Most Important Domains for Personal Knowledge Panels (James Dooley Interviews Jason Barnard)

James Dooley: Hi, today I'm joined with Jason Barard from Calube and I'm going to jump straight in Jason and ask you which domains are the most important at triggering and creating a knowledge panel, specifically for a person.

Jason Barnard: That’s a really good, specific question. I’m going to start off by saying don’t look at domain authority. There may be a lot of websites within the list that I’m going to give, and within the list on Caliq Pro that we make available, that happen to have high domain authority. But the correlation is there, the causation is not. Wikipedia is obviously super important. If you can get a Wikipedia page, you’re going to solve a lot of your problems in terms of getting a knowledge panel and making it stable. However, the story, the brand narrative, who you are, what you do, who you serve is then controlled by Wikipedia editors. Unless you’re famous, I would advise against it. Wikidata obviously comes pretty high up. LinkedIn is incredibly important. Here’s something that 99% of people don’t know. It’s not just one URL for your LinkedIn profile you should be paying attention to. There are eight.

James Dooley: You did indeed surprise me there. I didn’t know there’s eight. Can you send me the eight? I can put them in the description.

Jason Barnard: Yep. Absolutely. Crunchbase, especially in business. Similar idea there. Then there’s, I’m looking at my list here, Namu.wiki. I didn’t even know what that was until it cropped up in our dataset. 25 billion data points is how I get all this information. With those 25 billion data points, we’re able to identify which websites Google uses as reference. That’s from the Google leak of last year where they let the cat out of the bag. Some URLs are used as reference sources for specific entities. It’s boolean. True or false. Does this URL represent the entity we’re looking at. Do we trust it for information about that entity. That’s what Caliq Pro does best. We can tell you the 300 URLs that will make the difference for your personal brand in Google, and indeed in AI, because we built an algorithm that then expands that out to AI using the citations and a lot of other data we pick up from the AI.

James Dooley: Before you move down the list, I’ve got one or two questions on that. Caliq Pro gives me the 300 domains and you’re saying it’s not related to domain authority. So I’m presuming it’s based off trust and confidence related to who you are. Is that correct?

Jason Barnard: Yeah. We use data from knowledge panels, from Google search results, from AI mode results and the citations in there. Then moving away from Google over to ChatGPT, Perplexity, Claude, Grok, we’re tracking all of these engines. We’re tracking the citation sources and we can bring all of that together to actually understand which URLs are important to all of them. They all use the same data source, the web. They’re all looking at the same entities, which is us. They’re all trying to make sense of the same thing for the same people for the same purpose. They’re trying to understand James Dooley so they can give James Dooley to their audience when he can provide the solution to them. They’re using your digital footprint to figure that out. The other really good part about this is we have the biggest dataset of Google’s own knowledge graph outside Google itself. I’ve been collecting data from the knowledge graph for 10 years and Google are foolish enough to give it away for free. It hasn’t cost me a penny except in storage and some head space. We have 74 million entities collected from the knowledge graph and we track them all.

James Dooley: Do you know approximately, out of those 74 million entities, how much is made up of people versus companies or books or podcasts?

Jason Barnard: We focused on people and companies. It’s mainly people and companies. It’s about half people. I don’t know, about 25% companies. A lot of books. Some podcasts. Not a huge number of podcasts. Events at one point was quite popular but now has disappeared. So it’s really people, corporations, books. There are some products, not a huge number. And topics as well.

James Dooley: Before we expand back onto the domains for triggering a knowledge panel for a person, is triggering a knowledge panel for a person easier than a business, or is a business easier than a person?

Jason Barnard: A person is significantly easier than a business. There are multiple reasons for that. One is Google is focusing on people. People are difficult to track down whereas companies are easy to track down. Companies tend to have websites, people tend not to have websites. People tend to write and do things on multiple different platforms. Corporations do that a lot less. People are significantly more ambiguous. You’re going to share your name with 20,000 or 30,000 other people. I share my name with 10,000 other people. You probably share it with 100,000 other people, some of whom are famous. Google’s focused on people because it needs to disambiguate to understand who the authorities are. If you’re looking at E-E-A-T, it doesn’t want to apply your signals to another James Dooley and vice versa. Identifying people is a fundamentally important part of what Google’s doing today. Corporations, they use the website as a proxy for the entity of the corporation. They don’t need to worry about that so much.

James Dooley: You said it’s harder for Google to understand a person because generally a person doesn’t have a domain, whereas a company has a domain. So, on what domains are most important for triggering a knowledge panel, I’m presuming you would agree the number one domain is making certain that the person has got a domain and an entity home.

Jason Barnard: Yeah. I was actually having a chat with Google Gemini the other day talking about claim, frame, proof. Your own domain is where you claim and you can frame what you’re claiming. You can say I’m an expert in brand SERPs, and that means I’m exactly the right person you need for your online reputation management problems or your proactive online reputation management project. That’s my claiming and framing. I can say that on my own website. It’s more difficult to say that on other websites. So consider your website as where you claim and frame, and then find the proof on other websites that shows third parties agree with you.

