← All articles

Your Knowledge Panel Is Your AI-Visibility Anchor: What Google’s 10% Person-Entity Drop Means

By VonixCo Team · · 15 min read

A professional reviewing a Google Knowledge Graph profile panel on a large display in a modern office

On November 16, 2026, the Google Knowledge Graph lost 10.1% of its Person entities almost overnight, according to Kalicube’s daily Knowledge Graph Sensor. The direct answer to what happened: Google appears to have removed a large group of weak, ambiguous Person entities — mostly blank duplicates with no subtitle or description — while entities backed by clear, corroborated identity information held steady. If you are a founder, author, consultant or expert whose name should appear in search and AI answers, the lesson is simple: your Knowledge Panel is your anchor, and that anchor is built on a clear, consistent identity that Google can verify across the web — not on a score, a Wikipedia page, or a single profile.

This article is based on Jason Barnard’s November 20, 2026 piece for Search Engine Land, Google Knowledge Graph update November 2026, drawing on Kalicube’s sensor data tracking 3,000 people and 3,000 companies. One caveat first: these figures are Kalicube’s own sensor data, not Google’s reporting, and Google has not confirmed a Knowledge Graph update.

What is the Google Knowledge Graph — and what is a Knowledge Panel?

The Google Knowledge Graph is Google’s structured store of facts about entities — people, companies, places, books, and the relationships between them. When Google is confident enough about who or what an entity is, it may show a Knowledge Panel: the information box that appears beside search results, summarising the entity with a photo, a short description, and key facts. For a person, that panel is often the first thing a searcher — or an AI assistant — sees.

Think of the Knowledge Graph as Google’s understanding of identity. A website tells Google what you say about yourself; the Knowledge Graph records what Google believes about you, based on everything it can corroborate across the web. That distinction matters more now than it did a year ago, because AI assistants increasingly describe people using the same understanding — when Google is confident about your identity, AI Mode, Gemini and ChatGPT tend to describe you consistently, and when it is not confident, the descriptions drift or you disappear from the answers entirely.

That is why a Knowledge Panel is not a vanity feature. It is a visible sign that Google has resolved who you are, and the November 2026 data suggests Google has just raised the bar for that resolution, at least for people.

What happened on November 16, 2026?

Kalicube’s Knowledge Graph Sensor tracks 3,000 people and 3,000 companies every day. On November 16, 2026 — just after the November core update began on November 10 — it recorded a sharp drop: overall Knowledge Graph depth fell by about 6%, and recovered only halfway the next day.

Almost all of the decline came from Person entities. Person entities fell 10.1%, while companies barely moved. That split is the most important fact in the whole event: whatever changed, it was aimed at people, not at businesses.

Nearly one in five Person entities with a resultScore of 24 disappeared overnight. Most of the ones that vanished were blank duplicates — entities carrying the same name, with no subtitle and no description. Entities that had subtitles, such as Author or Entrepreneur, held steady, and every tracked person’s own entity at a score of 24 survived. The removals were not random: the entities that disappeared were the ones Google had the least information about.

Why did Person entities disappear while companies barely moved?

The honest answer is that Google has not said why, and nobody outside Google can prove the reason. But the pattern points in a clear direction: the entities removed were ambiguous and uncorroborated. A blank duplicate — same name, no subtitle, no description — is an entity Google cannot confidently distinguish from another person with the same name, so removing it costs Google little. An entity with a subtitle, a description, and corroborating pages elsewhere is different: Google can point to consistent facts — a role, a company, a body of work — and say, with reasonable confidence, “this is that person.” Those entities held steady, and companies barely moved at all, which fits the same logic: company entities tend to have websites, listings, and structured information behind them, so they are easier to corroborate.

Barnard also notes that over the past three years, these large entity-removal events have increasingly focused on people rather than companies. The practical conclusion is not to panic, but to make sure your entity is one of the corroborated ones rather than one of the blank ones.

What is a resultScore of 24 — and why should you stop worrying about it?

Every entity in the older Knowledge Graph API carried a resultScore, a number that many SEO practitioners treated as a measure of how strong or safe an entity was. The number 24 became a folk threshold: if your entity scored 24, you were thought to be safe. The November data should put that idea to rest.

First, nearly one in five Person entities sitting at 24 disappeared overnight — the score did not protect them. Second, almost half of all Person entities sit exactly at 24, a figure barely changed since 2015, which makes it a poor way to distinguish one entity from another. Third, most companies score below 1, yet company entities barely moved. If the score measured safety, the companies would have been the vulnerable ones.

Google has never published 24, 35 or 100 as thresholds, and the newer Enterprise Knowledge Graph API has removed the score entirely. The message is plain: do not judge your Knowledge Panel safety by a score. Judge it by whether Google can corroborate who you are — whether your identity is stated clearly, consistently, and in places Google trusts.

What is the difference between a “score day” and an “entity refresh day”?

