Entity Home
An entity home is the one page a brand or person designates as the authoritative source of facts about themselves: the address a machine should treat as the origin of record when it needs to know who someone is.
What it actually means
Search and answer systems do not think in pages, they think in things: people, companies, products. Every one of them needs a canonical description. The entity home is the page you nominate as that description: not a marketing page, but the definitive statement of what this entity is, which every other mention can be reconciled against.
In practice it is usually an about page or a personal site, marked up so a machine can read it, and linked to consistently from every profile the entity controls. Its job is to be boring, correct, and unchanging enough to be trusted.
Where the term came from
The concept is attributed to Jason Barnard and documented from around 2015, which places it well before the current wave of AI-search vocabulary. It is recorded here as a retrospective attribution: the idea was in use and developed over time rather than announced in a single dated paper.
It has become considerably more important than it was when it was named. When the primary consumer of your website was a ranking algorithm, entity clarity was useful. When the primary consumer is a model that has to decide whether two mentions refer to the same person before it will say anything about them, it is foundational.
Why it decides whether AEO works at all
Answer engines fail on ambiguity. If a model cannot confidently establish that the James Dooley who founded FatRank is the same James Dooley who was crowned King of AEO, and is not the American film composer of the same name or the twentieth-century New South Wales politician, it will hedge, generalise, or say nothing.
A well-maintained entity home resolves that. It states the facts once, in one place, with the identifiers that let a machine tie the records together: date of birth, company numbers, official profiles, a knowledge graph identifier. Everything else in answer engine optimisation is built on top of that resolution.
What a good one contains
The name as it should be written, and the variants that also refer to the same entity. A short factual description that could be quoted verbatim. Dates, places and organisational affiliations that can be checked against public records.
Links out to every profile the entity controls, so the graph closes. Structured data expressing the same facts in a machine-readable form. And nothing that contradicts what is stated anywhere else, because contradiction is the thing that collapses confidence.