404, 410 and Soft 404: Handling Removed Pages Correctly
Do 404s harm rankings, when should you use 410, why soft 404s are dangerous, and should a removed page be redi...
Read
Entity SEO means building content around real-world objects — entities — rather than keywords, and making those entities unmistakable to a search engine. An entity can be a person, a company, a place, a product, an event or a concept: anything you can talk about.
An example makes the difference obvious. As a string, "Apple" is five characters. As an entity it is at least two distinct objects: a technology company and a fruit. Google stores those separately, each with its own attributes (founder, founding date, products) and its own relationships to other entities. The entire logic of semantic SEO rests on that distinction.
An entity is a singular, unique, distinguishable object. It has three characteristics:
In the Knowledge Graph Search API documentation Google states that these objects use schema.org types: Person, Place, Organization, MusicGroup, SportsTeam, VideoGame and others. In other words, schema.org is not merely an "SEO plugin" — it is the vocabulary of Google's entity model. Applying that vocabulary in practice is covered in the schema markup article.
According to Google's own announcement, at launch the Knowledge Graph held more than 500 million objects and over 3.5 billion facts and relationships. Where does that graph get its data? Google's knowledge panel documentation lists the sources:
Google's Knowledge Graph Search API documentation is also explicit about three limitations: the API is read-only (you cannot "write" into the graph through it), it is not suitable for use as a production-critical service, and it returns only individual matching entities rather than graphs of interconnected entities. Each result carries a relevance value called resultScore.
The conclusion matters: there is no "add me to the Knowledge Graph" button. Entity status is earned through consistent, repeated, mutually corroborating information across the web.
1. Disambiguation. The moment a page opens, the reader — and the search engine — should know what it is about. That requires a precise <title>, the entity's name and type in the first paragraph ("WhiteSEO is an SEO agency based in Baku"), and consistent naming. Writing the brand differently from page to page ("White SEO" here, "WhiteSeo.az" there) is the most ordinary and most damaging mistake there is.
2. Corroboration. Google treats a fact cautiously when it appears in only one place. The same address, phone number and company name should repeat in identical form on your contact page, in the footer, in structured data, on social profiles and in directories. This matters especially for local SEO.
3. Connection. Link the entity to other recognised entities: sameAs pointing at your official profiles, internal links to related concepts in the copy. Your internal anchors should name the entity of the destination page — a subject covered in the content cluster article.
Step 1 — Fix a naming standard. Write the official brand name once and keep it identical everywhere. If alternate names exist, declare them in schema with alternateName while keeping the copy loyal to the primary form.
Step 2 — Build a strong About page. This page is the primary document of your brand entity: when it was founded, who runs it, what field it operates in, where it is located, what it has achieved. Google's Organization structured data documentation recommends exactly this page (or the home page) for the markup.
Step 3 — Make contact details identical everywhere. Name, address, phone — down to the character.
Step 4 — Establish sameAs connections. Put your official LinkedIn, Facebook, Instagram and YouTube profiles into the sameAs array of your Organization markup. It says: this site and these profiles are the same entity.
Step 5 — Make authors visible. Every article should carry a name, a role and a short bio. Google's helpful content documentation asks directly: "Is it self-evident to your visitors who authored your content?"
Step 6 — Collect external corroboration. Industry directories, press, partner sites, professional associations. The goal here is not link volume but the same facts repeating in independent sources — much as quality outweighs quantity in backlink work.
A knowledge panel is the information card shown beside search results. Google's official documentation stresses two points:
Panels are generated automatically. They cannot be ordered or bought. A panel appears only once Google recognises the entity with sufficient confidence.
Panels can be claimed. If you represent the entity you can claim the panel, get verified on Google and suggest changes. If you do not represent it, you can propose edits through the feedback link in the bottom right-hand corner of the panel; suggestions are applied after review.
Google's helpful content documentation is organised around three questions: Who? How? Why? The author entity answers the first.
In practice that means a dedicated author page on the site (bio, experience, specialism, profile links) with articles linking to it. The author's name should appear in the same form on social profiles, in talks and in other publications — that is what turns the author into a recognisable entity in their own right.
The last letter of E-E-A-T — Trust — is, in Google's own words, the most important one. Trust is built from verifiable facts: a real name, a real address, real case work, cited sources.
| Tool | What it checks | What to watch for |
|---|---|---|
| Rich Results Test | Validity of the page's structured data | Errors and warnings, completeness of Organization fields |
| Search Console → URL Inspection | Index status and the HTML Google sees | Whether the schema survives rendering |
| Knowledge Graph Search API | Whether the entity is recognised in the graph | resultScore and the returned type |
| A search for your brand name | How Google interprets the brand | Panel, sitelinks, related searches |
| Schema Markup Validator | General schema.org syntax | For types outside Google's own features |
The full checklist, prioritised, is a standard part of our SEO audit service.
1. Multiple spellings of the brand name. "WhiteSEO", "White SEO", "Whiteseo.az" — three strings, three candidate entities as far as Google is concerned.
2. Missing or dormant profiles in sameAs. A link to an empty, abandoned social profile adds no corroborating weight.
3. Hiding the author. Articles signed "Admin" create no author entity.
4. Markup that is not real. Google's structured data guidelines explicitly prohibit marking up invented information or content that is invisible on the page.
5. Squeezing entity work into one page. An entity is built through the consistency of the whole site, not by a single About page.
There is no direct submission route — the Knowledge Graph Search API is officially read-only. Entity status forms through consistent, mutually corroborating information across the web: a stable brand name, a substantial About page, sameAs connections and the same facts in independent sources.
No, it complements them. Keywords show how people search; entities show what they are searching about. The method that combines the two is set out in semantic keyword research.
Yes, and it tends to pay off relatively quickly. Consistent name, address and phone details, correct LocalBusiness markup and identical information across directories all improve visibility in local results.
No. Wikipedia is a strong corroborating source, but Google draws panel information from "various sources across the web" and from data partners. Industry directories, press coverage and official registries also serve as corroboration.
Technical changes (schema, naming consistency) are processed within weeks; full entity recognition usually takes months. The fastest visible effect is a cleaner brand-name search result — accurate sitelinks and description.
Do 404s harm rankings, when should you use 410, why soft 404s are dangerous, and should a removed page be redi...
Read
The difference between 301, 302, 307 and 308, Google's strong versus weak signal distinction, redirect methods...
Read
What HTTP status codes are, what the five families mean and how Google interprets them: indexing, canonical si...
Read