Knowledge Panels are a hot topic in SEO and in marketing, but very few people know how to trigger and manage them, and even fewer people know what information is included in them. Kalicube® has provided data for 900,000 knowledge panels in 6 countries across 29 industries that provides astonishing insights into what triggers knowledge panels for brands and people, and what information Google includes in them: the attributes (there are well over 1000 unique attributes), social channels, descriptions, data sources, related entities… and more! If you want to know how to trigger a knowledge panel for yourself or your brand, and also how to control the information Google shows, then this webinar is for you. Jason Barnard shares his in-depth analysis of the data from this massive study, then Andrea Volpini and Paige Hobart join him to share tips, tricks and give you the actionable takeaways for managing your knowledge panel on Google.

Published by: SEMRush. Guests: Jason Barnard, Andrea Volpini. Host: Paige Hobart. November 26, 2020.

Summary: How to Get and Control a Knowledge Panel for Your Brand or Yourself (SemRush)

Event: SEMrush Live Webinar Date: 26 November 2020 Participants: Jason Barnard (presenter), Andrea Volpini (WordLift, co-presenter), Paige Hobart (host) Video: https://www.youtube.com/watch?v=[ID]

Key Takeaways: How to Get and Control a Knowledge Panel for Your Brand or Yourself (SemRush)

Jason Barnard and Andrea Volpini delivered a detailed presentation on triggering and managing Knowledge Panels, with Paige Hobart hosting. This webinar contains critical prior art for the entity home concept, the three-step corroboration process, the child education metaphor, and Andrea Volpini’s linked data principles. It also documents the collaboration between Kalicube and WordLift that informed later patent filings.

1. The Entity Home - “The Single Most Important Thing in This Entire Presentation”

Jason Barnard stated explicitly: “That little globe icon with your site next to it - that’s the single most important thing in this entire presentation. Give your entity a home.” He explained: “It’s not the description. It’s the globe icon with the domain name or even a page. That says: this is where Google considers the home of the entity is. And that is the place it will look for information about the person, the brand, the event, whatever kind of entity it might be - and then go out and corroborate it.”

2. The Three-Step Process: State, Corroborate, Point

Jason Barnard articulated a clear three-step process: “Once you’ve convinced Google that your site is the home - a page on your site is the home for your entity - then you can state a fact on that home. It can be when it was founded, who the founder was, the area served, your social profiles. Then you need to get significant corroboration from multiple independent authoritative sources. And then you can helpfully point to them from the home.” He summarised: “I’ve given you this piece of information (step one: stated a fact). I’ve managed to get it corroborated by multiple independent authoritative sources. Here they are, Google - point to them. Show Google where it can get that corroborative information. It will believe you and it will give you a knowledge panel.”

3. Knowledge Panel = Google’s Understanding + Confidence

Jason Barnard explained: “Left-hand side - Google is recommending what it thinks the best answer or the best solution to your question or your problem is. On the right-hand side, it’s saying: this is fact. This is stuff we are sure about.” He added: “Getting that Knowledge Panel means you have to convince Google that it has fully understood facts about yourself or your brand.” And later: “The Knowledge Panel indicates that Google has understood who you are, what you do, and who your audience is. It’s a representation of Google’s understanding of you as a brand or you as a person. And in that sense, it’s the single most important thing in your entire SEO strategy.”

4. The Child Education Metaphor - Extended

Jason Barnard extended the metaphor throughout: “It doesn’t just believe the man in the street. The teacher would be the equivalent of Wikipedia - the trusted source. The parents would be Wikidata. The evening school teacher would be Crunchbase. The baker down the road is LinkedIn.” He added: “All of these sources that are more or less trusted - the child needs to hear this information from these multiple sources. And the more they hear it, the better they remember it. And the more it’s related to information they already know, the more it will stick.”

5. Google Doesn’t Care About Notable - It Cares About Understanding

Jason Barnard drew a clear distinction: “Wikipedia says you need to be notable. Google doesn’t care about notable. It just wants to understand. If you bear that in mind, you’re thinking: I’m just educating a child.” He gave concrete data: “44.8 percent of brands have Knowledge Panels. 36.6 percent of people have Knowledge Panels. This is from my data set - 70,000 brands and people.”

6. Google My Business ≠ Knowledge Panel

Jason Barnard distinguished clearly: “Google My Business is a business listing where you provide information and Google trusts you on your own good word. The Knowledge Graph is information it has understood on its own, through corroborating from third-party independent sources.” Andrea Volpini added: “Google My Business is a facet of your entity - it’s what Google will perceive about your business especially within the context of local queries.”

7. Consistency Is the Most Important and Most Difficult Thing

Andrea Volpini stated: “One of the biggest mistakes is really to get data wrong. When you start sharing data in structured form, you might not be consistent. And that’s what creates confusion. And then it takes a lot of time to reconcile stuff.” Jason Barnard confirmed: “I didn’t realise how confused the Knowledge Graph gets very easily with what I think is completely consistent information.”

8. Don’t Move the Home - Jason’s Wikipedia Disaster

Jason Barnard described the experiment: “I had three Wikipedia pages - the group I was in in the ’90s, the blue dog and the yellow koala cartoon, and myself. The editors deleted them all within two weeks because I’d messed with them too much.” He explained the key learning: “With myself, the Knowledge Panel disappeared because I moved the home - I moved all the information about me to a new page on the same site. Google lost its crutch. Whereas with the other two, I left the information where it was. Google had a crutch to fall back on, and it kept the information.” Lesson: “Once Google thinks it’s understood, as long as you don’t take away its crutches - the supporting information - it will stick with you.”

9. Andrea Volpini on Linked Data and Unique Identifiers

Andrea Volpini introduced the concept of giving entities not just a web page home but a “web data home” - a persistent URI that can be moved across pages without breaking entity resolution. He explained: “What we do with WordLift is that we try to publish structured data using this format called linked data. We provide this unique ID, this URI, that is persistent across the web. Having a unique identifier inside your structured data will also allow you to have more flexibility when you change your website structure.”

10. Authoritative Sources Are Industry and Geo-Specific

Jason Barnard emphasised: “When I say authoritative sources, I mean authoritative within your industry. If you can’t get Wikipedia or Wikidata, that dog association down the street from me in Paris - they are authoritative for dogs in Paris. So they would be a great source of corroboration. Google will use them as a reliable independent authoritative source.” He added: “Forget about DA [domain authority]. It’s not a Google metric. The dog association is obviously more authoritative on poodle shampooing in Paris than the butcher’s association.”

11. Relationships: Strong, Close, Long-Lasting

Jason Barnard described the quality of relationships that matter for Knowledge Graph presence: “The idea of relationships is strong, close, long-lasting. My relationship with Rand Fishkin is actually quite distant, so it doesn’t help me a great deal. I work a great deal with Andrea - that relationship is obviously closer and longer-lasting.” Andrea Volpini added that WordLift itself is “associated with lots of topics - voice search, JSON-LD, schema markup - topics that WordLift is specialising in. I haven’t seen another brand with that level of connectivity to topics.”

12. The Knowledge Graph Is the Foundation - Bet on It

Andrea Volpini stated: “The Knowledge Graph is a foundational element of modern search. You can go on the Google Assistant and ask information about myself, Jason, WordLift, and SEMrush - and the reason you can ask this question using your voice is that there is a Knowledge Graph behind it.” Jason Barnard added: “Don’t bet your house on the representation of the Knowledge Graph in Google Search. Bet your house on the Knowledge Graph. Work with what you’ve got today, so that tomorrow, if you’re in the Knowledge Graph, you’re solidly in it - and you’ll be able to leverage all the new things like Google Discover.”

Historical Significance

This webinar, recorded in November 2020, documents the collaboration between Jason Barnard (Kalicube) and Andrea Volpini (WordLift) that became foundational to Kalicube’s methodology. It contains the earliest detailed articulation of the three-step entity home process (state → corroborate → point), the entity home as “the single most important thing,” the child education metaphor extended to specific trust levels (teacher = Wikipedia, parents = Wikidata, baker = LinkedIn), the distinction between understanding and confidence, and Andrea Volpini’s linked data principles with persistent URIs. These concepts form core elements of The Kalicube Framework (TKF) and The Kalicube Process™ (TKP), and are referenced in Kalicube’s INPI patent filings.

