their list and yours is your content brief.

Free: The SEO audit checklist.

---

Related guides

• SEO when AI answers the question first — the wider picture
• 6 content formats that still get clicks — where entity density pays off
• Internal linking and topic clusters — entity coverage at site scale
• Schema markup for blogs — making your entity machine-readable