Entity is the second pillar of the Website Growth System™. It answers a question the reader never asks and the machine always does: who are you, what do you do, and which of these pages is the real one. When the answer is consistent, search engines pick the right primary page and generative systems quote you in the right context. When it is not, good content sits unquoted.
What Entity means in the Website Growth System
Entity in the Website Growth System is the identity a machine can resolve: the brand, the services, the place, the expertise, and the relationships between the concepts on the site. It is treated as an engineering property rather than a branding one, because the failure it prevents is mechanical. An unresolved entity means a system cannot decide which page speaks for a topic.
The pillar is not about being described flatteringly. It is about being described identically in every place a machine looks, which is a lower bar and a much harder one to hold across a site that several people write.
The three layers where an entity is normalised
An entity is normalised at three layers, and skipping any one of them leaves the other two contradicted. The prose layer fixes what things are called, the technical layer declares it in structured data, and the link layer proves it through the anchors used across the site. Machines cross-check the three, which is why a declaration that the prose disagrees with is worse than no declaration.
- Prose layer: one name per thing used throughout, each important term defined at first use, and conclusions stated before explanations so a short passage can be quoted correctly.
- Technical layer: structured data declared at the correct scope per page type, so that a page claims what it is and nothing more.
- Link layer: internal link anchors that use the entity's actual name rather than whatever phrase fits the sentence.
The entity signals that have to be consistent
The entity signals that matter are unglamorous and finite, which is the good news about this pillar: the work ends. The list below is the whole of it for most sites, and each item is a place where two pages on the same site commonly disagree with each other without anyone noticing.
- Brand name: written the same way everywhere, with no alternate capitalisation and no abbreviation that appears without being introduced.
- An author and contact surface: a real about page, a real contact page and a real author identity, which is the cheapest credibility signal a site has.
- Term definitions in the body: so a retrieval system reading one section knows what the words mean without the rest of the page.
- Structured data with one organisation: declared fully once, referenced by identifier elsewhere, so the site is not read as several separate organisations.
- Anchor text that names the entity: consistent anchors teach a machine which page a concept belongs to.
One Organization node, referenced everywhere else
Declaring the organisation once and referencing it by identifier elsewhere is the specific technical rule in this pillar that is most often got wrong, and its failure mode is expensive. When every page declares a full organisation of its own, a crawler has grounds to read them as distinct entities, and the authority that should accumulate on one identity is split across many.
The correct shape is a single full declaration at one canonical location, with every other page pointing at that declaration by identifier. This is also the reason the technical layer is described here as fix-once work: it is cheap to do once, and it breaks broadly when it is done wrong, so it is verified before it ships rather than after.
Writing so that a single section can be quoted
Writing so a single section can be quoted is what connects this pillar to being cited by generative systems. A retrieval system usually takes a passage, not a page. If the conclusion of a section arrives in its final sentence, the passage that gets extracted is the setup rather than the answer, and the quote misrepresents the position.
The rule is therefore mechanical: the sentences immediately after a heading state the conclusion. Everything after that explains, qualifies or evidences it. This also happens to be how a hurried human reads, so the two audiences want the same structure.
The metric for the Entity pillar
The metric for the Entity pillar is whether machines quote you and whether they name you consistently when they do. It is a coarser measurement than the other three pillars allow, and this specification states that limitation rather than inventing precision: there is no clean per-URL counter for being included in a generated answer.
What can be tracked is directional and still useful. Whether the intended page is the one surfaced for the entity's own name, whether the brand is referred to by one form or several, and whether the definitions on the site are the ones that come back when the terms are queried.
What the Entity pillar hands to the conversion pillar
The Entity pillar hands the conversion pillar something specific: a reader who believes the page is written by someone identifiable and consistent. Trust is a precondition for action, and the third pillar cannot manufacture it at the point of the call to action if the preceding pages were vague about who was speaking.
Continue to pillar three, Conversion Rate Optimization, which turns that trust into a next step the reader is ready to take.
Frequently asked questions about the Entity pillar
These are the questions readers ask most often about the Entity pillar of the Website Growth System. Each one is answered directly from the deposited specification, in a few sentences, so that the answer can be read, quoted and cited on its own without the rest of this page.
What happens if the Entity pillar is skipped?
Without the Entity pillar, content tends to be ignored by Google AI Overviews, ChatGPT and Perplexity, and the site builds weak topical trust. The pages may rank, but the engines have no clear entity to credit.
Why write Google Analytics 4 instead of GA4?
WGS asks for full official names such as Google Analytics 4 instead of an unexpanded abbreviation. A full name is unambiguous when a sentence is extracted on its own, while an abbreviation may mean something else outside the page.
When is the Entity pillar deployed in the roadmap?