---

Full Transcript: How to Get and Control a Knowledge Panel for Your Brand or Yourself (SemRush)

---

Paige Hobart: Hi everybody, welcome to “How to Get and Control a Knowledge Panel for Your Brand or Yourself.” I’m joined today by the wonderful Andrea and Jason, who I’ll get to introduce themselves in a moment. I just want to give us a little bit of time for people to join us on the webinar. So give us a hi, let us know where you’re from. I know it’s Thanksgiving today, I think, in the US - so I’ll be very impressed if we’ve got any Americans with us today, clearly avoiding their families.

Give us a hi, let us know where you’re from. And in fact, Paige wanted to hear herself back to make sure - I did the classic thing of not using my YouTube. But we are lovely. So you can see lots of people. Someone’s got coffee at the ready as well, so it must be a very different time zone for them.

But yes, welcome, welcome to our Knowledge Panels talk. Very excited to be joined by these guys. They are so, so knowledgeable on Knowledge Panels. I will admit that I will be asking all of the basic questions because I’m not an expert in this field, but very excited by the content to share with you today. I’m really hoping there’ll be a lot of actionable insights for you.

Jason, do you want to open up your deck and then we can get the proper intros on the way? Because I know we’ve got a few slides on intros.

Jason Barnard: Yeah, I’d love to open up your deck - sounds like we’re on a kind of cruise ship and we’re about to take a dip into the sunbathing session. Wonderful. Thank you very much, Paige. I’m really happy you’re here. I think you’re underplaying your knowledgeability and insights into Knowledge Panels by saying what you just said. We’ve had several discussions with Andrea and yourself, and you’ve helped me at least learn a lot by looking into my database and saying, “What’s that? Why is that there? Shouldn’t that number be smaller?”

Well, I’m Jason Barnard. Andrea and me are going to give this talk. We’re going to share the slide deck. This should be a lot of fun. I’ll present myself. I’m The Brand SERP Guy®. So I like to present myself using my Brand SERP. A Brand SERP is what appears when somebody searches your exact-match brand name or your personal name. And I am a great advocate of controlling it. And as you can see here, mine looks really kind of quite cool - it’s got lots of image elements all over it. It’s not just the blue links. Looking good is one thing. Controlling it so it looks good and is accurate and positive is the next step up.

And I can actually tell my life story through my Brand SERP. I used to make cartoons. I was a blue dog. My wife was a yellow koala. And that’s a bit mad, but it was a lot of fun - and it was on TV, which was great. ITV International produced it.

Now I’ve got the podcast, and I’ve just learned that podcasts can now be put in IMDb, which we’ll look at later on. So I’m getting more stuff into IMDb. That podcast has been a great boon for me in terms of learning all about this stuff. I publish all over the place - wherever anybody will let me. Danny Goodwin at Search Engine Journal lets me write and supports me, and he’s very kind about that. And I talk all over the world, giving talks about Knowledge Graphs and Brand SERPs and SERP features, which Paige loves too.

That’s about it, really. Andrea - your Brand SERP, how does that look?

Andrea Volpini: Yeah, it’s so far so good. I mean, there is still some work that we can do to optimise it. But so my name is Andrea. I am the CEO of WordLift. I focus on automating everything that Jason does. So I’m a Jason automator. And excited to get this started.

Paige Hobart: Brilliant. And Paige - thank you. I feel like my Brand SERP’s a little bit embarrassing compared to you two, but this is where we’re starting from. I’m sure there’s a lot of people in the chat with us today that might have a little bit about them. I’ve got another person with my name who’s a similar age and blonde - which is fine for her, but I don’t want her in my Brand SERP. So I need to make mine as cool as possible, basically. And I’m really hoping these guys help me out today.

So I’m an SEO Team Director at an agency called We Are Roast. And yeah, really happy to get this going.

Jason Barnard: Right. OK. In fact, it says just “Roast” instead of “We Are Roast” - so, and it was me who made the slide, so that’s my fault and I do apologise.

Paige Hobart: Right. What’s the plan today? What is the plan? So I made the guys really try and hone in on all these actionable insights today, so I really hope it’s beneficial. So we’re gonna do: What is a Knowledge Panel? One thing that me, Jason, and Andrea complain about is the language across SEO - we don’t always call the same things the same thing, which really bothers me. So like Knowledge Graphs, Knowledge Panels, knowledge things - what is a Knowledge Panel? What do they show? Who has them? Where do they come from? And how do you get one for yourself or for your brand - which I think is what everybody’s here and really excited about. Relatively easy ways to get into the Knowledge Panel and Knowledge Graph. How do you control the contents of your Knowledge Panel? And the beyond - so we’ll get into some really exciting 2021 and beyond topics at the end as well.

Jason Barnard: Brilliant stuff. Yeah, I mean, I think those first - I mean, we all want to know how to get one and how to control it. But in fact, you have to understand where it’s coming from. Like a lot of things, if you don’t understand the basics - if you don’t understand where this is appearing from, where Google’s getting this information - you’re never gonna get one or control it.

So I’m gonna start: what is a Knowledge Panel? I do all the simple stuff at the beginning. Andrea does all the complicated stuff at the end.

What is a Knowledge Panel? There you go. On a Brand SERP - basically I talked about Brand SERPs - on the left-hand side you’ve got Microsoft’s Brand SERP, on the right-hand side you’ve got my personal Brand SERP. And I’m trying to big myself up by saying I look almost as good as Microsoft. But what we’re focusing on today is the Knowledge Panel - that part on the right. What we call the right rail. So if you look on desktop, it’s everything on the right.

And if I like to consider it this way: on the left, Google is recommending what it thinks the best answer or the best solution to your question or your problem is. On the right-hand side, it’s saying: this is fact. This is stuff we are sure about. Obviously sometimes it gets it wrong. But left-hand side - advice. Right-hand side - fact.

So getting that Knowledge Panel means you have to convince Google that it has fully understood facts about yourself or your brand.

And here I’ve got some figures out. I had loads and loads of data - we’ll see that in a moment. And in fact I just boiled it down to this: 44.8 percent of brands have Knowledge Panels. 36.6 percent of people have Knowledge Panels. This is from my data set, and the data set is 70,000 brands and people. So it’s fairly representative. I’ve made a really big effort to make it representative - to have big brands, small brands, medium brands, famous people, unfamous people. Paige being in there as well - I’m tracking you, Paige, I’m afraid. Andrea’s in there. William Shatner’s in there - so incredibly famous guy from Star Trek who’s obviously got this amazing Knowledge Panel.

And now we’re going to look at exactly what Google puts in there. And it depends on what it’s understood.

If you look at Microsoft - that’s a lot of information. Google’s saying, “I know all of this stuff about Microsoft, and I’m confident enough to put it up there and show it as fact.”

Then we’ve got WordLift, which is Andrea’s company. Less facts, but still quite a lot of facts. Interestingly, the profiles there - the social profiles - are very important, because then people can go and obviously interact with your business. And then “People Also Search For” - are related entities. Entities that Google thinks are related to WordLift. So SEMrush is in there, which is really quite groovy - it must be because of this webinar.

SEMrush - they’ve got a lovely Knowledge Panel. And you can see there’s lots of small pieces of information, like the founder, the founded date, who the CEO is. That’s information that Google’s collected from around the web and has become confident in that information and is presenting it to us.

And then we’ve got the smallest Knowledge Panel of all, which is Kalicube - which is my company. But as you can see, the point here isn’t so much to say, “Oh look, WordLift is loads better than Kalicube, SEMrush is better than Kalicube.” It’s that Google’s understood more and is more confident in what it’s understood, and is willing to stick its neck out for them more than it is for Kalicube - because I haven’t explained it very well.

And then we look at people. On the left-hand side, we’ve got JoJo Siwa, who I had never heard of until recently - we’ll see why. She’s got a wonderful Knowledge Panel. I was going to use William Shatner, because I love his, but Siwa is better, because we’ll see in a moment.

Rand Fishkin in the middle there - we all know Rand. Google’s obviously understood quite a lot about him. And it’s got “People Also Search For” down at the bottom, which are other people within the same sphere - people related to him in some way. There’s his wife there. There’s also Cindy Krum, who’s another SEO person. So it’s pulling relationships together for us.

