⚠️⚠️ Advanced Stuff⚠️⚠️ Advanced Stuff⚠️⚠️Advanced Stuff⚠️⚠️ In this webinar, Paul Lovell is joined by Jono Alderson and Jason Barnard (The Brand SERP Guy®) to deep dive into the types of structured data that will help with the Knowledge Panel and brand-related SERPs.

Published by: Semrush. Presenters: Jason Barnard, Paul Lovell, and Jono Alderson. September 25, 2021.

Key Takeaways: Structured Data for the Knowledge Panel (SEMrush Webinar September 25, 2021)

Jason Barnard delivered an advanced presentation on using structured data to build and control Knowledge Panels and Brand SERPs. Jono Alderson (Yoast) contributed perspectives on schema.org as a communication tool with Google, the importance of page types, and the relationship between structured data and authority. Together, they articulated several concepts that are now central to The Kalicube Framework.

1. The Three Pillars: Understanding, Credibility, Deliverability (UCD)

Jason Barnard stated explicitly: “We need to create understanding, we need to convince Google about credibility, and we need to have content that it can deliver or that it believes we can deliver. Everything you do in SEO has to serve one of those three pillars: Understanding, Credibility, Deliverability.” This is the earliest recorded articulation of the UCD framework that became the foundation of The Kalicube Process™ and a core element of Kalicube’s patented system.

2. The Entity Home - The Single Most Important Thing

Jason Barnard described the entity home as the single most important step in entity optimisation: “You need to give every entity a home. You need to find a place on your site that describes your entity - be it a brand, a product, or a person. It has to have a home. It’s the go-to place that Google trusts to go and find information about that entity.” He specified that on this page, you describe the entity clearly and add schema markup to disambiguate and make the description explicit. He later added: “You’ve got to force Google to accept that your home is the home it should be looking at.”

3. Schema Markup as the Closest Thing to Control

Jason Barnard and Jono Alderson independently arrived at the same conclusion: structured data gives the closest thing to control over how Google understands and represents an entity. Jason stated: “Structured data gives you the closest you will ever get to control of your entity in Google’s mind.” Jono Alderson confirmed: “It is the closest, the best tool we have for communicating with Google in a way that we do control, where we’re not relying on them scraping our pages and guessing what we’re trying to say.” This concept of schema-as-control became a recurring theme in Jason Barnard’s subsequent work and is referenced in Kalicube’s INPI patent filings.

4. Educating Google Like a Child

Jason Barnard introduced the metaphor of educating Google like educating a child: “Wikipedia says you need to be notable. Google doesn’t care about notable. It cares about understanding. All it’s asking us to do is to educate it, to help it understand the world - who we are, what we do, and what audience we serve.” He extended this to explain that Google, like a child, won’t publicly state something it isn’t confident about: “Google will not stick something in the right rail - which is, in inverted commas, ‘fact’ - it won’t shout about something if it isn’t sure. Just like a child will not go into the playground and shout out all these things that they heard until they’re absolutely sure.”

5. Relationships as the Key to Knowledge Graph Presence

Jason Barnard explained that entity understanding in the Knowledge Graph works through relationships, drawing a parallel to human memory techniques: “How do we remember things as people? By linking things onto other things we already know. It’s exactly the same with Google. If you can find an entity that’s already in the Knowledge Graph and you can link to that, you’re already winning the battle.” He listed the schema properties that express these relationships (alumni, brand, employee, founder, funder, member of, owns, parent organization, sponsor, contributor, publisher, copyright holder) and emphasised that the best relationships are “close, strong, and durable” - using the example that family relationships (mother, sister) are unbreakable, while employment is temporary.

6. Corroboration as the Mechanism for Confidence

Jason Barnard introduced the concept of “accumulation of corroboration” - borrowing from Wikipedia’s standard of “multiple independent authoritative third-party corroborations” and applying it to entity optimisation. He explained that schema markup functions as a signpost to corroborating sources, and that Google builds confidence through consistent, corroborated information across multiple independent authoritative sources. He distinguished between sameAs (linking to parsable data sources) and url (linking to web pages about the entity).

7. Consistency and Duplication - Small Discrepancies Sink Your Boat

Jason Barnard devoted a significant section to demonstrating how tiny discrepancies create duplicate entities in the Knowledge Graph. Examples included: subdomains vs. root domains creating separate entities, accented vs. unaccented characters failing to match, word order differences (e.g., “Sal Cedrio” vs. “Cedrio Sal”) producing different Knowledge Graph scores, and time zones causing event duplicates. He stated: “Small discrepancies can sink your boat faster than a bag of rocks.” He also demonstrated that first impressions in the Knowledge Graph are persistent: “Your first shot is going to stick. Don’t get it wrong.”

8. Brand SERP as Control - Not Just Making It Look Sexy

Jason Barnard described a shift in his own understanding of Brand SERPs, from aesthetics to control: “I thought it was about making it look sexy, but it’s not. It’s about control. If I can control that Brand SERP, I control my message. And if I’m controlling my Brand SERP, by extension I’m controlling pretty much everything - I’m telling Google it means Google’s understood me.” He described a hierarchy of control: start with your Brand SERP, then C-level employee SERPs, then product SERPs, then content SERPs.

9. Google’s Annotation Layer - Why Consistency Matters

Jason Barnard referenced a conversation with Fabrice Canel (Bing) about how search engine bots create an annotation layer on content before storing it: “Bingbot functions on anything that’s standardised. It’s going to understand it better, digest it better, label it better. That annotation layer is phenomenally important. Whatever you’re doing on your site, do it consistently, because that’s what Bingbot and Googlebot are using to classify your content.” He warned against bastardising WordPress to the point where it loses the standardisation advantages that Yoast provides.

10. Wikipedia Is Not Required - And May Not Be Helpful

Jason Barnard demonstrated through his own experience (three Wikipedia pages deleted, Knowledge Panel rebuilt from scratch in six weeks) that Wikipedia is not required for Knowledge Graph presence. He stated: “I actually now don’t like [Wikipedia] very much - not because I’m vindictive, but because it’s actually more fun doing it without it and easier.” He showed examples of people who built Knowledge Graph presence entirely through their own sites with simple structured data and corroborated information.

11. Niche Authority Matters More Than Scale

Jason Barnard showed that his own site, kalicube.pro - “one of the tiniest sites in the entire universe” - was cited as a source in Knowledge Panels. He explained: “You want to look for the ones that are really, really authoritative within your niche.” He argued that in any niche situation, a small authoritative source can compete with large sites because Google’s Knowledge Graph system is “incredibly granular with where it’s looking.”

12. Google My Business Is Not a Knowledge Panel

Jason Barnard drew a clear distinction between Google My Business and a Knowledge Panel: “A Google My Business is a business listing where you provide the information to Google. It’s the equivalent of the Yellow Pages. A Knowledge Panel is a fact that Google has understood from third-party independent information, from crawling the web, that it understands independently who you are, what you do, and who you serve - without you having to tell it. And that’s much more powerful.”

Historical Significance

This webinar, recorded in September 2020, contains the earliest known recorded articulation of the UCD (Understandability, Credibility, Deliverability) framework that became the core of The Kalicube Process and a key element of Kalicube’s INPI patent filings. It also documents Jason Barnard’s entity home concept, the control hierarchy (Brand SERP → C-suite SERPs → product SERPs → content SERPs), corroboration as confidence mechanism, the “educating a child” metaphor for Google, and the annotation layer concept (credited to Fabrice Canel of Bing). Jono Alderson’s independent confirmation that “structured data is the closest, the best tool we have for communicating with Google in a way that we do control” provides third-party validation of Jason Barnard’s schema-as-control thesis.