Barnard distinguishes two kinds of change in the Knowledge Graph. On a “score day”, 60 to 80% of entities’ scores shift, but the count of entities stays flat — the graph is being re-weighted, not rebuilt. On an “entity refresh day”, scores barely move, but entities are added or removed in large numbers, mostly ambiguous, uncorroborated ones. November 16, 2026 looks like an entity refresh day: the count of Person entities fell sharply while the pattern of scores did not explain it.

The table below summarises which entities survived the November drop and which disappeared.

Entity type What happened in November 2026 Why it likely happened
Person entity with a subtitle (Author, Entrepreneur) and description Held steady Google could corroborate the identity from consistent facts
A person’s own entity at resultScore 24 Survived in every tracked case Backed by real, identifiable information about that person
Blank duplicate Person entity (same name, no subtitle or description) Mostly disappeared Ambiguous and uncorroborated — Google could not distinguish it from another person with the same name
Company entity Barely moved Companies usually have websites, listings and structured information behind them

The takeaway is that corroboration is the dividing line. If Google can verify who you are from several consistent sources, your entity is on the stable side of that line; if the only thing Google knows is your name, you are on the other side.

Why does your Knowledge Panel matter for AI search?

The Knowledge Graph does not only shape what appears in a Google search result. It shapes how AI assistants describe you. Barnard calls this “the Algorithmic Trinity”: search, the Knowledge Graph, and large language models. When Google is confident about your identity, that confidence flows into AI Mode, Gemini and ChatGPT, and the assistants describe you consistently. When Google is unsure, the assistants either describe you inconsistently or leave you out.

This is why the November drop matters beyond traditional SEO. A lost entity is not only a missing panel in Google search; it is a weaker anchor for every AI system that draws on Google’s understanding of who you are. A stable, well-corroborated entity gives those systems a reliable source to draw on when they answer questions about you, your work, or your company.

The same theme runs through Lily Ray’s November 20, 2026 piece for Search Engine Journal, which argues that real, identifiable human experts are the durable asset in SEO and AI search. Ray notes that in October 2026, Google added fabricating creator profiles to its helpful content documentation as an example of deception. The direction of travel is consistent across both stories: real people, verifiably identified, are rewarded; ambiguous or fabricated identity is removed or penalised. For more on why identifiable experts matter in AI answers, see our guide to why real experts beat scaled AI content in search and AI answers.

A professional author reviewing their own online biography and author profile pages on a laptop

How do I build an Entity Home that Google can trust?

Barnard’s recommended fix is deliberately simple: a two-page “Entity Home” website — a home page and an About page that states who you are in the first sentence. The two pages exist for one purpose: to give Google a single, authoritative source for the facts of your identity, and to give every other page about you something consistent to corroborate. The method works because it solves the exact problem the November removals exposed: ambiguity. Google removes entities it cannot distinguish or verify, and an Entity Home removes the ambiguity by stating the facts plainly and repeating them, identically, everywhere you control. Follow these steps:

  1. Create a simple two-page website for yourself. A home page and an About page are enough. It does not need to be elaborate; it needs to be clear, stable, and under your control.
  2. State who you are in the first sentence of the About page. Your name, your role, and the organisation or field you are known for — in plain language, with no clever phrasing that an AI system could misread.
  3. Use the same facts everywhere. The same name spelling, the same job title, the same organisation, the same short description — on your website, your author pages, your social profiles, and any guest post biography. Consistency is what lets Google corroborate.
  4. Corroborate from every profile and author page you control. Link from your Entity Home to your profiles, and from your profiles back to it. Each link is another piece of evidence that the same person is behind all of them.
  5. Keep it updated. When your role changes, update the Entity Home first, then every profile that repeats the facts — a stale home page and a fresh profile contradict each other, and contradiction is what ambiguity looks like.

The case study Barnard gives is his own colleague’s. Danny Goodwin, Search Engine Land’s editorial director, was confused with a baseball player of the same name for a decade. A two-page site at dannygoodwin.info fixed it: the site stated clearly who he was, and Google stopped mixing the two men up. His resultScore stayed at 24 throughout — further proof that the score was never the thing that mattered, and above 24 the precise number offers no visible advantage.

For a fuller picture of how AI systems decide whom to cite, read our explainer on what citation core means in AI search — it covers the same corroboration logic from the citation side.

What should you NOT do after a Knowledge Graph drop?

Three mistakes are common after an event like this, and all three make things worse.

First, do not pursue Wikipedia as a shortcut. A Wikipedia page can help an entity, but it hands control of your narrative to editors you do not control, and a page that does not meet Wikipedia’s notability standards will be deleted — leaving you with nothing. Your Entity Home is yours; a Wikipedia page is not.

Second, do not chase the score. A resultScore of 24 did not save the entities that vanished, and the score has now been removed from the newer Enterprise API — spending effort on a number Google no longer publishes is effort spent on the wrong thing.