Then the next one is Daddy Koala, who’s a yellow koala in the cartoon series I created. And I’ve been experimenting to see if I can push all of this information into Google. So I’ve taken this family - Daddy Koala, Mummy Koala, Grandpa Koala, Grandma Koala, Boowa, Kwala - these two families, dogs and koalas, and explained to Google who is who, who is related to who. For example, Daddy Koala’s significant other is Mummy Koala - which is genius naming of characters, if I may say so myself.

And then on the right-hand side, we’ve got Léonardo Joe Barnard, who’s my daughter. I do experiments on anybody who will let me - I do ask permission. And we got this Knowledge Panel, which is actually pretty good. It only took me two months. Two months to get from nothing - Google didn’t know who she was - to Google now showing, when you search for her name on Google, knowing who her grandmother is, who her parents are, and a description which is pulled from my site. So I have control over what Google shows about my own daughter.

And here we have the interesting thing - down at the bottom you can see “Siwa.” And it’s because my daughter played the role of the blue dog called Siwa, which is why I found out about JoJo Siwa in the first place. And she’s getting in the way of Google understanding who this blue dog is. So I would really like it if JoJo Siwa would stop singing and dancing for just a couple of years so I can build up the blue dog.

And you can see there - Rand Fishkin. The reason I put that in there: it’s the same person, but a different Knowledge Panel. Google is confused. Google thinks there are two Rand Fishkins - one of whom was in “SEO: The Movie,” and the other of whom writes books. But it’s in fact the same Rand Fishkin. And that’s an important part of what we need to do with Knowledge Panels - to disambiguate and prevent Google from duplicating in this manner. Because that indicates misunderstanding.

And coming soon - that was all the data. And Andrea made me take it out. He said, “There’s too much data, it’s not interesting, nobody’s interested - write an article instead.” So I’m going to write an article instead, and you will be able to find that soon.

Where do Knowledge Panels come from? The Knowledge Graph. On the left-hand side, you can search in the Knowledge Graph using the Kalicube tool, Knowledge Graph Explorer. If you look at Knowledge Graph Explorer, you’ll find that. And that’s Google’s Knowledge Graph - that’s explicitly full of all the information it has fully understood.

And that’s a great question from Rivas Capital: what’s the difference between Google My Business and the Knowledge Panel?

Google My Business is a business listing where you provide information and Google trusts you on your own good word. The Knowledge Graph is information it has understood on its own, through corroborating - we’ll see that in a moment - from third-party independent sources.

So Google My Business is simply a business listing. It isn’t fact. It isn’t something Google has understood. It’s something you have told Google, and Google is willing to push it out there as a business listing.

But you’ll see on the right-hand side, we’ve got Google Books, we’ve got YouTube, we’ve got Google Podcasts, and we’ve got Google My Business. Sometimes Google will feed things directly into Knowledge Panels without going through the Knowledge Graph at all. And these are examples of Google-owned properties that will sometimes feed a Knowledge Panel. But when I say “Google My Business,” I don’t mean the Google My Business panel - that’s just the business listing. I repeat: it’s just that Google My Business will sometimes be able to feed directly into Knowledge Panels.

But indirectly, there are millions and millions and millions of sources. Google isn’t saying everything has to come from a Google property. It looks in Wikipedia, Wikidata, IMDb, LinkedIn, Biography.com - whatever that is - Discogs, Financial Times, all sorts of sites all over the world. We’ll look at some of those later on.

So you have to remember: the Knowledge Graph isn’t just Wikipedia or Wikidata. And it isn’t just Google taking information from its own properties and showing it in Google as fact. It’s actually corroborating all this information, learning from the massive information there is online. And Andrea is going to talk about how complicated that can be.

Now - how do you get a Knowledge Panel? Far and away the most important thing is that little globe icon - a little globe icon with your site next to it. And that’s the single most important thing in this entire presentation, as we’ll see in a moment. Give your entity a home.

Now I’ve blown this one up to really big so you can see it. It’s that little globe icon. It’s not the description. It’s the globe icon with the domain name - or even a page, as in the case of my daughter. That says: this is where Google considers the home of the entity is. And that is the place it will look for information about the person, the brand, the event, whatever kind of entity it might be - and then go out and corroborate it, as we will see.

So once you’ve convinced Google that your site is the home - a page on your site is the home for your entity - then you can state a fact on that home. It can be when it was founded, who the founder was, the area served. It can be your social profiles - that was a big one not very long ago. Anything you want to inform Google about.

Then you need to get significant corroboration from multiple independent authoritative sources. And when I say authoritative sources, I mean authoritative within your industry. So Wikipedia, Wikidata are obviously important for everybody if you can get them. But if you can’t get them, that dog on the right-hand side represents the dog association down the street from me in Paris. They are authoritative for dogs in Paris. So they would be a great source of corroboration, and Google will use them as a reliable independent authoritative source - in the case of a dog shop.

Down there you’ve got a link to Kalicube, which actually has a list of the ones that I have found. I found five or six thousand there, that I break down into different industries and different regions. Oh, and that’s the second thing: authoritative independent sources, broken down both by your industry and by your region. So I said the Paris dog - not dogs in general.

And then you can helpfully point to them from the home. Now I’ve got my home there, which is jasonbarnard.com. And basically, using schema markup - which Andrea’s going to describe in a moment - you can just say: “OK, I’ve given you this piece of information - step one, stated a fact. I’ve managed to get it corroborated by multiple independent authoritative sources. Here they are, Google - point to them.” Show Google where it can get that corroborative information. It will believe you, and it will give you a Knowledge Panel - or change information, or add information to a Knowledge Panel once you already have one.

Andrea.

Andrea Volpini: Yeah. So how do we get the information in the hands of Google? We do this by building structured data on our website. And so every web content that we publish can be augmented with the use of structured data. So here we have a blog post, and the blog post is mentioning a few elements. And we do this by using this form of metadata that allows the search engine to understand who we are and what we’re talking about. And that’s exactly the process.

Moving forward, I think it’s super important to understand that Google My Business is an entry point for the Knowledge Graph. And if you are a business owner, that’s where you should probably start. So there is nothing wrong with starting with Google My Business. This is actually my suggestion. But this is a facet of your entity - this is what Google will perceive about your business, especially within the context of local queries.

So we’ve seen, in the case of WordLift, the representation of the Knowledge Panel that Jason can get from his computer. But then if I do the same query here in Rome, then I would probably get the local panel first, before the display of the entity. So don’t think of Google My Business as something else. Google My Business is the way in which Google is rendering everything that is locally focused and related to the entity. So the two things are connected.

And I read another question which was interesting, about what are the biggest mistakes that people make. I think one of the biggest mistakes is really to get data wrong. As Jason was saying, when you start sharing data in structured form, you might not be consistent. And that’s what creates confusion. And then it takes a lot of time to reconcile stuff.

And Google My Business is super important - because if you close, for instance, your Google My Business listing and then you reopen, then you might confuse the Knowledge Graph. And then you might get different results. So that’s one issue. I think we can move on.

Jason Barnard: Brilliant. Sorry - I just love the idea of confusing the Knowledge Graph. And it’s incredibly important, and it’s underestimated. I didn’t realise how confused the Knowledge Graph gets very easily with what I think is completely consistent information. And the Rand Fishkin example is a good example of it - it now thinks there are two Rand Fishkins, and we now have to explain to it that there aren’t.

Paige Hobart: Yeah, yeah. And I love - oh, sorry. I was gonna say that was a great question from Azim Digital, and then we can get to the next section. Yeah, around the biggest mistakes you’ve seen people making trying to get a Knowledge Panel. There we go. Right, well -

Jason Barnard: One of the biggest ones is messing with your Wikipedia page too much, which is what I did. Got deleted. Created me lots of problems. Another mistake is trying to get a Wikipedia page when you don’t deserve one. Wikipedia is an encyclopedia for people. The aim of Wikipedia is to show information about notable people, notable entities, notable companies - and not everything. Whereas Google isn’t being difficult about it - saying, “I want to understand everything.”

So for example, Paige, I’ll use you as an example. You don’t have a Knowledge Panel. Trying to get a Wikipedia page would be a pointless exercise. A, because you wouldn’t get one. B, if you did get one, they would just delete it. C, it isn’t actually useful to the human community.