---

Full Transcript: Structured Data for the Knowledge Panel (SEMrush Webinar September 25, 2021)

---

Paul Lovell: Countdown - there we go, and we’re live. And yeah, we’ll give it a couple of minutes for everyone to jump in. If anyone’s already in the live chat, let us know where you’re from. Ah, there we go, just opened up here. We’ve got a few - Peru, Canada, Mexico, Texas, Omaha, Seattle. Anton’s in there. Mongolia. Cornwall - again, yeah, I’m going down there, I’m looking at a house down there. So yeah, so we’re here. Thank you everyone for joining. We’ve got a lot of people here. So we are here to talk about structured data for the Knowledge Panel. We’ve got two great guests: Jason and Jono. I’ll let you guys introduce yourselves, as I’m sure you’ll do a much better job than me. And we’ll let Jason take the lead.

Jason Barnard: Well, I mean, I’m Jason - thank you very much. I’m Jason Barnard, The Brand SERP Guy®. But I will actually present myself during the presentation, so I won’t waste time on that. Jono, who are you and what do you do?

Jono Alderson: God, I wish I knew the answer to that. I’m Jono. I work as Special Ops at Yoast, who hopefully some of you have heard of in the context of WordPress and SEO. I do research and development for Yoast and a whole bunch of other stuff and SEO things. But my real, real, real passion throughout all of that is schema.org and structured data. So this is an absolute treat to be here and ready to nerd out with Jason.

Paul Lovell: That’s it. And I’m really looking forward to this one, because me and Jason had a little bit of back-and-forth on this one and been playing around with the SERPs a little bit. So yeah, let’s let Jason take the lead and we’ll get the show underway.

Jason Barnard: Right. Well, I mean, before starting - one thing in general, I remember you saying: when you joined Yoast they said, “Can you come and work for us? But what do you want to do?” And you said, “Oh, I want to do schema markup.” And they said, “Yeah - isn’t that the most brilliant thing in the world?”

Jono Alderson: Yeah. And look where we’ve come to from there. We’ve got all this cool stuff going on. Schema’s everywhere. I still saw a review recently saying something like “the amount of schema on the whole web has radically increased.” And I think some small part of that is because it’s so prolific and used now, which is awesome. Really, really cool.

Jason Barnard: Yeah, and it’s just - I’m addicted to this stuff. It’s so fascinating, really. Yeah. I mean, isn’t it kind of - 30 percent of the web is WordPress and half of that is Yoast, so that’s 15 percent of the web. So that effect is enormous. I remember you saying to me that you dreamt about it. And that incredible increase - I mean, it really does strike me that Yoast, anything that Yoast will implement, will make a phenomenal difference to the web as a whole.

Jono Alderson: That’s the dream. And it’s such an incredible position of privilege to be able to do things like this. Which I think - there’s a lot of SEO that’s tactical, it’s reactive, it’s responsive, it’s competitive. Whereas a lot of the schema stuff and the structured data stuff is so much bigger and more important than that. Like, it paves the way for new types of technologies and a more inclusive and democratic internet, and new types of competitors to Google and other services, and things we haven’t even dreamt of yet - because we’ve never had that kind of level of depth of description. And now we do. It’s an absolute dream to be part of this journey and to be pushing it forwards.

Jason Barnard: Brilliant. Right, OK. Well, I’ve actually got a screen to share. One thing I’d like to say about Jono is one thing I love about him - he can say these sentences that are like 350 words long, not pause in the middle, and they still make sense at the end. Which is a phenomenal talent.

Anyway, I’m going to talk to you about structured data for the Knowledge Panel. Ken Shaw from France has just joined us, so I’m terribly keen because I’m French. It doesn’t sound like it, but I am.

Please do follow me on Twitter or LinkedIn if you’re interested in all this stuff. I post my experiments on Twitter incessantly, to the point of boredom for everybody else. And YouTube - I’m doing loads of experiments on there and posting stuff there as well. So if you’re interested in what I talk about over the next kind of 30 minutes or so, do follow me, do join the YouTube channel, because it’s a lot of fun.

Now - structured data for the Knowledge Panel. Before starting, I’ll present myself. I’m the Brand SERP Guy. And this is what I mean: when you search my name, my personal brand name, what comes up is pretty much exactly what I want. Not 100 percent, obviously, but it’s not far off. The control - we’ll talk about later on - that Jono mentioned to me, which I love.

I used to do cartoons for kids - Boowa and Kwala. I made a movie, and I can get that up on the Brand SERP. I’ve got a podcast which is up there on my Brand SERP, saying how many wonderful - 150 amazing guests on my groovy podcast. I write lots of articles for Search Engine Journal, SEMrush - and we’ll see that later on. And I talk at webinars and conferences around the world, including Yoast.

So basically, my Brand SERP tells the story about me - who I am, what I do, and who might be interested in me. And we’ll see why that’s important later on.

And here’s a lovely disaster story that actually turns out all right. And this actually feeds in a great deal to what’s going to come up. I had a Wikipedia page - in fact, I had three Wikipedia pages about different elements of my life. They all got deleted within the space of about a week, due to - I messed with them too much, which is fair dues. And it absolutely killed my Knowledge Graph API score. And it killed my Knowledge Panel, which absolutely completely disappeared. And it’s taken me six weeks to build it back up again from zero - pretty much, obviously based on previous knowledge. And that’s a whole different discussion.

I used WordLift, which I absolutely love, on my website for building the structured data and pushing that out - more than just the structured data, pushing it out into different Knowledge Graphs. So I’ve managed to rebuild it. So that disaster is now over and we can now relax. Or I can relax, a little bit more.

Here’s the plan. Number one: Knowledge Graph. Number two: Knowledge Panels. Number three: Knowledge Panels and Brand SERPs - I’m the Brand SERP Guy, so Brand SERPs will come up a lot. Educating Google. Using schema to keep Google on track - I think this is phenomenally important and really interesting. I’m doing lots of experiments on this. And then: beware - closing the loop. Consistency and duplication, which is an underestimated problem in my opinion. And the more I work on this, the more I experiment - I’ve got about 50 experiments running right now on the Knowledge Graph and Knowledge Panels - and I see how important this is and how easy it is to duplicate, to get things wrong. And from my point of view, it isn’t even Google being stupid - it’s me being incomplete, I would say. And then some fun examples at the end. So stick around for the end - if you’re not interested in the rest of it, at least the fun examples will keep you entertained.

So - the Knowledge Graph. What it is. And I’ll do this very quickly because this is an advanced webinar. It’s Google’s understanding of the world. It’s an encyclopedia that a machine can read natively in order to produce results for search in billions of a second. So that’s a reasonably good explanation of it.

And I found this example, which is from 2013, because I used the Leonardo da Vinci example that everyone else uses and got it completely wrong on a presentation I did the other day. So I tried this one, and that’s Daft Punk in the middle, who did the music for Tron. They’re from Paris, which is terribly appropriate for a French person. And interestingly enough, this is an example from 2013. And if you type in “who did the music for Tron,” it brings up a featured snippet and not a Knowledge Panel. So small note there as we go forward.