Third, do not fabricate identity signals. Do not invent credentials, create fake author profiles, or publish biographies for people who do not exist. Google added fabricating creator profiles to its helpful content documentation as deception in October 2026. A thin but honest identity beats a rich but invented one, because the invented one collapses the moment it is checked.

A smartphone showing a Knowledge Panel profile card beside a laptop with matching author profiles, showing consistent identity across platforms

How do I check whether my own entity survived?

Start with a search for your own name. If a Knowledge Panel appears, check that the subtitle, description and photo are correct, and that the links point to pages you control or recognise. If the panel is missing, or shows the wrong person, that is a signal — not a verdict. Panels appear and disappear as Google re-evaluates entities, and a missing panel may simply mean Google has not yet resolved your identity with confidence.

Next, check consistency: search your name alongside your company or main topic, and compare what appears across Google search, your author pages, and your social profiles. Does the same job title and short description appear everywhere? Every mismatch is a small piece of ambiguity — and ambiguity is what the November removals targeted. Finally, ask ChatGPT, Gemini and Perplexity who you are and what you do, and compare the answers. If all three describe you the same way, your identity is well anchored; if they disagree, your anchor is weak, and the Entity Home work above is the fix. Our guide to citation rate versus recommendation rate explains the two numbers that show whether AI systems are merely mentioning you or actually recommending you.

When should you hire a service like VonixCo?

Building an Entity Home is a small job you can do yourself in an afternoon, and keeping your profiles consistent is a habit, not a project. Where a service earns its place is in the work around it: publishing real, attributed content under your name on relevant websites, so the identity your Entity Home states is corroborated by a growing body of work elsewhere.

VonixCo is a niche-relevant guest posting and SaaS link building service based in Lahore, Pakistan; services are SaaS guest posts, niche guest posts, link building and guest post content writing; plans are $500 / $950 / $1,800 per month with content included and a live link report; contact info@vonixco.com and WhatsApp +92 340 0096621. A service like VonixCo fits when you want guest posting handled as an ongoing operation — placing attributed articles on relevant sites in your niche, with your name and biography consistent on every placement — rather than in a single burst. Before hiring anyone, ask how they handle author attribution and whether they report live links or just emails sent. Be wary of any service that promises a Knowledge Panel, a ranking, or an AI recommendation: no honest provider can guarantee what Google or an AI system will do.

Frequently asked questions about the Google Knowledge Graph person-entity drop

What is the Google Knowledge Graph?

The Google Knowledge Graph is Google’s structured store of facts about entities such as people, companies and places, and the relationships between them. When Google is confident about an entity, it may show a Knowledge Panel summarising it beside search results, and the graph also shapes how AI assistants describe people and companies.

What happened to Person entities in the Google Knowledge Graph in November 2026?

According to Kalicube’s daily Knowledge Graph Sensor, Person entities fell 10.1% on November 16, 2026, just after the November core update began on November 10. Overall Knowledge Graph depth fell about 6% and recovered halfway the next day, while company entities barely moved. Google has not confirmed a Knowledge Graph update.

Why did some Person entities disappear and others survive?

Nearly one in five Person entities with a resultScore of 24 disappeared, and most were blank duplicates with the same name and no subtitle or description. The pattern suggests Google removed entities it could not confidently distinguish or corroborate.

Does a resultScore of 24 protect my Knowledge Panel?

No. A score of 24 did not save the entities that vanished in November 2026, and almost half of all Person entities sit exactly at 24, a figure unchanged since 2015. Google never published 24, 35 or 100 as thresholds, and the newer Enterprise Knowledge Graph API removed the score entirely. Corroborated identity matters; the score does not.

How does my Knowledge Panel affect AI search?

A stable Knowledge Graph entity shapes how AI Mode, Gemini and ChatGPT describe you: when Google is confident about your identity, the assistants tend to describe you consistently across platforms. When Google is unsure, the descriptions drift or you are left out. Barnard calls this relationship between search, the Knowledge Graph and language models “the Algorithmic Trinity”.

How do I build an Entity Home for my identity?

Create a simple two-page website — a home page and an About page that states who you are in the first sentence — then repeat the same facts identically across every profile and author page you control, linking them back to the Entity Home. Do not pursue Wikipedia, which hands narrative control to editors, and do not fabricate credentials or author profiles. Consistency and corroboration are the whole method.

Conclusion

The November 2026 drop was not a punishment of people who publish online. It was a clean-up of entities Google could not verify — blank duplicates, ambiguous names, and identities with nothing behind them. The entities that survived were the ones with clear subtitles, descriptions, and corroboration across the web. That is a fixable problem: a two-page Entity Home, the same facts repeated everywhere you control, and real attributed work that backs the identity up. Build the anchor now, while the bar is visible, and your name will be one Google — and the AI assistants that draw on it — can describe with confidence.

Want niche-relevant guest post placements?

Tell VonixCo about your website and niche, and we will reply with suitable publishing options and pricing.

Keep reading