Paige Hobart: Yeah, that is useful. I tried a Wikidata entry and they deleted me because I wasn’t notable enough. And I was like, “Yeah, OK.” But the thing is -

Jason Barnard: You don’t need one to get a Knowledge Panel. Google doesn’t care that you’re notable or not. It just wants to understand. And I think if you bear that in mind - you’re thinking, “I’m just educating a child. I’m educating a child.” And that’s why I love this.

I posted out my wonderful structured data for the 53-song album I made in 2007. And I could actually do a lot more than that - that’s quite a limited example. But you can see how quickly this builds up, from one music album that I made with 53 songs, how many different elements there are related to it.

And then - Jordan B. Peterson, ebooks - whoever that might be. I love that: a child who responds to every answer to her previous question with “why?” Except you keep answering forever. And that’s Google. You show it the album and it says, “OK, what were the songs?” You show the songs, it says, “Who wrote them?” And then you tell it who those people are. “What else do they do? Where do they live? What’s their mother called? What’s their father called? What are their children called?”

And the more you do it, the more you realise: Google is an insatiable child who just wants to learn and learn and learn, and who has the capacity to remember everything.

Paige Hobart: Yeah, yeah. I think they actually said that Google’s artificial intelligence has roughly the intelligence of a six-year-old.

Andrea Volpini: So if you think about that when you’re trying to -

Jason Barnard: I think in some cases, I mean, from others probably it’s a lot wiser than ourselves. But I think - when you hear “the intelligence of a six-year-old,” people tend to think, “Oh, not very intelligent then.” But in fact, six-year-olds are incredibly intelligent, and they’re learning very, very fast. They’re learning much, much faster than we truly appreciate.

Language is a great example - how quickly a child will pick up a language. How easily they digest all this information. How much of it sticks within their brain. And the stuff you want to stick in their brain - like “green beans are good for you” - is stuff that they don’t keep in their brain because they’re not interested. It doesn’t suit their purposes. Difference here is that Google will remember that green beans are good for it.

So - the easy entry. We talked about Wikipedia earlier on. It’s the obvious entry, and a lot of people go for that. And I would suggest, unless you are truly notable, it’s a waste of time. It isn’t a question of “how long will it take me?” - it’s “how long will it take me to waste my money and time?” Having done it, it will then just get deleted and there’s no point.

Wikidata is kind of - you said you weren’t notable enough. You don’t need to be incredibly notable. Wikidata has a much lower entry bar. It has a lower entry bar that doesn’t really - Andrea is going to disagree. I’ll just finish this point and then - Google itself, as I said, just wants all the information.

Now, Andrea, why am I wrong about Wikidata?

Andrea Volpini: No, no, no - you’re totally right. You just broke the connection for a second. So, Wikidata has a lower entry bar - you’re correct.

No, no, no. I believe that Wikipedia and Wikidata serve the purpose of caring for global knowledge. So as you said, you have to be notable enough to become encyclopaedic. So what is the secret value of an individual? I mean, if they have done something that’s notable enough for others to know about it.

So you need to have enough references around that justify the presence on the general and universal Wikipedia. That’s the thing. I would suggest as well the idea that somebody will spontaneously wake up one day and think, “Oh, I wonder about that particular person or that particular brand” - and spontaneously search for you. If that isn’t going to happen - you know, from a band that you used to love in the ’90s or the year 2000 or whatever, and you suddenly wake up and go, “I wonder what happened to the singer from the Kinks.” There’s somebody who’s notable, who will get that spontaneous search on Wikipedia. And you would go to Wikipedia and look them up, just like you would open up an encyclopaedia.

Paige Hobart: Yeah. Oh, go ahead. There is a good question - I think there is a good question. That’s a good one.

Jason Barnard: I will - in fact, no. Because if you are notable, then making sure that you have a decent Wikipedia entry is incredibly important. Because Google still believes Wikipedia pretty much on its own good word. So if you are a notable brand, it’s a good idea to get involved in the Wikipedia page. Not by editing it yourself - which is the mistake I made - but by talking to the editors, by talking to the administrators, to try to get the facts straight, to try to make sure that that Wikipedia entry actually represents you.

And if you don’t want to go down the Wikipedia route because you don’t feel you’re notable enough, then Crunchbase, LinkedIn, Bloomberg, Google Books, Financial Times, Irish Times - and the dog shop down the road. They’re all great places to get yourself corroborated.

That figure’s quite old, actually - 1,650. Actually, it’s now nearer 3,000 different sources that Google has cited in Knowledge Panels, which we can assume are reasonably well trusted. Once again, there’s that list. But I’m only tracking 70,000 brands. There must be 20 million, billion, trillion brands in the world that I could be tracking. So it really is only the tip of the iceberg.

Paige Hobart: Yeah, we’ve got an interesting question from G. Hill around notability - do you need a certain amount of links?

Andrea Volpini: It’s really based on the knowledge domain where you exercise your presence or your activity. So it’s very relative. I mean, we are notable probably within the small context of Knowledge Panels and Brand SEO. I don’t even want to say SEO. But so - being notable is within the context of potential queries that cover a knowledge domain.

So that’s Wikipedia and Wikidata - it’s human beings that are deciding if they’re notable or not. Right? Where is the end? Well, there are some rules that the community follows. And the rules of notability are very well defined within the code of Wikipedia. So again, you have to have enough evidence and sources that talk about you, so that generally people will look for you and then they will find an answer in Wikipedia.

The reason that we SEOs spend a lot of time talking about Wikipedia is that Wikipedia allows us to understand also the way in which the Knowledge Graph is shaped - because that’s how a lot of the facts have been seeded in the beginning, and continuously are seeded. So the reason we SEOs spend time on Wikipedia and Wikidata is that, of course, we want to understand the way in which knowledge is organised inside the Knowledge Graph.

And so one of the most important aspects - as we say that Google My Business is a facet. So it’s one angle at which we can look at, for instance, an organisation. And that angle provides everything that is related to directions and address information and hours in which your service is available and service areas.

So what’s interesting is that you want to find an entry point. So Google My Business can be an entry point. But you can also kind of attach yourself to something that is notable. So if you have written a book, probably a book is already well structured as an item. And so the Knowledge Graph might already have evidence of the book. And then you, as an author, might kind of derive the authority from the book.

And that’s how you can also get into the Knowledge Panel. So if there is something that you’ve done that’s already inside the Knowledge Graph, then you can use structured data to let Google know that you are the contributor or you are the author of that book. And that’s exactly what Jason does with his podcast, right.

Jason Barnard: Well, in fact, I started doing it with the blue dog and the yellow koala, because it was produced by ITV International. It was in IMDb. It was already semi-present in the Knowledge Graph.

And then the group I was playing in before. And with the podcast, I’ve been working with WordLift - which has been a wonderful experience - of saying: I have a podcast series. The series has multiple episodes. 151 episodes - I counted them this morning. And each episode has a topic. Each episode has a guest. Each episode has a producer - Kalicube. Each episode has a director - myself.

And pushing this information to Google. So, for example, Rand Fishkin was on the podcast, and he is in the Knowledge Graph. So it’s relatively easy for me to say, “Oh, look - hey, Google - he was on this podcast episode. That podcast episode was hosted by me.” I’m actually showing the relationship between myself and Rand Fishkin.

Same with Andrea Volpini. And we were talking with Andrea about the link tax at one point, so there’s a relationship between myself and the link tax - which is going to be in the Knowledge Graph too. Remember that the Knowledge Graph also contains concepts, like economics, the link tax. So it’s not just things, people, brands, places - it’s also concepts.

Sorry, go ahead, and explain - entities have facets.

Andrea Volpini: Yeah, so, I mean, these concepts are in most cases articles in Wikipedia, again. So when you kind of describe what you do, you might reference concepts that are inside Wikipedia and that Google can understand. And that’s what we do when we, for instance, create structured data of an article - we try to encode the meanings of the article by attaching the article and the metadata of the article with the metadata of the concept.

And so we kind of have to think of the Knowledge Graph as, you know, like a human brain, where we want to feed information. And we can feed information in different ways. We can use structured forms - like structured data on your site, or entries on Wikidata, or entries on Crunchbase. Crunchbase is a marvellous structured database for companies and startups. So you can curate your presence on Crunchbase, and Google sees Crunchbase not as a website but as a database.