But on the right - it’s a brain. Google understands the world like we do as human beings, through entities and relationships and attributes for those entities. And if we think about that - think about how we educate children, how we educate anybody - and if you think about educating Google like you would educate a child, things become much, much, much clearer.

So - Knowledge Panels. Where they come from. A lot of people think they just come from Wikipedia. If you look at those, they’re all Wikipedia-based. Not true. There are lots of sources that can generate and trigger Knowledge Panels. And I’ve got a collection of them - or a collection of sources that Google is willing to cite in a Knowledge Panel as being trusted. 3,184 - that was yesterday. And you can actually have a look at that list at the link at the bottom.

But in fact, it’s not just the Knowledge Graph. Google Books, Google My Business, YouTube, Google News, and others. I mean, I would say IMDb, but examples of that are relatively rare. But certainly Google-owned products can feed the Knowledge Panels without actually any presence in the Knowledge Graph. And the best example is Google Books - just try looking up books you like or authors you like, and you will find that they’re not in the Knowledge Graph, at least through the API, but they are producing Knowledge Panels. And Google My Business is a new example that’s coming up more and more. Andrea Volpini from WordLift talks about that a lot, and I think he talks a lot of sense about that. And YouTube is the new one I found, as we’ll see later on.

Brand SERPs - what is a Brand SERP? This is my favourite topic, so I’m going to spend lots of time on this. That’s a Brand SERP - it’s what appears when somebody Googles your brand name. And I’m obsessed by how important this is. A, to a brand - because people who are searching your brand name are people who are going to be doing business with you or who already do business with you. They’re people who are important to your business.

But it’s also phenomenally important for insights into Google - how Google perceives you, your brand, your content strategy, how much it understands you.

And this is a wonderful example of how much that Brand SERP is an insight into how much Google matters. With the raw blue links - if you look on the left side there, that’s StartPage. That’s the raw blue links from Google itself. But StartPage just uses the results through Google itself. And you see how much Google is messing with those blue link results. Now, on Brand Search, this is spectacular. True, on other kinds of results it’s probably less true. But here you can see - I mean, one example that really hit me when I saw this: three ads for Microsoft. Google takes them off. Google is removing those ads because it doesn’t believe they bring value. And it’s presumably giving up some money, some income, for that.

So immediately you can say: well, actually, Google’s filtering out an awful lot of ads that people - foolish people with badly organised or optimised campaigns, or badly targeted campaigns, or unrealistically targeted campaigns - they’re getting filtered out.

And then you can also see how those rich results - as I call them, the universal and the extended search results - really do take over. To the point at which there’s only one result from the StartPage result that remains on the right-hand side on Google’s final result - that’s Microsoft’s homepage.

So if we look at the top seven organic there on the left, you can see how much Google is messing - how much it’s adding, how much it’s taking away, and moving things up. For example, office.com or outlook.live.com. And it’s adding that Knowledge Panel on the right, which demonstrates a great deal of understanding.

And one thing that a Brand SERP represents or shows or demonstrates is how much understanding Google has about who you are, what you do, and who you serve.

If you want more information about how that SERP is built - it all started with Gary Illyes, who told me what they tell new engineers at Google about how the whole thing functions, with candidate sets. And it’s basically Darwinism, in a sense. It’s a beautiful, beautiful description that I wrote up in Search Engine Journal - all credit to Gary Illyes there.

Then I went round the world to figure it all out by talking to the people at Bing, who were very open and honest about how it all functions. You’ve got another five articles there. And it all confirms how this functions - which is candidate sets taking over from blue links when they have more value or bring more value to the user. And how much the whole-page algorithm plays into this. And really messes up those top seven in this case organic results.

If you want to track your Brand SERP, your Knowledge Panel, your Knowledge Graph - there’s a free way to do it there. I’m not asking for anything in return. This is my personal private project, and I’m only doing it because I want to understand. I want to understand how all this functions. How Brand Search functions. What functions. How much insight that can give us into SERPs in general. How the Knowledge Panel appears, when it does appear, when it doesn’t appear. What information appears for what type of Knowledge Panel. And also what the Knowledge Graph contains - what the API throws back and how that changes over time. We’ll see later on how much it can change.

But basically, this is my fundamental idea: to build a Brand SERP, Google has to understand who you are, what you do, and who your audience is. Then it can say: this is the brand, this is what they do, and this is information that is valuable to their audience.

And if you can get that, you can have a great Brand SERP. And also, I would argue, in your SEO strategy, you’re going to be very well served by the fact that Google understands who you are, what you do, and who your audience is - because that’s all they’re aiming to do: present the best answer to their audience, their users. And if they can understand that you have that best answer, and that your audience is that subset of their users, you’re away. And you’re probably winning.

And here’s a good measurement of whether or not your brand is well understood. You rank number one. You have rich sitelinks. You have rich elements like video, images, news, podcasts. You have People Also Ask - which implies Google will put up those questions people ask about your company but also about your industry. It has to understand you in order to be able to figure out what those questions might be. Positive and accurate choice of blue links - seems stupid, but if you’ve got a bad review ranking on your Brand SERP, think: Google thinks that’s valuable to your audience. If that’s the case, it sees it as important. And if it is actually representative of your brand, then you’ve got a problem. And if it isn’t, then it hasn’t fully understood how wonderful you really are - if it had, it would be ranking a much better set of reviews on your Brand SERP.

And then entity boxes, which I love - appearing more and more. The entities that it associates with your brand.

But of course, the big one for today is the Knowledge Panel. A Knowledge Panel is a clear indication that Google has understood who you are and what you do. And they come in all shapes and sizes. I mean, I have this thing - there on the left-hand side, Perceptive Fittings Limited. No idea who they are. But that’s the smallest Knowledge Panel I could find. It’s a company - which is pretty vague. Then you move along to the right - that’s Kalicube. I’m working on that one - work in progress. WordLift there. SEMrush - lovely Knowledge Panel. They’re Microsoft, obviously the big players.

They come in all shapes and sizes and they contain all sorts of information. And I was stunned: I have a collection of 40,000 brands, 75,000 entities in a database, and I found 3,000 distinct attributes that are cited in Knowledge Panels across these 45,000 brands. I think it’s about that number.

There is a proviso, though, and that’s quite interesting: some of them are obviously just misspellings of other ones. Which means that the machine is being let loose, from what I understand. And so a lot of them are duplicates. So it’s not really 3,000 distinct attributes that Google is willing to show about brands - maybe it’s 1,500. I would need to actually sort through the data. But it does mean there’s an awful lot of different types of information that Google understands about different brands and will put in Knowledge Panels.

And then there are all sorts of formats. I took this from a presentation - I’d been on YouTube the other day. Obviously you’ve got the brand on the left - everything. You’ve got Oakley. Products - they’re astonishingly complex, and they’re very different to the brand ones. Then you’ve got the Blues Brothers - the films, that’s very different again. The Beatles - group. Keira Knightley - obviously, if you don’t know them, you’ve been asleep under a rock for a very long time. Keira Knightley - an actress. And a book on the right there.

And you can see how different - I mean, I just stated how different it can be between different brands. If you look at that. But then if you look at different types of entities, it becomes mind-blowingly difficult to follow and mind-blowingly interesting. And you understand why Google needs machine learning to drive all this stuff.

