Here is the part most brand teams find hard to believe: the page that shapes how Google and ChatGPT describe your company is rarely your homepage, and almost never your best-performing landing page. It is usually the About page, the one your analytics say nobody reads. Machines read it. They read it more closely than any buyer does, because they are trying to answer a question your buyers already think they know the answer to: who exactly is this?
That page has a name. Jason Barnard, founder of Kalicube, popularized the term “entity home” to describe it, and Kalicube defines it as “the single, authoritative page on the web that defines an entity, acting as the primary source of truth for algorithms.” This guide walks through how to pick that page, mark it up, connect it to the rest of the web, and check whether Google and AI assistants are reading it the way you intend.
What is an entity home page, and why does Google need one?
Think of a street corner in the City of London. Two signs, Old Bailey and Green Arbour Court, tell you exactly where you stand. Nobody has to guess. Your entity home page does the same job for a search engine. It is the fixed address where the facts about your company or your person live, and it is the place every other mention can be checked against.
Google’s Knowledge Graph does not store web pages. It stores entities (people, companies, products, places) and the facts that connect them. When Google meets a new mention of your brand on a podcast page, a Crunchbase profile, or a trade article, it has to decide whether that mention refers to you or to a different company with a similar name. Without a stable reference point, it reconciles those mentions by majority vote, and the majority is often wrong, out of date, or written by someone who copied your 2019 boilerplate.
An entity home page gives the algorithm a tiebreaker. It says, in plain language and in structured data, what you are called, what you do, who runs you, where you are based, and which profiles elsewhere on the web belong to you. When the rest of the web agrees with that page, Google gains confidence. When it does not, Google hedges, and hedging is how you end up with no Knowledge Panel, a panel showing the wrong logo, or an AI answer that confuses you with a competitor.
Score your setup with the HOME test
At Instant Press we audit entity signals before we pitch a single journalist, because press coverage only compounds when the coverage points somewhere coherent. The model we use is the HOME test. It has four parts, and you score each one from 0 to 2, for a maximum of 8.
H is for Hub. Do you have one URL, and only one, that you treat as the source of truth? Score 0 if facts about you are scattered across a homepage hero, a press kit PDF, and a LinkedIn bio. Score 1 if an About page exists but contradicts other pages on your own site. Score 2 if a single page holds the canonical description and everything else on your domain agrees with it.
O is for One description. Is the core sentence about you the same everywhere? Score 0 if your LinkedIn, Crunchbase, X bio, and website each describe a different company. Score 1 if they rhyme but differ on facts like founding year or headquarters. Score 2 if the same plain description, with the same facts, appears on every profile you control.
M is for Markup. Does the hub carry schema.org structured data that names the entity and lists its other profiles through sameAs? Score 0 for none, 1 for generic plugin output that marks up the website but not the organization, and 2 for a deliberate Organization or Person block with a stable identifier and a complete sameAs list.
E is for Echo. Do third parties point back? Score 0 if no independent source links to the hub. Score 1 if a few directories do. Score 2 if trusted sources (a Wikidata item, industry databases, press coverage, speaker pages) link to the hub and repeat its facts.
A score of 6 or above means your entity home is doing its job and the next gains come from earning more coverage. Below 4, fix the foundation before you spend money on visibility, because new mentions will scatter into the same confusion as the old ones.
Which URL should be the hub? Google’s own documentation for Organization structured data says: “We recommend placing this information on your home page, or a single page that describes your organization.” For most companies the About page wins, because the homepage gets rewritten every time marketing changes the campaign. An About page tends to stay put, and it already exists to answer the identity question.
Write the schema that tells machines where home is
Structured data is filing. You are taking facts that a human reads in a paragraph and putting each one into a labeled folder that a machine can open without guessing. On the entity home page, the folder is a schema.org Organization block for a company, or a Person block for an individual.
Google made this easier to justify in November 2023, when it published “Expanding markup support for Organization details, including Logo structured data” on the Search Central blog. The current documentation is blunt about the barrier to entry: “There are no required properties.” That means you have no excuse to skip it, and also no checklist to hide behind. You choose what matters.
The properties that do the heavy lifting are name, url, logo, description, foundingDate, founder, address, and sameAs. Give the block a stable @id (the hub URL plus a fragment such as #organization) so other pages on your site can reference the same entity instead of creating duplicates. Then mark the page itself as an AboutPage whose mainEntity is that organization. A minimal version looks like this:
{
"@context": "https://schema.org",
"@type": "AboutPage",
"url": "https://example.com/about",
"mainEntity": {
"@type": "Organization",
"@id": "https://example.com/about#organization",
"name": "Example Co",
"url": "https://example.com",
"logo": "https://example.com/logo.png",
"description": "Example Co makes accounting software for independent bakeries.",
"foundingDate": "2019",
"sameAs": [
"https://www.linkedin.com/company/example-co",
"https://www.crunchbase.com/organization/example-co",
"https://www.wikidata.org/wiki/Q000000"
]
}
}The sameAs list is where most teams get lazy. Google’s documentation describes it as “The URL of a page on another website with additional information about your organization” and notes that “You can provide multiple sameAs URLs.” Include every profile you control and every authoritative profile about you: LinkedIn, Crunchbase, your Wikidata item if one exists, your Wikipedia article if one exists, YouTube, and industry directories. Leave out profiles you abandoned years ago, since a sameAs link to a dead account with old facts works against you.