James Dooley: So these domains you listed are third party sources that prove your claim and your frame with who you are, what you do, and who you serve.

Jason Barnard: Exactly. Gemini was saying your website is by far the most important aspect of all of this, but it isn’t the one that’s going to be the dominant proof. It’s where you claim and frame. Without the claiming and framing, the proving doesn’t mean anything. First thing is set up your website. Make sure you’re making your claims and framing them in a way that serves your purpose today. Then go out and find the proof. Number one is you’ll often find you already have a lot of proof. Find it, link to it, and reach out to the people who created the content to ask them to reframe what they’ve written. If you can get the proof of the claim framed the same way on the third party as it is on your website, you’re off to the races.

James Dooley: If there’s a third party source that explains me exactly how I want to be described, should I link from the entity home to that article, or can that just be done in schema?

Jason Barnard: I would do both. Schema markup is simply supposed to reiterate what’s already on the page, with some exceptions. Don’t consider schema markup to be standalone. It should re-explain the page’s content in a format the machines can digest. Schema markup is also limited. So don’t obsess about schema. HTML is just as powerful if it’s done well. Use schema because it reiterates what’s already said in the HTML and increases the confidence level of the machine. Link out to all the proof. We started off with Wikipedia, Crunchbase, LinkedIn. Twitter is up there. Amazon is huge. As an author, that’s a great opportunity. Self-publish your book on Amazon. Amazon is a trusted source by Google and the other AI. Google Scholar is incredibly powerful. It’s a Google property. You’re injecting information right into the heart of the beast. You don’t need a book. You can put articles on there. You manage it yourself. Muck Rack is hugely important if you’ve written any articles. You create a profile, you write your own description, you link out to your preferred articles, and it will collect new articles automatically. Goodreads is great if you’ve written a book. The problem is you can’t edit it yourself, you need a librarian at Goodreads. At Caliq Pro we have Alisa who is a librarian and she handles this.

James Dooley: You’ve mentioned Amazon, Goodreads, Google Books. All three are about publishing a book. How important is it, if someone wants a powerful knowledge panel, to publish a book? Even if they’re a roofer.

Jason Barnard: Start with articles because they’re shorter. If you’re a builder, write on the Building Association of the UK or the Building Association of Glasgow if you live in Glasgow. You want it as relevant as possible. I’ve been naming the big hitters. At Cali Cube, we share the hundred top domains that support your entity in Google’s knowledge graph and in AI. But there are 62,000 in our dataset. We track 23 million URLs, 1.7 million domains on a regular basis. 62,000 are reference domains for Google. That’s 3.58%. Less than 4% of domains out there are going to bring seriously powerful value for building your place in the knowledge graph. We share 100 for free, but we have 62,000, and it gets very niche. That’s where the 25 billion data points comes in. If you tell us what you do and who your competitors are, we can tell you exactly which domains will help you, which ones you already have, and which ones you should prioritise.

James Dooley: That’s interesting because in link building agencies we used to obsess over metrics like DR, trust flow, toxicity. Now this feels like a different dimension. We’re looking at trusted sources that feed the knowledge graph and LLMs. A few more questions on domains. You mentioned Google Scholar, Wikidata, Crunchbase. These are important, but I’ve heard horror stories about people trying to tweak Wikidata or paying for Wikipedia. Can anyone create their own Crunchbase, Google Scholar, and Wikidata? Which ones should people avoid editing themselves?

Jason Barnard: You can create your own Wikidata. You can create your own Crunchbase. You can create your own Google Scholar. You can create your own IMDb. The one you’re likely to get stung by is Wikidata. The admins are very careful. If you create a Wikidata profile and they don’t feel you’re deserving, they delete it. People come to us saying their knowledge panel disappeared. Often it’s because the knowledge panel was triggered by Wikidata, the Wikidata got deleted, and the panel disappeared. Once you’ve been deleted, if you create a new one, it’ll get deleted again, even if you are notable enough, because you’re supposed to reactivate the old one. There’s lots of stuff like that in Wikidata and Wikipedia. If you don’t know it, you’ll get stung because the admins and editors don’t take prisoners. Crunchbase is less risky but there are human editors. IMDb has human editors too. Any platform that’s respected will have humans in the loop. Bear that in mind.

James Dooley: If an entrepreneur watching this doesn’t know whether an old agency created a Wikidata page that got deleted, how do they find out?

Jason Barnard: We track Wikidata and we track a lot of people. If we don’t have the data, Alisa will figure it out. She’ll also often say don’t create Wikidata yet, not because you’re not notable enough, but because we haven’t built the proof. She will spend time building the proof first and then create it.