So if we look at this: number four - educating Google. What we’re doing is educating Google. If we hark back to that Wikipedia idea - Wikipedia says you need to be notable. Google doesn’t care about notable. It cares about understanding.

So Wikipedia will throw your pages out if you’re not notable for a human being - and that’s logical. It’s an encyclopedia for human beings, which is fine. Google is not an encyclopedia for human beings. It tries to bring together information that human beings will find interesting, that they will search for off their own back. It’s trying to understand the world.

So all it’s asking us to do is to educate it. To get it to understand. Help it to understand the world - who we are, what we do, and what audience we serve. And that keeps coming back, and I’ll keep saying it until everyone gets bored.

First thing to do: give every entity a home. This is the single most important thing - before the other thing, the single most important thing I’m going to say in about 10 minutes’ time. You need to give it a home. You need to find a place on your site that you describe your entity - be it a brand or a product or a person, it doesn’t matter. It has to have a home. It’s the go-to place that Google trusts to go and find information about that entity.

On that page, you obviously need to describe it clearly and add some schema markup to make sure that it’s disambiguated and extremely explicit. And we’ve got a disaster story coming up - which you actually saw the start of - about what happens when you create a home for your entity and then you mess with the schema markup.

So you need to use schema to indicate attributes and relationships. Now, I’ve actually just picked these off - I did it early this afternoon. And it’s an example of an organisation.

Now, why have I listed those? On the left are relationships that I can potentially push to Google. If we go back to that learning idea - educating Google - we’re saying, I’ll take a little sidestep here: how do we remember things as people?

I talked to a guy in Prague who was called Chester the Memory Man - International Memory Man. And he was talking about how we remember things. And he was talking about linking things onto other things we already know. And all those memory techniques are based on entities we already know. And the more you can link things that you want us to understand to something we have already understood, with a relationship, the easier it is for us to remember it. It’s exactly the same with Google.

So if you look at: alumni, brand, employee, subject of, founder, funder, member of, owns, parent organization, sponsor, organization affiliation - if you can find an entity that’s already in the Knowledge Graph and you can link to that, you’re already winning the battle.

And on the right-hand side - and this is good news for anyone who runs a podcast like me - sponsor, sponsor a podcast, you link yourself to an entity that’s already recognised and you build up your own presence in the Knowledge Graph and Google’s understanding of you. Contributor, publication, production company, copyright holder, organiser, participant, publisher, funder - all of that pushes towards creating great content, but also using schema markup to explicitly tell Google that you created that content, your brand created that content. So it’s taking credit for something you’ve done and linking yourself to it. Pushing all of that into the Knowledge Graph.

So if you’re the copyright holder of an amazing film like the Blues Brothers, it’s going to be really easy to get into the Knowledge Graph simply by linking yourself to that.

One important aspect here, of course, is you want to do it accurately. And we’ll see why that’s so important in a moment. But you also want to get it reciprocated as much as possible. Obviously a one-way link - you’re saying, “I did this.” Who’s going to believe you? You can lie. But if the link comes back saying, “Yes, he did do that,” or “That brand did do that,” you’re away. Google’s going to go, “Right, OK, that’s truth, I’m away.” And in the Knowledge Graph you go.

The best relationships: find those that are close, strong, and durable. The length of time is important. The strength of that relationship. The closeness of the relationship. If you think about families, for example - sisters and mothers. Very close, very strong, very durable. In fact, unbreakable. But if you think about employee - not so durable. So you want to find the relationships that are close, strong, and durable. Obviously, not every relationship you can find from this list is going to actually fill that need. But then, the more the merrier. The more you can link onto things, the easier it is to remember as human beings - but the easier it is for Google to actually get a grip on it.

Here we’ve got the Knowledge Graph Explorer, where you can find entities - that’s on my site. And here we’ve got the example of SEMrush and WordLift. And you can see how their presence in the Knowledge Graph has built up over time. On the left-hand side, if you look at the Budapest Knowledge Graph Update - that’s the reason for the big leap for SEMrush there in September 2019. If you’re interested in that, look it up on Search Engine Journal. There was a big change in the Knowledge Graph in August 2019 where all the scores went up phenomenally. It was a really major update. And I haven’t seen one since, but I’m keeping track.

So once you’ve established a relationship, you use schema to signpost corroboration. And what I like about it is the accumulation of corroboration. I thought that was a brilliant phrase when I said it the other day. It doesn’t seem quite so clever right this moment, but it doesn’t matter.

This comes back to Wikipedia. What Wikipedia has is multiple independent authoritative third-party corroborations. Now, I said you don’t need Wikipedia, but you can use that as a really good guide. If you’re going to point to something using your schema markup as a signpost, you want to point to as many multiple independent - not you - authoritative - big sites, big brands (in fact, I should say not big sites) - third parties. Once again, not you.

And why I put “independent” in brackets - it’s because you can actually corroborate with things like LinkedIn, Twitter, on your own account. Google isn’t completely exclusive about the fact that it needs to be corroborated by someone else. But that does mean you need to be incredibly consistent across all the sources that you control. Which we’ll also come to later.

And at the bottom there, I’ve got a little explanation which has been slightly cut off by my bottom right-hand corner - sameAs. I like the idea of saying: if you’ve got any doubt about whether to use sameAs or url - sameAs is supposed to be a link to data with other data that it can then parse. So it would be something like DBpedia data. And WordLift have also got a system for that. And url is meant to signpost a page that talks only or mostly about the entity. It can contain schema markup. So don’t say, “I’ll link to that page as sameAs because it contains schema markup.” Because the page is actually a web page. You want to link to databases or sources that actually contain parsable data. Otherwise, use url. SameAs was added, as far as I understand, after url, in order to provide that possibility. And people have been cheating on it, as usual.

And there you go - the sources quoted in the Knowledge Panel in my super-duper set of 45,000 brands, 75,000 entities: 3,184 unique domains. And that’s only a tiny part, obviously. I’m only tracking 45,000 brands. There must be 25 million. So that number is very small, but it’s very niche.

So what you need to do is actually not so much look for things like Wikipedia, or Crunchbase, or Inc.com, or The Guardian, or whatever it is you think is authoritative - that’s brilliant, that’s such a big site. You want to look for the ones that are really, really authoritative within your niche. And if you look at that list, you’ll see some really tiny sites. In there somewhere, I think, is kalicube.pro, which is my site - one of the tiniest sites in the entire universe. And yet I get cited in Knowledge Panels, because I’m authoritative for certain things.

Next - keeping Google on track. It just shows how important that home is. Once you’ve got that home, you can really keep things on track.

And - oh, sorry - discrepancies. Discrepancies that seem unimportant can make all the difference in the world. Here we’ve got three different Knowledge Graph IDs - and M and two Gs - and they show exactly the same thing. And it took me quite a long time to actually manage to get that to happen. That’s my ex-wife. She ended up with three Knowledge Panels. Part of it was my fault, part of it was already there. But I’ve managed to disambiguate to the point at which it now shows the same thing for all three. Next step is to get rid of the other two that we don’t want anymore.

These guys - characters - Boowa and Kwala. If you look for “Boowa and Kwala”: it’s a person - it’s weird. Two people: Boowa is one person, the blue dog in a cartoon. Kwala, another person in the cartoon, yellow koala. Obviously. TV series. It’s a movie. It’s a music group. It’s a music album - there’s a music album called “Boowa and Kwala.” There’s a music group called “Boowa and Kwala.” There’s two songs called “Boowa and Kwala.” And there’s web content called “Boowa and Kwala.”