So the more relationships you build inside this large database that Google uses, and the more facets you kind of expand and develop about your entity - and I have this thing on my Brand SERP from an experiment called Google Cameos. And these videos are built with questions that Google makes to people like me, and I answer these questions. And these questions are automatically generated by looking at the things that I do the most.

So all the questions that I receive are related to startups and starting a business and, you know, that kind of boring stuff - because that’s kind of the facet that I was able to create.

But then if you look at Jason, then he’s a musician, and then he’s a digital marketer, and then he’s a cartoon maker. And so he has multiple facets. And then he can attract different types of queries.

So the question is: what type of queries are you looking for, and how do you want to develop your Knowledge Graph presence?

Jason Barnard: I think that’s a brilliant point. Because when you said, “I’m a cartoon maker and a musician and a digital marketer” - but right now I’m a digital marketer. And I have a problem with the confusion that Google isn’t sure which one it should be prioritising for me. And at the moment, it’s prioritising musician - which slackers my ego no end, but it doesn’t actually help with my day-to-day existence.

And I’d like to just come back to one thing - thinking about educating a child. It is: how do we learn? We learn by attaching new information to information we have already understood. And if we, as a child, get information and it says, “OK, I can relate that to something I already know” - it will stick. If we can’t, it won’t.

And the other one is: if you think about a child, it doesn’t just believe the man in the street it happens to have met, on everything he says or she says. The teacher would be the equivalent of Wikipedia - the trusted source that’s always given true information, and that you trust. The parents would be - oh, let’s say Wikidata. And then the evening school teacher would be Crunchbase. Obviously I’m pushing this a bit too far and it’s beginning to get a bit silly. But the baker down the road is LinkedIn, let’s say.

But all of these sources that are more or less trusted - the child needs to hear this information from these multiple sources that are more or less trusted. And the more they hear it, the better they remember it. And the more it’s related to information they already know, the more it will stick.

Paige Hobart: Thank you, Rob T. I think I overdid it. Your siblings are probably like The Onion, because everything they tell you is wrong.

Just, Andrea - back to something you were talking about on Crunchbase. We’ve got a question from Jeannie Hill about potential resources on how to optimise or make the most of your Crunchbase page. Do you have anything?

Andrea Volpini: So my tip is really as simple as building relationships. So Crunchbase, like Wikipedia or like LinkedIn - the more you build relationships, the more you provide data points that the Knowledge Graph is looking for.

OK, so - like a child, the relationship is what creates the information. So on Crunchbase, try to tell the story of the different companies that you’ve been through, and the roles that you’ve been through. And try to articulate the information about the companies and the partners and the investment rounds. Again, it’s a relationship business. The more relationships you make, the more these relationships are consistent with what you’re looking for, the more effective you become in building a rich Knowledge Panel.

Jason Barnard: Yeah. And I’d like to point - I mean, the idea of relationships is: strong, close, long-lasting. Yeah, that’s what you’re looking for. Strong and close, particularly.

I mean, my relationship - I’ve mentioned Rand Fishkin. It’s actually quite distant, so it doesn’t help me a great deal. I work a great deal with Andrea. That relationship is obviously something that is closer and longer-lasting, and it is potentially better for me - especially as Andrea is so good at getting himself attached to entities that I’m interested in, like voice search and - what else is it? - Knowledge Graphs, JSON-LD.

Sorry - WordLift itself. If you look at WordLift in the Knowledge Graph, it’s associated with lots of topics, which I find stunningly interesting. Topics like voice search, JSON-LD, schema markup - topics that WordLift is specialising in. Now, that’s incredibly strong. I haven’t seen another brand with that level of connectivity to topics.

And if we talk about the topical layer - which we probably better not do today because we’ll never get through it - it’s the future of search. Google understanding who you are, what you do, who your audience is, and what are your specialist topics.

Andrea Volpini: Yeah. So that’s an example of how the data gets into the graph. OK, so there are multiple ways in which information gets into the Knowledge Graph. But one way is that Google will use the recent advancements in natural language processing to synthesise information that it gets across multiple news outlets.

So here, when Eddie Van Halen died, of course they were able to group the information about the event on Google News. And they can kind of compact a lot of articles into one single sentence.

But if we move to the next slide, we see that they didn’t ask The Onion - which was Paige’s example - which would have said he’s still alive and living in Brighton.

But so this information that’s been constructed from the summarisation of text coming from different news outlets then becomes a fact - because there are multiple sources that are stating the same information. And then that becomes a triple. That becomes a fact inside the Knowledge Graph.

So this is one way in which you can build data for an entity. Of course, it takes a lot of effort, because you have to publish a lot of information in order to create a triple out of these. But just to give you the idea of the diverse sources of information that get into the graph. And Google My Business is one source. And Wikipedia is another source.

So when Daydays is asking about, “I had a beautiful Knowledge Panel and then finally the wiki, and now things are getting worse” - I totally understand. Because of course, now Wikipedia is there, and Google is trying to understand if it can reconcile the information that it had before with the information that’s been obtained from Wikipedia.

And Jason here has done a pretty tremendous experiment, when his Wikipedia page was removed - so the Knowledge Panel completely changed all of a sudden.

Jason Barnard: Yeah. Well, in fact, it was three pages - because I had the group I was in in the ’90s, the blue dog and the yellow koala cartoon, and myself. And basically the editors went through them one by one and just deleted them within the space of two weeks, because I’d messed with them too much. So fair enough.

It upset my ego, I think, more than it did anything else. And with the three different entities, I did three different experiments.

With myself - the Knowledge Panel disappeared. And I had to rebuild it, because I moved the home. I moved my home - i.e., all the information about me - I moved to a new page on the same site. And that’s important. The whole thing just disappeared. Google lost its crutch.

Whereas with the other two, I left the information where it was. Google had a crutch to fall back on, and it kept the information. And interestingly, the confidence score in the Knowledge Graph itself - which you can look up on Kalicube - stayed the same for one of them and actually went up for the other one.

And I’ve seen two other examples - Wikidata was another example. Somebody who’s doing an experiment, his Wikidata page got deleted, and his confidence score has gone up and up. And his Knowledge Panel suffered a little bit but not a great deal.

So I think it’s important to remember that once Google thinks it’s understood, as long as you don’t take away its crutches - i.e., the supporting information - it will stick with you. And you won’t have this problem of a diminished Knowledge Panel. Or you will have less of a problem.

Control and influence - so definitely don’t delete the home. Don’t delete the home. No. Don’t delete that.

Paige Hobart: We’ve got a question actually on the home concept. I think that resonated with quite a lot of people in the chat. Giving your entity a home - how do you establish where your home is? And like, what’s the best kind of home for your entity? Especially - you mentioned your daughter only has a page, rather than a whole site.

Jason Barnard: That’s yet another part of yet another one of my silly experiments. I mean, it’s on my own site. But because she’s related to me, that home seems reasonable to Google. It’s understood that my site, jasonbarnard.com, is my home. Therefore, my daughter reasonably has a home which is one page on that site.

And when I said I moved my home for my own entity - I moved it from my homepage to an About Me page. And I should have either had it on the About Me page all the time or stuck with the homepage.

And the blue dog and yellow koala experiment, which I’m doing with WordLift, is - I’m pushing all this in because Google sees me as the creator of these characters. My site is therefore a reasonable home for them. But in fact, what’s interesting is that the original site where they’re still hosted is also considered a home for some of them. So Google’s kind of trying to make its mind up - am I the authority, or is the original site the authority? So in fact, you can steal the home, as it were.

And it’s an interesting kind of question - how do you make it understand what the home is? But we’re back to corroboration. The more things point to it - I mean, it’s link building, but it’s mentions. And it’s pointing to your site or this page on the site as being the place that is the reference.

And also, what I’ve noticed is: the more you tell the truth - it’s, once again, we’re back with kids. If you keep telling the truth to a child, it will believe you pretty much every time after that. But if you start lying to it - like parents do - like, “If you don’t go to bed, you won’t get any food for the next five days,” and then they feed them for five days. If you can keep telling the truth, you will build that home. You will build that trust. So keep telling the truth. Be honest. Be clear.