James Dooley: You mentioned IMDb earlier. Is that another important profile, alongside Crunchbase and Muck Rack?

Jason Barnard: IMDb is interesting because it was used as training data for a long time. It lists films, so anyone in a film gets an IMDb page. But podcasts count too. If you’ve got a podcast, or you’re a guest, you can get an IMDb page. It can be helpful and powerful.

James Dooley: For tradespeople like roofers, they are not going to be on IMDb. Is this where people can work with you and you can say here are 250 sources specifically for a roofer in Glasgow?

Jason Barnard: First thing we do is look at what you already have. People overlook that. They immediately say they want lots of profiles. Sort out what you already have first. The machines are trying to put all of this fragmented web together and they’ve probably missed 90% of what you’ve done. We make sure they see it, understand it, and we frame it to your advantage using your website. Then we find where your competitors or peers are that you are not. We use a Venn diagram approach. You give us your 20 competitors, we find the sweet spot where they all overlap. Start there and work your way out. Start with the most effective and don’t waste time.

James Dooley: Could I then add those sources into Google Scholar?

Jason Barnard: You can add anything you’ve written to Google Scholar. It’s you saying you write articles, books, or papers, and here they are. Then link to that from your own website. Do not worry about saving link juice in this context. Link juice is not the priority here. The priority is pointing to the right places so the machine can say that’s you. Publishing books is a great example. If you do it wrong, it will duplicate you. You publish a book, it creates a new James Dooley. Then you don’t get the authority. If you do it right, every book adds authority to the correct entity, not a duplicate, and not another person with the same name.

James Dooley: That’s where entity recognition and disambiguation comes in. It’s confusing.

Jason Barnard: Exactly. I gave a talk 10 years ago called Empathy for the Devil. If you have empathy for Google’s problem, which is a messy web, and you sort out your corner of the web and make it clear, it will reward you. We had a client who spent £60 on a tool that created 100 social profiles. That created 100 junk profiles with no value as reference sources. Then when they pivot or change their bio, they have to update 100 profiles manually. Even if they’re not important, they create confusion. It becomes a ball and chain.

James Dooley: I’ve got a devil’s advocate question. Some people will talk about link sculpting and keeping page rank. When you link out to these proof sources, do you use dofollow or nofollow?

Jason Barnard: Dofollow. I won’t argue that link dilution is not a thing, because it is. But in the context of the algorithmic trinity, links are a much smaller part of the picture now. E-E-A-T and other signals matter. Knowledge graph is not about links. LLMs are not about links. The trade-off is simple. You might lose a tiny bit of link juice, but you gain credibility and understanding. The machine becomes confident. It doesn’t guess. That confidence is the win.

James Dooley: Where’s the line between self-corroboration everywhere versus wasting time on sources that are not trusted?

Jason Barnard: I’m writing an algorithm that weights first, second and third party corroborative power. First party is what I control. Second party is where I control the content but not the website. Third party is where I have no control. JasonBarnard.com is first party. Kalicube.com is technically second party because it’s a different entity, even though I run it. An article written by an employee on Kalicube.com can have more value because it is not me writing about me. LinkedIn is second party. I control my profile but not the platform. It’s very valuable. Muck Rack is second party too. Third party is when somebody else mentions you without you controlling it. If FatRank mentions Jason Barnard and says he’s great at online reputation management, that’s third party corroboration. The machines don’t care that we know each other. They can’t know. I wouldn’t waste time creating 100 junk profiles. I would spend time getting a real mention on a relevant page that has visibility. One sentence from a trusted third party can tick all three boxes in our model, understandability, credibility, deliverability.

James Dooley: One last question. People often can’t get Wikipedia or struggle with Wikidata, so they go to services selling pages on WikiAlpha, WikiGenius, and similar sites. They can rank and feed LLMs now. Do they have value?

Jason Barnard: You’re right about what you remember me saying. That was five or six years ago. The world changes. I wasn’t wrong at the time, but quoting that advice today would be bad advice. Back then, creating those pages could trigger a knowledge panel, then when you tried to claim it, it got deleted because Google used it as a spam signal. Wikidata wasn’t very good then. Now the calculation is different because Google is not the only game in town and those sites have improved. They feed multiple algorithms. Use them as part of a wider strategy. Self-corroboration is fine, but if it’s all you have, you’ll get caught out in the future. You still need real third party corroboration. The entity home website remains the centre. It is the control point for your digital footprint and where you claim and frame, then you point to proof. A little bit of everything, not too much of anything.

James Dooley: You taught me the infinite loop of self-corroboration and multifaceted corroboration. I used to look at domain authority. Now I look at trust, LLM citations, rankings, conversions, and knowledge graph strength. Hopefully this has given people a strong list of domains that help trigger a person knowledge panel. Jason, it’s been an absolute pleasure again. Hope to get you on again soon.

Jason Barnard: It was delightful.