In my own defence, it was before the Knowledge Graph. And I just kept calling things “Boowa and Kwala.” And I didn’t realise how ambiguous that would become - and how ambiguous, in fact, it already is in our own minds. When I say to you “Boowa and Kwala,” which one do I mean? The two characters? The TV series? The movie? The music group? The album? The actual recording of the song? Or the web content? So obviously, that ambiguity is incredibly difficult to deal with. And you have to deal with it. And it’s very difficult for us as humans to understand that it even exists, because in our minds we associate one particular string of characters to one specific entity, more often than not.

Clarifying - once again, tiny differences can make all the difference in the world. Here it’s my ex-wife again. I do lots of tests on anybody who gets anywhere close to me. So if you don’t want to be tested on the Knowledge Graph, don’t ever communicate with me.

Véronique Barnard - with the accent, in French. Without the accent - it just doesn’t recognise her. Which is stunningly wild, because it’s got a very good confidence level with the accent.

And then this is my daughter. And what I did was use schema markup using alternateName: Léonard, Joe, Leonard without the accent, and Leonardo Barnard. And as you can see, the confidence Google has that that actually matches the person I’m talking about drops very dramatically as soon as you remove part of the name. So that difference in name for my daughter is actually quite problematic in terms of actually getting Google to understand when I’m talking about her.

Next is stabilising. This is the disaster. And it was, at the time, quite frustrating and disappointing. And my ego took a pretty big hit. But as you can see - three Wikipedia pages: the cartoon I created with my ex-wife, the band I was in in the ’90s, and myself. And the Wikipedia pages for all three got deleted at the same time, because I’d been messing with them too much. Fair dues, no problem.

What happened was: the Barking Dogs went from strength to strength. Boowa and Kwala - absolutely no problem at all. And mine completely plummeted. I lost my Knowledge Panel. Everything went out the window.

Why? Because I moved the schema markup from the page it had been on to a new page, in an attempt to clarify everything. And it was a completely stupid mistake. Put it all back, rebuilt it all, and it’s taken me six weeks to build that all back up again.

But what’s very important here for me is: once Google’s got a grip on you - on your entity - through Wikipedia, Wikidata, or any other source (Crunchbase, LinkedIn, or your local dog society if the niche market is dogs in your hometown), it won’t drop you out of the Knowledge Graph as long as it has that schema markup as a reassuring crutch.

Next is creating. And this is what Paul was talking about earlier on. This is a podcast I created using Eventbrite. Took an hour - less than an hour, a few minutes, basically. Eventbrite had this educational event markup, and it creates, all on its own, that Knowledge Panel, without any support from any other source.

If you look at YouTube - same thing. And actually, it doesn’t quote a source, because it’s a Google-owned source. And YouTube - if you see at the bottom there, a publication is “live broadcast start date.” You wouldn’t have thought that would be an event, but it is. And YouTube is a really great way to push stuff into the Knowledge Graph, if it’s a live event, for example.

And you’ll see that the Knowledge Graph ID is different. And as you can see, there is a problem with the fact that I’ve created multiple entities with the same thing.

But in short, that whole thing is: structured data gives you the closest you will ever get to control of your entity in Google’s mind.

And Jono showed me this - it’s his new talk. And he showed it to me and I was convinced I’d already seen it. Tell me.

Jono Alderson: I’m delighted that we came to the same conclusion in parallel. It’s lovely. Thank you very much for pulling this out.

No, I was listening to - where did I catch you talking? Oh, we were talking about this. We were preparing for this presentation, and you mentioned the word “control.” I was like, “Yes, this is the point.” This is - because so often the question comes up about why should I be doing all this schema stuff beyond just getting the immediate rich snippets? What is the value beyond just the tactical “get the rich results and the Brand SERPs”?

And I’ve struggled with that answer for a long time. Because I believe it’s profoundly important for a whole bunch of reasons, but I’d struggle to convince a CEO of that.

And the answer I’ve come to is that it is the closest - the best tool we have for communicating with Google in a way that we do control. Where we’re not relying on them scraping our pages and guessing what we’re trying to say. Now they’ve given us the tools to say explicitly what we’re trying to say. We can describe our value propositions, our relationships, our products, our prices.

So I think, yeah - as a tool for control, there is nothing else like it. And as the world changes and Google gets better and better at relying more on this, and better and better at understanding things, this is the way we communicate with Google. And it allows us to start to think about Google as somebody we have to convince of our value, in the same way that we’ve always had to convince consumers. This is the language that we do that in.

Jason Barnard: I have a three-step process - or the three things we need to do. We need to create understanding. We need to convince Google about credibility. And we need to have content that it can deliver, or that it believes we can deliver. And that’s it. Everything you do in SEO has to serve one of those three pillars: Understanding, Credibility, Deliverability - which is a cut word, but I’m going to keep with it because I can’t think of a better one.

And you were saying - Brand SERPs and the word “control.” This is phenomenally interesting. Because that whole thing about Brand SERPs - I thought it was about making it look sexy. But it’s not. It’s about control. If I can control that Brand SERP, I control my message. And concepts are phenomenally important to anybody searching our brand names. But beyond that, if I’m controlling my Brand SERP, by extension I’m controlling pretty much everything I’m telling Google - it means Google’s understood me.

And if I can push that out - I say my Brand SERP, then I say the C-level employee SERPs. If I can control them, then the product SERPs, then my content SERPs. If you search my content, what turns up? That’s all about control.

And we do get that opportunity. It’s not 100 percent control. But it’s the closest we’ll get. And schema markup is wonderful for that.

But I remain convinced that I have seen that slide before. And I stole all your ideas.

Duplicates - and this is the one that I got really - I really got off on it, because it’s quite a good laugh. Closing the loop seems so simple to a human being. You think you have done that. But a machine doesn’t give you any quarter - is that what we say in the pirate world? It won’t let you off with not closing the loop. You have to close every loop and make sure that everything is the same. Or you end up - excuse me - with duplicates.

It seems so simple, and yet any tiny discrepancies can make all the difference in the world.

Here we go. On the left, we’ve got “Boowa and Kwala” - that was the example. In fact, all of those are the music group. But it keeps spreading - music recording. But they are the same thing. But we’ve got four entities in the Knowledge Graph that represent that one thing. And if I may defend myself, that wasn’t my fault.

On the right, however, it’s completely my fault. Two entities for the “April Dunford, Kalicube” event. The Eventbrite number, the title is the same, the time is the same. It’s created one from Eventbrite and the other one from YouTube. I have yet to figure out how to stop it doing that, because I made sure for this one that everything was the same. So there must be a difference. There must be something in there that Google has latched on to and said, “That’s not the same thing.” Even though it is.

And this is the example that really got me started on that, a couple of weeks ago. Paul’s markup. He started off without a time, and then Google thought it was at midnight GMT. Then he added a time, but he forgot to add the time zone. And that really brought to me two things.

One of which is: your first shot is going to stick. It’s been three, four weeks and we haven’t managed to get that time of midnight to change - even though the structured data has changed on the site, and even though it was that structured data that had created the Knowledge Panel and the presence in the Knowledge Graph. So your first shot sticks. Don’t get it wrong.