And schema markup also helps. I mean, Andrea knows much more about that than I do. But if you actually just state in schema markup “this is the home,” it will go out and check. Does LinkedIn say this? Does Crunchbase say this? Do all these sources point back to this place?

Now - to control it. First thing to do is claim it. When it appears, you claim it. You click on that button and you can claim it. Depending on whether or not it has found your home - which is that little globe icon - if you have that, you can claim it through Search Console. It’s really easy, very quick, absolutely beautiful.

If you don’t, you have to send lots of paperwork. And if you’re a person, it’s pictures of you logged into your social accounts and all sorts of strange things. Documents - you have to prove who you are. Which is fair enough. They don’t want to give this kind of control to somebody or a brand that isn’t the legal representative of that brand or that entity.

And so once you’ve claimed it, you can suggest an edit. And you click on the button, you say, “Actually, I’d like to add this information or change this information.” And I have a friend who wanted to add his shoe size. And he asked them, “Can you add my shoe size? I’m shoe size nine.” They added it. Brilliant. He was really pleased.

And that’s wonderful. But you have to - A, you’re asking people, so it takes a certain amount of time. B, you have to give supporting evidence. People won’t just believe you - like the child won’t believe you.

And sometimes they simply cannot change it. Things like what is the home - a human being can’t change it. What it says - “British-French musician” - I cannot get a human at Google to change that. Because it isn’t the humans who decide, it’s the machine who decides. And in certain things, like the “British-French musician” - it won’t treat me as a digital marketer or an author, which is what I want. The machine is the one who controls it, not the human beings.

And if I say to the human beings, “Can you please change - for example, I wrote this song, or I released that album, or I was in the Barking Dogs and I was in another group called Stanley the Counting Horse” - which is a very silly name for a group, but it was fun - if the machine thinks they’re wrong, if they’ve corrected it for me - the human being has corrected my Knowledge Panel for me - and the machine comes along and says, “Actually, that’s wrong,” it will just switch it back.

So if you cannot correct the piece of information or the pieces of information that are making for this misunderstanding, you will never get it changed long-term. The human beings can change certain information in the Knowledge Panel, but if you don’t correct the source of the mistake, the machine will just switch it back.

So basically, this is the home concept. I’ve got a home for my entity. I state a fact on the entity’s home. I get significant corroboration. And I helpfully point from the home to the corroboration - to prove what I’m saying. To say to Google, “Look - Bloomberg says this. IMDb says the same thing.”

And as Andrea said, the most important thing, once you’ve got your home, is to be consistent across the web. And it’s also, unfortunately, the most difficult thing for humans - to be consistent in everything they do and say.

So - taking things further. This is where I take a big step back and listen to Andrea.

Andrea Volpini: All right. So - structured data. Well, I think I can start from one of the questions: what type of structured data can we use for helping Google?

One thing for sure that we can use is the sameAs links. So these are the most basic property that you can use for connecting the entity that you have inside what Jason called your home, with the same entities on Crunchbase or on LinkedIn. And this is super important. So if you have a Wikipedia page, then you should immediately add a sameAs link from the structured data of your homepage - or whatever home page you have chosen - to the article on Wikipedia. And the same applies to your LinkedIn profile, your Twitter account. Because for a machine, it gets super simple to connect things up. So you want to build relationships. And in order to build relationships with structured data, you can use things like sameAs.

But at the same time, you can also use properties like “knows” - to let the search engine know that I am friends with Paige and friends with Jason. And then we can build this relationship within the structured data.

Now, one important aspect here is that - as we talk about the home and the problem of moving home (because maybe I have a page, but then all of a sudden I leave the agency, or I change business, or I change my site, my domain, and I move my home somewhere else) - now, this can be managed when I start to create the home for the entity.

So what we do with WordLift is that we try to publish structured data using this format called linked data. So we provide this unique ID - this URI, so-called - that is persistent across the web. So Jason has his own URI, which is something like data.wordlift.io/slash-whatever/slash-Jason-Barnard. And this is exactly the entity representation in the form of data that can be moved from one page to another.

So having a unique identifier inside your structured data will also allow you to have more flexibility when you change your website structure and you decide, “Hey, this entity is the same that is represented there.”

There’s also the consistency problem - it’s there when you start publishing all these structured data. You might have, for instance, a page for a product, but then the same product is mentioned in another page. And how does Google reconcile this? We can make this reconciliation simple by using this unique identifier.

So one thing that we can do to make a step forward is not only to use structured data, but to start thinking in terms of linked data - in terms of connecting data and having your own home for the data, not only for the web part. I know it’s not super simple to grasp in the beginning, but then it gets simpler.

So in this example - I kind of extended a little bit of the best practices from Jason into more kind of advanced ideas. Not only give your entity a home in the sense of a web page, but also in the sense of a web data home. OK? And your web data home can be the entity that you publish with WordLift, or can be the entity that you have published with other linked data platforms like Wikidata - or other sources that create machine-friendly information that can be digested by a search engine that is building a Knowledge Graph like Google.

So in this example here, once I’m starting to use these URIs - these unique identifiers for the entities - then I can start linking these identifiers with other identifiers. And by doing this, then I can make my information queryable.

So in this example, I’m making a query on Wikidata to see: where is the entity home for that person that is born in Leeds and who does digital marketing as his main occupation? And I get Jason as a response. So I’m running a query on Wikidata in order to get the entity home for its URI. And that’s, I think, the evolution - the way in which search engines need to think.

Jason Barnard: I think - sorry - what you just said is actually incredibly important. I mean, Bill Slawski uses the example: where does the phrase “throw the napalm on the forest” - or something - come from? And it knows the film. And what it’s doing is looking at these different aspects of an entity and being able to identify the entity from its aspects.

Who wears a red shirt and City glasses and has a white beard in the digital marketing industry? It’s Jason Barnard. All those aspects, all those properties that I have, make me uniquely identifiable.

And Jeannie Hill’s just asked a question about “knows” and sameAs. I mean, my experience with “knows” - I’ve tried it, I’ve done some experiments - it’s a very weak relationship. It’s very weak and very ambiguous. So it actually doesn’t help very much. “I know Andrea Volpini, but I work with Andrea Volpini” is a stronger relationship than just knowing him. And sameAs - sameAs is obviously saying it’s exactly the same thing. So that’s the strongest possible kind of identification.

What’s your thoughts on those, Andrea?

Andrea Volpini: I think they serve just different purposes. Because sameAs is an equivalence - I’m saying, “This is the equivalent entity there.” But “knows” - it’s something that, of course, it’s weak, in a way that it’s super simple to see that we work together. And for a search engine, getting everything that is related to us - published on the web - and creating the embeddings that are required for extracting the relationship is probably simpler than taking into account information like schema “knows.”

So yes, it’s pretty weak. I still use it. I always think that when we build data, we don’t build data just for the search engine - but I’m also building data for using this data myself, within the context of whatever application we can build. So for me, it’s helpful. But I believe that it doesn’t really make a big difference in SEO terms.

Jason Barnard: I love that. Because with the podcast, I’m trying to push the podcast out to different - I’m pushing it into IMDb, which now accepts podcasts, apparently. Well, apparently - I know it does because I’ve just done it. Yep.

And one of the problems I have is: all the people who’ve been on the podcast, I don’t have all the information in a database. So it’s actually quite slow. So what I’m now doing is building a separate database which is much more - which I can then leverage to push it out to all these different places on the web.

Andrea Volpini: But you do have the database already, because you have the graph. So the database that you build with the structured data with WordLift, it’s already accessible. And you can query the graph to send the data wherever you need to send it.

Jason Barnard: Very good point. Well done. We’ll do that later on. Thank you very much.

Andrea Volpini: Yeah. That’s a little bit of another evolution, but it just proves the idea that when we create data in structured form, there is a lot of usage of this data - and not only in the context of search engines.

So there is a very interesting project that is coming up within the context of the Wikimedia Foundation. And it’s called the Wiki of Functions, or the Abstract Wikipedia. So the idea is that with the data that is available inside Wikidata, with Abstract Wikipedia, new content will be built in different languages - for articles - to make a sort of universal version of an entity. So it’s kind of a language-independent way of constructing data - constructing sentences from data.

So try to think - when you add this information in structured data, when you create these relationships, you are actually providing the machine with a way to create text. So with Abstract Wikipedia, new pages will arise from the information that’s contained inside Wikidata and the information that’s captured across the different languages for that same entity.