For a founder or personal brand, swap Organization for Person and add jobTitle, worksFor (pointing at the company’s @id), alumniOf, and knowsAbout. If the person and the company both matter, give each its own entity home and link them to each other. That two-way link is how Google learns that this person founded that company, rather than inferring it from a stray press quote.
Make Wikidata and every profile point back
Schema on your own site is a claim. Claims from the entity itself carry limited weight until independent sources confirm them. That is the job of the echo, and Wikidata is the most structured echo available.
Wikidata is the free, open knowledge base run by the Wikimedia Foundation, and every item carries a Q-number identifier. An item can hold an “official website” statement (property P856), which is the cleanest possible pointer from a trusted public database back to your entity home. When your sameAs includes the Wikidata item and the Wikidata item names your site as official, you have built a closed loop that machines can verify in both directions.
Be careful here. Wikidata has its own notability rules: an item needs to describe a clearly identifiable entity supported by serious, publicly available references, or fill a structural need. Creating a thin item about a two-person startup with no independent coverage invites deletion, and editors notice self-promotion. The honest path runs through earned media. Coverage in real publications gives you the references a Wikidata item needs, and the item then gives Google the confirmation your entity home needs. This is one reason press placement and entity work belong in the same plan.
Wikidata is only one echo. Every profile you control should carry the same one-sentence description and link to the hub, not to the homepage. Update the boilerplate in your press kit so journalists copy the right facts. Fix the founding year on Crunchbase. Rewrite the LinkedIn About section so it matches word for word. None of this is glamorous, but each corrected profile removes one contradicting vote from the pile Google is reconciling.
Do AI assistants read your entity home the same way?
The same consistency problem now shows up in ChatGPT, Gemini, Perplexity, and Claude, and the stakes are higher because you cannot see a panel to know what went wrong. An AI assistant answering “what does Example Co do?” draws on what it learned in training and, increasingly, on pages it retrieves at answer time. Either way, it is summarizing what the web says about you.
If the web says five different things, the assistant either picks one at random, blends them into something false, or declines to say much at all. If the web says one thing, repeated across your entity home, your profiles, Wikidata, and independent coverage, the assistant has far less room to drift. Consistency is not a ranking trick. It is a way of reducing the number of plausible wrong answers.
Two practical habits help. First, write your entity home description as a sentence an assistant could lift verbatim: subject, category, audience, differentiator, in under 30 words. “Example Co makes accounting software for independent bakeries” survives summarization. “We reimagine the future of financial clarity” does not, because it contains no facts to repeat. Second, test it. Ask three assistants who you are, note the exact phrasing they use, and trace each wrong fact back to the source that probably planted it. That source is usually an old directory listing or a stale bio, and fixing it is often a ten-minute job.
Your entity home page also gives you a place to correct the record on your own terms. When a product line changes or a founder leaves, update the hub first, then the schema, then the profiles, then pitch the change. That sequence means the newest coverage confirms the newest facts instead of fighting them.
Look at how the person who named it built his own
The cleanest working example we found belongs to Jason Barnard himself. We checked his site, jasonbarnard.com, in October 2026, and it practices exactly what the entity home concept preaches.
The root page carries JSON-LD that declares the page an AboutPage, with a Person as its mainEntity. The description inside the markup opens with “Jason Barnard is an award-winning entrepreneur, inventor, speaker and author.” The sameAs array lists 82 URLs, from his Kalicube author page and Bluesky profile to an ISNI record and his Wikidata item, Q23416738. On the Wikidata side, that item’s official website statement points back to jasonbarnard.com. The loop is closed in both directions.
Notice what he did not do. He did not rely on Kalicube’s corporate site to describe him, even though he founded it. The person and the company each have their own home, and they point to each other. He also chose his homepage as the hub rather than a buried About page, which works because the whole domain is about one person. A company with a product homepage would put the same structure on its About page instead.
Run the HOME test on that setup and it scores well on all four parts: one hub, one description repeated across profiles, deliberate Person markup with a long sameAs list, and an independent Wikidata item that echoes the hub. That is the standard to measure your own setup against. Pick your URL this week, write the sentence, mark it up, and then start earning the coverage that makes the rest of the web agree.