And the other is that tiny difference. I mean, for Paul - I would imagine he’s in the UK - it’s 1800 hours. It doesn’t necessarily occur to us that there are other time zones. Then we think about other time zones, and the first thing we do is say OK, GMT. GMT is spectacularly difficult, because GMT and BST are two different things, and UTC is actually - as far as they understand it - GMT. Try digging into that. Then you get your times right, and then you can stop duplicating these events with different times.

But here are some examples of small discrepancies that I’ve seen, and I keep seeing them. You’ve got namesakes, obviously - different people with the same name, rather than the other way around. But I’ve also seen a sub-domain where Google has seen one domain with a www and another one without it, and taken it to be two different entities. So obviously that’s really important. That was a problem of incorrect redirects that Google couldn’t understand because it couldn’t follow the link. Time zones - we just talked about. Accents - we talked about earlier on, with my daughter and my ex-wife. The order of the words - I have a client called Sal Cedrio, and “Cedrio Sal” brings up very different result scores in the Knowledge Graph. So we need to disambiguate. We need to make sure that Google understands, and we need to use the same terminology - “Sal Cedrio” rather than “Cedrio Sal” - every single time we mention it. Especially in the structured data.

And my new opinion about this, that I made up this afternoon, is: it could be a specialisation all on its own. Because you have to dig down every single time - there’s a different reason. I had an author, and it was a search result that was making it wrong. If you search for her name on her publisher’s website, she had another author with the same name who came from New Zealand. She was from Australia, and it was mixing them up. And it was one search result on one site, because that one site was authoritative. And that’s the primary thing you should be looking for - what is authoritative? What is driving that result? And in this case, what is driving that mistake?

Fun examples to end with. I promised you - because this has gone on a little bit. This webinar was a fun example. I love this. Forget featured snippets. Knowledge Panels - you can jump everybody and get right to the top. You don’t need featured snippets anymore.

And I tried to reproduce this, as you will see, on a sunny Sunday afternoon. I did “What is the Knowledge Panel? Google: Jason Barnard Explains.” Got myself a Knowledge Panel. It took me two minutes from the moment I created it. But basically, I invented the event and created it on YouTube. And within two minutes of creating the event in my own brain, Google had already understood what it was. Which is phenomenally cool. Quite a good laugh.

But I didn’t get that Knowledge Panel - which was a big disappointment to me, because I really thought I would. But I got lots of visibility. You can see: the video immediately ranks. “What is a knowledge panel in Google” - my website ranks. The site I put the event on - and this “what is a knowledge panel” - that’s a lot of good visibility, very simply, very quickly. And that took a couple of days to rank the video and to rank the web page.

And there’s a couple of people doing this really rather well. This guy - Italian artist. And I love it that they say “Did you mean genius?” He loves that too. He’s been working on it bit by bit. He commented in May, so I started tracking his presence in the Knowledge Graph in May. Look how it goes up. And all he’s done is placed information all over the web that confirms what he’s saying about himself. And that whole thing is based on dangritz.com, which is his own site, which has the simplest structured data you can possibly imagine - but points to all the right places.

And this guy - Read Bob Brewster - who basically created himself and his entire family in the Knowledge Graph. And he even has his shoe size, which I love. And the hospital where he was born. And that’s just placing information, being very methodical about it. And as you can see, he actually had all his Wikidata stuff deleted - and that was in July - and he’s gone from strength to strength. And I really hope that my Wikipedia experience didn’t make that happen to him, which would be really unfortunate. But he’s done a great job of keeping that going, pushing it forward.

And this is my favourite example, because it’s my super-duper fun characters - Boowa and Kwala. I’ve managed to create the entire family structure in the Knowledge Graph. And it’s taken me two or three months. And - oops, sorry, that was the wrong way - as you can see, the Knowledge Graph API score is going up and up. And when you search for that, it associates all these family members with it.

And on the site, I’m using WordLift to actually create the structured data that’s allowed me to do that. And if you can see there, my site jasonbarnard.com is actually cited in the Knowledge Graph for four of the characters. So that structured data is not only getting me cited in Knowledge Graphs, getting me a presence in Google - it’s also allowing me to inform Google of this particular circumstance, or this particular set of characters, and control it. Because those descriptions are mine. The relationships between them - who is Mummy Koala’s significant other? It’s Daddy Koala. But I had to explain it through structured data.

So here’s - to end up - things you might want to retain in your brain. Knowledge Panels vary enormously by entity type and entity, so you need to look at them case by case. Wikipedia isn’t necessarily helpful - I actually now don’t like it very much. Not because I’m vindictive, but because it’s actually more fun doing it without it, and easier. Give your entity a home. Force that home on Google - you’ve got to force Google to accept that your home is the home it should be looking at. And then you can, as Jono said earlier on, control. Schema.org markup is your best shot at control. Small discrepancies can sink your boat faster than a bag of rocks - as you saw with mine, which sunk like a bag of rocks from one day to the next. Brand SERPs are the most interesting thing in digital marketing - I’ll put that in white so that it really stands out. And track all this stuff, because it changes so very fast.

More information: visit my site, follow me on Twitter or LinkedIn, go to the YouTube channel - there’s absolutely loads about this on there now. And if you want more information about Brand SERPs, Knowledge Panels, Knowledge Graph, and those examples - just go to kalicube.pro. I’m sharing all this information. It really is just kind of an attempt to understand. I really just want to understand. So please do go along, put your brand name in there so that we can track it. And we, as a community, can understand better.

Thank you very much. That was it. Done it.

Paul Lovell: Hello, Jason. That was great. No, I loved it. It was - I mean, the information there is solid. We’ve had lots of good questions. Before we start talking, if anyone else - if anyone watching - has got any questions, please do ask them. We’ll try to get to as many as possible.

I think one of my questions, Jason, to you: you’re tracking so many different entities by your site. Would you say that if you are able to gain any of those types of panels, graphs, whatever you want to call it - would you say that Google would then understand you as sort of like a trusted source or a trusted site? As in, they’re triggering that based off of you just placing structured data, for example? Would, for example, a spammy site be able to place structured data and get one of those enhanced features, I suppose we’d call them, in the Google SERPs?

Jason Barnard: Well, it’s almost certainly open to spam. And I wouldn’t ever say - what I managed to do was get Google to trust me about myself, my rock group the Barking Dogs, and Boowa and Kwala - the cartoon characters that I created with my ex-wife. And once I did that: A, I was cited in Knowledge Panels. But B, I could actually push any information I want. But more importantly, perhaps, is that I could claim the Knowledge Panels - because it recognised my site as the authoritative site. And in fact, then you just need to click on “Claim the Knowledge Panel.” If your Google Search Console is linked to that site, then you can claim control of the Knowledge Panel. So I can actually control all of these things through that.

So building up your own authority for any specific entity is the single most important thing you need to do. That’s it. And yeah.

Paul Lovell: Jono, I don’t know if you’ve got anything to add to obviously what Jason’s just been sharing, before we get to any questions.

Jono Alderson: Nothing particular. That was awesome. It was a really good overview, I think. The only thing I’d like to echo - which we’ve touched on in a previous one of these - is just how many, and you touch on this, Jason, just how many different ways there are to validate and to showcase your authority for things like: which certifications does this brand have? When was it founded? How tall is the founder? What is his shoe size? Where was he born? He mentioned things like that sort of thing. It exists. You can use schema.org to describe it. Nobody does it. It’s hugely powerful.