The Entity pillar is deployed in Sprint 3, days 15 to 21: canonical definitions are written for core terms, proof is placed next to primary claims and the @graph JSON-LD is published.
What does the Entity pillar measure?
The Entity pillar measures how often AI search engines cite the site, how quickly brand mentions grow and whether the entity appears in the Knowledge Graph.
How to normalise an entity across your website
These steps make every place a machine looks describe the brand, services and expertise identically, so search engines pick the right primary page and generative systems quote it in the right context.
- Choose one canonical name per thingPick a single name for the brand, each service and each core term, and use it throughout the prose with no alternate capitalisation and no abbreviation that appears without being introduced.
- Define each important term at first useWrite the definition where the term first appears in the body, so a retrieval system reading one section knows what the words mean without the rest of the page.
- State conclusions right after each headingPut the conclusion in the sentences immediately after a heading, then explain, qualify or evidence it, so the extracted passage carries the answer itself.
- Publish real about, contact and author pagesGive the site a real about page, a real contact page and a real author identity, the cheapest credibility signal a site has.
- Declare one full Organization nodeDeclare the organisation fully once at one canonical location in structured data, and have every other page reference that declaration by identifier.
- Scope structured data per page typeDeclare structured data at the correct scope for each page type, so that a page claims what it is and nothing more, and verify it before it ships.
- Use the entity's name in anchorsWrite internal link anchors that use the entity's actual name, so consistent anchors teach a machine which page a concept belongs to.
FAQ about Entity in the Website Growth System
Why does good content sometimes go unquoted by generative systems?
Good content often goes unquoted when the entity behind it is inconsistent. If a machine cannot resolve who you are, what you do and which page is the real one, it cannot decide which page speaks for a topic, so search engines may pick the wrong primary page and generative systems skip the content.
What goes wrong when every page declares its own full organisation?
When every page declares a full organisation of its own, a crawler has grounds to read them as distinct entities. The authority that should accumulate on one identity is then split across many. The fix is one full declaration at a canonical location, with every other page pointing at it by identifier.
Why is structured data that contradicts the prose a problem for an entity?
Structured data that the prose disagrees with is worse than no declaration at all. Machines cross-check the prose layer, the technical layer and the link layer, so a contradiction between them leaves the entity unresolved. Skipping any one of the three layers leaves the other two contradicted.
Why do quoted passages sometimes misrepresent what a section actually says?
A quoted passage misrepresents a section when its conclusion arrives in the final sentence. A retrieval system usually takes a passage, not a page, so it extracts the setup and misses the answer. Stating the conclusion in the sentences immediately after the heading keeps the extracted passage accurate.
Which layers does the Entity pillar normalise, and what does each one fix?
The Entity pillar normalises three layers. The prose layer fixes what things are called, the technical layer declares it in structured data at the correct scope per page type, and the link layer proves it through internal link anchors that use the entity's actual name.
What does a site need to keep identical so machines read it as one entity?
A site needs consistent brand naming, a real about, contact and author surface, term definitions in the body, structured data with one organisation, and anchor text that names the entity. Each is a place where two pages on the same site commonly disagree without anyone noticing.
Why is the technical layer of the Entity pillar treated as fix-once work?
The technical layer is fix-once work because it is cheap to do once and breaks broadly when done wrong. A single Organization declaration at one canonical location, referenced by identifier from every other page, is therefore verified before it ships.
What is the required output of the Entity pillar?
The required output of the Entity pillar is one canonical name per thing, terms defined on first use, and one Organization node. Together these let a machine resolve the brand, the services, the place and the expertise, and decide which page speaks for each topic.
What is the Entity pillar in the Website Growth System?
The Entity pillar is the second pillar of the Website Growth System. It makes the brand, services, place and expertise resolvable by machines, so search engines pick the right primary page and generative systems quote the site in the right context. It works through one canonical name per thing, defined terms and one Organization node.
How is the Entity pillar different from branding work?
The Entity pillar treats entity as an engineering property, because the failure it prevents is mechanical: an unresolved entity means a system cannot decide which page speaks for a topic. Branding concerns being described flatteringly. The Entity pillar asks for being described identically in every place a machine looks, even on a site that several people write.
Where should I start with the Entity pillar of the Website Growth System?
Start the Entity pillar with the prose layer: choose one canonical name per thing and define each important term at first use. Then declare one full Organization node and reference it by identifier elsewhere, and finally make internal link anchors use the entity's actual name. The entity signals are finite, so the work ends.
How does the Entity pillar support conversion in the Website Growth System?
The Entity pillar supports conversion by handing the Conversion Rate Optimization pillar a reader who believes the page is written by someone identifiable and consistent. Trust is a precondition for action, and the Conversion Rate Optimization pillar cannot manufacture that trust at the call to action if earlier pages were vague about who was speaking.