Jason Barnard: Wonderful stuff. Yeah, that’s - you didn’t play the video, but it’s OK. I mean -

Andrea Volpini: Yeah. So just to give you - I mean, the video simplifies what I’m saying. That basically, with all these functions, different sentences are built so that we will have the same concepts in different languages. Because at the moment, we do have a lot of concepts inside the English version of Wikipedia, but maybe if you go on the Italian version, there is less information available. So with Abstract Wikipedia, these will be leveraged, and we will have this universal representation of knowledge.

And if you go to the next slide - that’s even kind of more fancy and more futuristic. Because this is how the function will look. So this is a function that will allow the machine to create a sentence - like the one that we see here, about “Succession” - but using the data. So the function is kind of a template that grabs the data and creates something that can be written in Italian, in English, in Spanish, in Arabic. So it’s kind of a universal grammar.

But why is it important in this context? Because it’s important that we understand that the data that we publish is gonna create the information that people will consume in the next coming years. Right?

Paige Hobart: Amazing. Andrea - so, Jarrel Ervins just said: is there a good reference list you can share for structured data or linked data identifiers? Don’t know if you’ve got a resource that you might be able to share, maybe on Twitter afterwards, if you’ve got a link.

Andrea Volpini: Yeah, yeah. I will. I’m actually writing up some articles about it. But yeah, there is a lot of stuff here. I will definitely - I mean, the Five Star Linked Data Principles from Tim Berners-Lee is the main reference. But I will share it on Twitter.

Paige Hobart: Fantastic. Yeah. And I appreciate there’s a lot of questions that we’re probably going to run out of time to answer. We’ll try and get a few more in. But if there’s any questions that we’ve missed, please just tweet at us - handles should be later on in the deck.

But just to wrap up this section. So, in conclusion: you need a Knowledge Panel. I certainly need a Knowledge Panel. I think that is my homework for today - to go and get myself a Knowledge Panel, and to give my entity a home, because it doesn’t really have one.

You can trigger yours. So I can trigger mine - let’s go away and do it. Let’s go and get all that corroboration, all those links, and just everything to make sure that we’re not “notable” but at least a fact in Google’s mind.

You can mostly control it once you’ve got it. And Google needs to understand a few things - like who you are, what you offer, whom you’re relevant to, who you’re linked to. Remember the guys talked about hanging nodes off other pieces of information.

And: communicate, corroborate, and convince. Oh, that was a mouthful, Jason - I’m glad you put that on there.

And this is the single most important weapon in your digital strategy. Think of your Brand SERP. How much traffic do you get on brand - for you as an individual, if people are searching for you, or for your business? And also -

Jason Barnard: I mean, the Knowledge Panel indicates that Google has understood who you are, what you do, and who your audience is. And if it’s going to answer people’s questions, if it’s going to provide solutions - and that solution is going to be you, that answer is going to be you - it has to have understood who you are, what you do, and who your audience is.

And the Knowledge Panel - and I think that’s kind of a nice thing that we haven’t really said - it’s a representation of Google’s understanding of you as a brand, or you as a person. And in that sense, it’s the single most important thing in your entire SEO strategy.

Paige Hobart: We’re sorry - we’ll do the questions. We can - I think we can overrun by five minutes, and Anton won’t mind. I’ll look out for an angry Anton in our private chat. But yeah, I’m happy to run by five minutes. So if we skim through the next slides and then just spend however much time we’ve got left to do the Q&A.

This is just a slide on me - link there to do with SERP features. There’s a few of you that have been chatting about Google My Business versus Knowledge Panels. They are two different things, for the most part. We won’t talk about the hybrid ones, Jason. They are generally two different SERP features - until Google just decides they’re not and they’re going to smoosh them together.

Jason Barnard: Well, I like, Paige, that you know me so well already. That as soon as you say something like that, I’m going to go, “Oh, what about that time?” You’re like, “I know - hybrid SERP features.” They’re the bane of my existence at the moment. But I love it, and it’s all different.

But yeah - so Google My Business, different to Knowledge Panels. And triggered by two different data sets. Because Google My Business is that Maps information that you submit. And that article that you wrote is the reason we actually met, because we were doing a talk together - and it’s a brilliant article.

And with my PA - Paige, you’ve actually dug up an immense number of different features. Amazing stuff.

Andrea.

Andrea Volpini: Yeah. So that’s what we do. We’re actually providing a quick link to the Cyber Week offer, which should start tomorrow. But just for you - you can go on WordLift, use “Cyber Week,” and start building your Knowledge Graph or your structured data for your site.

Jason Barnard: Brilliant stuff. I’ll just end really quickly, and then we’re going to do Q&A. So do stick around, and you can ask your questions. Keep asking your questions, because we’ve actually got the right, now - thanks to Anton - to go over by 10 minutes.

I do Kalicube. It’s basically a research programme for me to try and understand Knowledge Panels and Brand SERPs. But I actually teach and consult on Knowledge Panels and Brand SERPs. I’ve got a set of courses for Knowledge Panels and Brand SERPs.

So my main job now is to research madly, work with Andrea to figure out how all this stuff works, and then teach everybody else. Give a person a fishing rod, and they will eat fish for the rest of their life - or something like that.

Wonderful. Brilliant. So - questions, Paige.

Paige Hobart: Yes. So I want to go all the way back to the beginning, because I was trying to make notes as we went along. So going back to the beginning - there’s a great one from Hider Ali: is a Knowledge Panel there to stay for a long time? Because we’ve seen some Google products and experiments coming and going after a few years.

I mean, my opinion on this is: this is Google’s database, so they’re definitely investing in this. But I’d love to hear what you guys think. I think Andrea’s probably got the best opinion on this.

Andrea Volpini: I think it’s going to stay. But how it’s going to stay - that’s the real question that we should answer. Because I think that the Knowledge Graph is a foundational element of modern search.

So now you can go on the Google Assistant and ask information about myself, Jason, WordLift, and SEMrush. And the reason you can ask this question to the Google Assistant using your voice is that there is a Knowledge Graph behind it.

So how the Knowledge Graph is then represented - nowadays we see this Knowledge Panel. But we already see the use of the Knowledge Graph in the context of voice search. Similarly, we see the impact of the Knowledge Graph in the context of query-less search. So Google Discover - it runs on these topical layers built on top of the entities of the Knowledge Graph. And if you don’t have an entity, it’s much less likely that your content will be pulled and recommended to users using Google Discover. Because -

Jason Barnard: So - which is the reason, Andrea, I’m desperately working to get this podcast into the Knowledge Graph. Because that’s my future. And I know it’s my future. And it’s my future of being discovered within the industry of digital marketing - especially with the query-less searches.

But one thing I do love about what you just said is: the Knowledge Graph is here to stay. The Knowledge Graph is the foundation of everything that Google is now doing, pretty much.

Knowledge Panels are the representation of them. So Knowledge Panels are important for your users, and they’re a representation of how well you’re doing in informing Google. So that’s two roles it’s playing.

But then you also have entity boxes - related entities. If you search for a brand name, you’ll see Microsoft, you’ll see Xbox, you’ll see Google, you’ll see Apple. And it knows which ones are related. And it needs to have you in the Knowledge Graph in order to appear there.

And I think, Paige, you were saying - a great way to appear on your competitor’s Brand SERP. If you’re in that carousel, then on your competitor’s Brand SERP, you’re appearing as a related search.

Paige Hobart: Amazing. I think I’ve also seen it playing a role in the new Google Shopping update. Because I was surprised to see a brand not appear in the brand dropdown, and I was like, “Why? This is clearly a product of our brand.” And yet it’s not considered a brand by Google.

Jason Barnard: Let’s dig into that, of course. So maybe here what we’re saying is: don’t bet your house on the representation of the Knowledge Graph in Google Search. Bet your house on the Knowledge Graph. And work with what you’ve got today for today - so that tomorrow, if you’re in the Knowledge Graph, you’re solidly in the Knowledge Graph. You’ll be able to leverage all the new things - like Google Discover, as Andrea was saying - whatever’s going to be happening in the future.

But right now, today, that representation shows you whether or not you’re doing a good job. And if you’re not - get to work. Well, that sounded horrible. Sorry. I don’t like talking to my daughter on a bad day.