And the other thing you mentioned, which really got me thinking, was things like: OK, you start off by taking control of the Brand SERP, and go great, OK, we’re now managing this landscape and our messaging. And then you step back and you say, OK, now we’re going to start to take control of the C-suite, and their landscapes and reputations. And actually what you’re doing, through engineering your Brand SERPs and managing your schema, is improving your business. And you’re earning all those results. Because you can’t really do one without the other. You can’t say, “Look how authoritative our authors are, and look at all the incredible university degrees they have,” without having incredible authors with university degrees. So you start to have to actually improve your thing before you can describe it in a way that proves that it’s authoritative - which I think is lovely.

Jason Barnard: Yeah. Look, it’s astonishing how much we think all of this is obvious. The number of people I’ve had talk to me and say, “Well, it’s obvious that I’ve got a degree.” And - you know - isn’t obvious at all. You have to actually explain it. And also, as you said, it’s linking all this stuff together that you kind of think, “Well, yeah, all right.” And so you link it, and then you big each element up. And each element that’s been bigged up supports the other elements and bigs them up. And it’s this kind of self-fulfilling prophecy - and sorry, a mutually supporting set of entities. And that mutually supporting set of entities is incredibly important. Sorry, go ahead.

Jono Alderson: No, just agreeing with that.

Jason Barnard: One thing I’ve noticed is: pushing things into the Knowledge Graph that are related to me has become increasingly easy over the last four or five months, since I’ve really made an effort. And basically, if I can link it to me, it will probably get in the Knowledge Graph. And some of that, in fact, is the way that you phrase your sentences. It’s semantic triples that Dawn Anderson talks about all the time. Bill Slawski loves them as well. But you know - “Jason Barnard’s father is John Barnard” - that will get John Barnard into the Knowledge Graph. And I did that, and it took me a couple of days.

Jono Alderson: Nice. I think it’ll be really interesting to see what happens as this space becomes more competitive, right? So you have some authority - don’t give me that authority score, for once of a better phrase - like, you’re a reliable, reputable node in the graph now. And things which are closely connected to you are probably more reliable and trustworthy. As other people start building these pools, and there’s the intersects between them - are my reviews more trustworthy than your reviews? And what happens if they’re different? That’s the game, isn’t it? That starts to get really interesting.

Jason Barnard: But then it’s also a question of whether - as I was saying earlier on - it’s very niche. I mean, you see what information, or Google’s Knowledge Graph system, or the machine - is incredibly granular with where it’s looking. And so that competition is actually very limited. It’s the dog parlour from next door against the dog association from the town next door.

Jono Alderson: Yeah. Sorry, it’s not you versus Amazon or giants. It’s very distributed.

Jason Barnard: I mean, obviously Amazon will have a great deal of weight in terms of products and in terms of books and authors. But in any kind of niche situation, you can still play the game, as it were, simply by being incredibly authoritative within a very niche market. And that list I gave earlier on wasn’t broken down into niche markets. But I think I might do that, because that would be very, very, very interesting.

Jono Alderson: Oh, that would be really cool - categorised by industry.

Jason Barnard: Yeah. Well, that’s my Sunday afternoon sorted.

Paul Lovell: Right, so let’s get to some of these questions. So, Diana Richardson’s asked the question: for a personal brand site, which schema properties would you recommend to boost one’s presence for brand searches and in the Knowledge Panel? So yeah - what are those top schema types that you guys would recommend for your brand searches, for a personal brand? So not my usual, like, company corporation - a personal brand itself.

Jason Barnard: Well, I mean, obviously the first thing is to write a decent description about yourself. Use semantic triples. Keep it simple. Don’t do what I do, which is trying to describe everything about yourself because you’re so proud that you were in a band in the ’90s that nobody cares about, or you made a cartoon. That was a big mistake on my part. If I was focusing on digital marketing 100 percent, then my Brand SERP would be even better than it is today. But it took me much more effort to get there because I was being multifaceted, and I wanted to prove to everybody in the whole world how wonderfully multifaceted I am. Complete mistake. Google isn’t that smart yet.

So keep it simple. Keep it factual. I link myself to John Moores University, which is a - well, let’s say - modest university in Liverpool. Luckily, John Lennon went there. That’s why I get linked to John Lennon, which is my favourite thing in the world.

But that kind of thing - awards and certifications. Jono mentioned them as well.

Jono Alderson: Yeah, that’s great. It makes me look clever. I guess the easiest answer is: exactly what you do in the real world. You would say, “Here’s the university I went to, here’s my rich, famous parents, here’s…” Like, there are properties for many of these things. Rather than trying to cherry-pick individual things and say, “How do I mark them up?” - actually, no. You want to be the answer. “How do I describe how authoritative I am?” is to describe how authoritative you are.

Jason Barnard: Brilliant. And also - and look at, as I said earlier - look at the close, strong, and durable relationships you have that are already in the Knowledge Graph. That’s a good place to start. And try to keep them on one line of thought. I.e., once again - coming back to - linking to my mother as a jazz singer, linking to my sister as a film director, it’s easy peasy lemon squeezy to get into the Knowledge Graph. But then it’s very difficult to get Google to understand that I’m not a musician - I’m actually a digital marketer.

Paul Lovell: Yeah. And you can find all of those schema property types, if it’s for a personal brand, under “Person” at schema.org. And you can just go on and on.

Jason Barnard: I know, I know, I do. I do try. Honestly, the smartest people come up with the simplest answers.

Paul Lovell: Yeah. So just go to schema.org and look up the Person structure. It will guide you. And try and be as complete as you can. And like Jason said, don’t try to give too much - just focus on what it is that you believe is important for the niche that you’re targeting and for your audience. Because I think one of the things I like, Jason, was when you were going on about audience and trying to understand and getting that Brand SERP right. What you need to understand is: people that are looking for your brand are probably looking for your service or your product. So you want to try and make sure that you’re sending those signals to as many different places as possible. Because the chances are, they’re going to be searching for something else that’s related to you, that you may well be able to trigger one of those panels quite easily for, if you mark everything up correctly. So I don’t know - where are you getting that bigger bit of visibility that you wanted?

So let’s have a quick look here. What else we got? So, Mika - I won’t pronounce that surname because I’ll probably get it completely wrong - hello from Malaysia. “I work for a company that offers a platform-based service. Which would you recommend that I prioritise: schema for a corporation, or for a platform-based service?” So that’s getting a bit more specific.

Jason Barnard: Well, Jono, you’ll correct me if I’m wrong - but that’s two different things. One of them is the product and the other is the organisation. Sorry, go ahead. I’ll let you explain it because you’ll do it better than I do.

Jono Alderson: Right. Like, understanding and defining the relationship between those is really important. It’s an organisation that offers a service. And you can describe that relationship. One of the things that people often don’t do very well with schema is they say, “I’m going to pick one thing for this page - this is a service.” No, actually - you want to describe the relationship between those things. This is critical.

And you can do that in schema. In fact, the Organisation schema has properties like “services offered” - and then you can describe those things inline. And you can map that specifically. So yeah, you can be both of those. Or you might be both of them. You might be an organisation that you subscribe to. I don’t think that’s a great example. But things can be multiple types. The one that Google always explains is: a book is both a book and a product. It can have an author and a price. So when things truly are in that kind of grey area, you can double up and describe them with properties of both things - which is really nice.