Paige Hobart: Nice. A really interesting question from Harry Hawkins: would creating a Google My Business panel or account decrease your chances of getting a Knowledge Panel?

Jason Barnard: No. Really, no. Quite the opposite. Quite the opposite.

Andrea Volpini: But there is a problem. Because you have to help Google reconcile the entity from the Google My Business with the entity in the graph. And this can be done in different ways. So from the structured data that you’re building, you can use specific attributes in order to connect your entity with the Google My Business entity.

You can even contact the Google My Business support, as we’ve done with Jason on a case. And they’re pretty helpful in understanding the way in which things work. So they are able to reconcile your entity in the main Knowledge Graph with the entity coming from My Business.

Jason Barnard: I think that’s a really interesting point. I mean, the speed and the accuracy with which Google employees answer Google My Business questions - and that they spent hours on the phone with me for a stupid, tiny detail - and the Knowledge Panel, the fact they will correct things within a day - shows how important it is to Google. They’re investing in people - which is rare.

Oh, sorry - I shouldn’t have said that. They’re investing a lot in people to get these things done quickly. Because they know that if they start getting it wrong now, it’s going to get more and more wrong in the future. So they’re trying to nail it today. And they’re willing to invest in people - human beings - to double-check this information and help us to get the information right today. Because it’s the future. And they need the future to be solid.

Paige Hobart: Yeah. Particularly when we’re seeing those hybrid Knowledge Panels appearing, where it is the Knowledge Panel and Google My Business smooshed together. You told me not to talk about that - I know, I know. But yeah, so it shouldn’t impact getting one meaning not getting the other one.

A really interesting question from A Digital Vibe: what if there are a few service areas for a business? Should the business panel be built separately for these locations, or maybe like a main one?

Andrea Volpini: So, “service area” is something that you can model with structured data. So, first thing - on your pages, you should be able to tell the search engine that that particular branch is serving that particular area. And there is a specific property in schema that allows you to add this.

And then, of course, within the context of Google My Business, you can cover up the territory with the different branches within the same account. So you should do both - work centrally and then work locally. And clearly define the service area within the Google My Business and within the structured data for the page that represents each branch office.

Jason Barnard: Wonderful answer. Yeah, I think you’ve got to be really careful about trying to create multiple versions of yourself. Really very good point. Because you come back into this duplication.

I mean, I’ve started to say, “You’ve been duped” - which is a really cool way of saying there’s two of you in the Knowledge Graph. And it’s a phenomenal problem. I’ve got 350 experiments on the go at the moment in the Knowledge Graph - 350 entities that I’m trying to mess with and see what I can do. And one of my biggest problems is duplicates that keep appearing, however careful I am. They still appear, and then I have to go in and sort it out. And sorting out takes a lot of time.

So multiple service areas and pretending to be in multiple places all at the same time is liable to create more problems than you can deal with in the future.

Paige Hobart: Yeah. And I think we’ve got time for just a couple more. There’s a really interesting concept happening in the chat about DA. And about how to get them - Harry Hawkins to summarise: so, the best way to get your content featured in the Knowledge Panel is to have the content corroborated by high DA websites, as opposed to low DA websites? But of course, DA is an artificial metric.

Jason Barnard: Well, it’s not a Google metric. It’s something that Moz made up. And no - forget domain authority. Domain authority is links.

I mean, Singhal - don’t - not you, sorry, Paige. Yes, completely agree - who actually created and started the Knowledge Graph at Google and has now gone off to Amazon to work. He makes the great point: it’s trust-based knowledge.

And if you take an example of, you know, The Onion - which is your example - nobody trusts it. It’s complete nonsense. So it might have a lot of links, therefore it would have a high domain authority. But it’s rubbish as a trusted source for the Knowledge Graph.

So it’s a really bad idea to get anything on The Onion, even though it’s a high DA site. So forget about DA. And go for the dog association in the street in Paris where I live, if you’re a Parisian poodle shampooer.

It needs to be a geo-localised, authoritative source that Google trusts. And a lot of it is just common sense. I mean, the dog association is obviously more authoritative on poodle shampooing in Paris - I like this - than the butcher’s association. That was an unfortunate example. The butcher’s association for dogs in Paris.

Paige Hobart: One last question, I think, to wrap this up. And you’ll like this one, Jason. “How’s Kalicube?” - this is from Keith Patterson. Has Kalicube had much success with Knowledge Panels showing in search results for smaller brands, on a local or regional brand level?

Jason Barnard: Yes. Because Kalicube now - what I’m doing is tracking. I’ve found some very surprising ones. I think it was the Oxford Cheese Factory, which is actually just a small cheese shop in the middle of Oxford. For some reason I’m tracking them - and I don’t know why. I think they might have added themselves to the tracking tool.

They don’t - they have the Google My Business, obviously, but not the Knowledge Panel when you search their brand. But if you look in the Knowledge Graph, they have a little sprout. What I call “sprouts” - and it’s a tiny little Knowledge Panel that says “The Oxford Cheese Company.”

And so yes, even very small companies. And I would really, really like to reiterate: Wikipedia, Wikidata - good examples of places where they’re looking for notability. They’re looking for you to be important. They’re looking for you to be interesting to the rest of the world. Google doesn’t care how boring or uninteresting or unimportant you are. It just wants to understand. And it’s up to you to explain and convince.

Paige Hobart: Perfect, perfect, perfect. Oh, I was going to wrap it up, but Antoine sent us another one. The information on the right - how now - is it a Knowledge Panel or Google My Business information? How do you know if it’s Google My Business or a Knowledge Panel?

Jason Barnard: Google My Business has the little map. And the Knowledge Panel has the little three-pronged cherry icon thing. If it doesn’t have the three-pronged cherry icon, it isn’t a Knowledge Panel. It isn’t knowledge as such.

So - you’ll often see “Claim this Knowledge Panel” or something right at the bottom. You might, yeah. So if nobody’s claimed it, it will say “Claim it” at the bottom. But if they have claimed it, then you don’t get that claim button.

So just look for the little three-pronged icon, and that will tell you you’ve got a Knowledge Panel. But also, go along to Kalicube and look at the Knowledge Graph Explorer. Type your name in there, and you will see that you might actually have a Knowledge Panel, but it simply isn’t appearing on your Brand SERP or your personal Brand SERP.

And it isn’t because Google hasn’t understood you that it’s not showing - it’s because Google doesn’t think you’re relevant to the searcher. And “people” is a great example for that. It will show different people the Knowledge Panel for different people, depending on region.

And the example I was using the other week, when I was researching it, was - what was her name - Mary Tyler Moore. I was looking at “Mary Moore.” And there’s a Mary Moore in Australia who’s a judge. So she obviously shows up - big Knowledge Panel for her - when you’re in Australia. She doesn’t get a mention when you’re in America - it’s Mary Tyler Moore the actress. And in the UK, it’s Mary Moore the writer.

So geolocation actually shows you a great deal. Google will change a lot according to your location. Dawn Anderson brilliantly says it’s all about the probabilistic element of all this - what is the probability that you’re looking for that specific Mary Moore and not another Mary Moore?

And with brands - because brands tend to have unique names within their region, within their industry - that happens a lot less. So you will tend to see brands across the world, even for smaller brands, once they get that Knowledge Panel.

But if Google doesn’t have much to show, it won’t show a Knowledge Panel - if it doesn’t have much information. So yeah - build up that information. Make yourself look glorious. Get Google to be confident. Explain. Communicate.

Paige Hobart: Thank you, Harry Hawkins - that was a great webinar.

Jason Barnard: Thank you. I thought it was good. I just got a bit overexcited. I’m gonna have to calm down now.

Paige Hobart: Yeah, I love the fact that finally DA has a new meaning - Dog Association. Thanks to you, Jason.

Jason Barnard: Oh yes. Moz will be happy, won’t they?

Paige Hobart: Awesome. Thank you so much, guys. And thank you all for joining us. This has been such a great session. I hope you’ve learned a lot. The YouTube will be available to replay if you’ve missed anything or need to go back. So look out for that in your inboxes. Thank you so much.

Jason Barnard: Thank you, Paige. Thank you, Andrea. That was brilliant.

Andrea Volpini: Wonderful. Thank you. And saying goodbye.