Jason Barnard: Oh, if I can make a point about books. I had a client the other day, and the book is a concept. It’s the publication. The edition of the book - that’s the thing.

Jono Alderson: Interesting.

Jason Barnard: Yeah. So the author is linked to the concept, but each individual release of the book - and it gets really complicated really quickly - each of them will have an ISBN number. So you can actually identify the different years and different things. But the book in itself is actually just a concept that is an entity.

And if we come back a step to the platform we were talking about, I would assume they’ve got the same name - and that’s a big problem. If you look at Boowa and Kwala: it was the TV series, it was the web series, it was a song, it was an album. You’ve got the same problem. You’ve got the same name that is both the corporation and the software. So you have to disambiguate. And unfortunately, it’s quite difficult to do. I talk from experience.

And I think the thing I like that you just mentioned, Jono, was the subscription - because that’s coming soon, isn’t it? Recurring payments and variable stuff.

Jono Alderson: That’s really interesting. Like, I imagine a significant proportion of businesses around the world are subscriptions. It’s not like - there are so many businesses where it’s £9.99 a month, or there’s a free trial, or it’s insurance, it’s a mortgage. And there’s not really been a good way in schema.org to describe those yet.

But yeah, that’s coming soon. So you can say, “This is £9.99 a month, and there’s a trial period, and in month three it goes up to whatever, and you do 24 payments.” Things like mobile phone purchases, which often happen over a 12-month period - now we can describe all of those. So yeah, if you’re in a subscription space, go check that out on schema.org. That’s going to be really cool.

Paul Lovell: Yeah. And it’s going to be interesting to see how, when that’s nested within a product, how that shows in the SERPs rather than just showing a price of £9.99. So that’d be good.

Right, next question. Is it best practice to use one schema format across all web pages?

Jason Barnard: You’re talking - I was talking about Fabrice. He was actually - not the French singer - Fabrice Canel, who is Mr Bingbot. Who isn’t the French singer.

And he was saying basically: Bingbot functions on anything that’s standardised. It’s going to understand it better, it’s going to be able to digest it better, it’s going to be able to label it better. And that labelling - as he was talking about, an annotation layer that Googlebot actually puts on all your content before it sticks it in the database - is what all the algorithms use to pull that data out. That annotation layer is phenomenally important.

So whatever you’re doing on your site, do it consistently. Because that’s what Bingbot and Googlebot are using to actually understand how it is they should be classifying. And that’s a really interesting concept.

I mean, if you look at two pages and they’ve got two different systems of schema, it’s going to go, “Well, I can’t classify them the same way.” And it’s going to be thrown off. It’s going to be looking for a system. And WordPress - he uses that as an example. WordPress, however good or bad you think it is, it’s systematic. Yoast is systematic. And so anything like that is already going to give you a big advantage, simply because it’s already seen it so many times.

And the further - this is my pet hate - people who take WordPress and then bastardise it to the point at which it isn’t recognisable. The further you get away from it, the less you can actually draw the advantages that it brings in that standardisation. That standardisation that - and I’ll repeat it - Bingbot and Googlebot use to understand, to digest, to annotate. Which allows the algorithms then to access this data. Sorry, I got a bit carried away.

Paul Lovell: Yeah, that’s good. So Mike’s asked the question - and Mike, if you’re there, just give us a little bit more clarification - he said: how do we recognise errors, like 404s, with schema?

Jason Barnard: Oh, that’s Jono.

Jono Alderson: There isn’t really a good way. This is really interesting. It’s something we struggled with in Yoast SEO. There isn’t really a good set of concepts for errors. There aren’t any concepts for types of error pages. You can’t describe a thing as a 404.

But I don’t know if you need to. Because all of the contexts where external systems like Google are going to be trying to consume schema - “this is a product page,” “this is a blog page” - don’t really apply if it’s an error page. And actually, just not having the schema there might be enough.

So in Yoast, I think - off the top of my head - we still output the Organisation schema and we output the Website, so we say, “This is yoast.com by Yoast.” But we don’t output anything about the page, and we don’t apply anything about it. Because there isn’t anything to describe.

Jason Barnard: Well, it’s kind of cruel, because it’s like, “This is an error - buy this website, owned by this company.”

Jono Alderson: And there are some discussions historically in schema.org about more specific types of pages. But I don’t think there’s much conversation about errors, because the assumption is: if it’s an error, then why does it need schema? Why should it have it?

Jason Barnard: Can I ask a question, please, Jono? All the different page types - like About Us and Collection and Contact Us pages and whatever - how important are they?

Jono Alderson: I think the honest answer is: at the moment, not particularly. Because there has been very little use of them historically. That’s something I want to see change. We recently shipped a feature in Yoast which allows you to change them on a per-page level. So hopefully people will start to push those, Google will react and start to treat them a bit more differently.

But I think, in the grand scheme of things, they’re not hugely different. I lost an argument a while ago about trying to introduce a concept of “home page,” which doesn’t exist in schema - which drives me mad. There was a huge amount of pushback. So I suspect trying to also say, “Could we have ‘error page’?” might be a dead end as well. But maybe Google will change all that, if they say, “We’re doing stuff with this now, this is important.” And then that opens the floodgates.

Jason Barnard: Well, identifying what kind of page we’re talking about, right at the top level, seems to me to be a no-brainer. And what I hear from you, Jono, is: you’re about to change the world again.

Jono Alderson: You, me, and the team - and then all the many people involved. But yeah, that’s a hill I’m willing to die on. I want to be able to say “this is a home page.”

Jason Barnard: Right. OK. Yeah, no, I mean, I love the idea of being able to say what kind of page it is. And I think that could be incredibly useful, granularly, moving forward. And I’m jolly well hoping Google are listening.

Paul Lovell: That’s it. So let’s get to our last question, from Gonzalo. “Should I have a Knowledge Panel for my company? I already have one by GMB, and another one as a marketing consultant. So should I have one as well for my company?”

Jason Barnard: Right. OK. This is one of my pet moans. A Google My Business is not a Knowledge Panel. Google My Business is a business listing where you provide the information to Google. It’s the equivalent of the Yellow Pages in the UK - “les Pages Jaunes” in French. It doesn’t actually mean anything other than the fact you’ve gone to the effort of filling in a form.

A Knowledge Panel is a fact that Google has understood, from third-party independent information, from crawling the web - that it understands independently who you are, what you do, and who you serve. Without you having to tell it. And that’s much more powerful.

But that isn’t to say that Google My Business isn’t important. Google My Business is a great platform for leaping into the Knowledge Graph. Because Google will use that as a source of information about you. You just need to corroborate it. And use schema markup using sameAs, for example, to identify which Google My Business corresponds to your company, and then leverage that into the Knowledge Graph.

But it’s that independent understanding that Google has - of who you are, what you do, and who you serve - that’s going to drive your SEO into the 21st-and-a-half century. I suddenly realised we’re already in the 21st century. I’m sorry. But yeah.

Paul Lovell: So - yeah. If we want, we can just grab this last one here. There haven’t been too many questions, so let’s just grab it. This is from Nor: “What would be good schema practice for a business operating locally, which I think you’ve just touched on, as well as providing services nationally - online service-based businesses that operate locally but provide services nationally?”

Jason Barnard: Go on, Jono.

Jono Alderson: I guess a combination of GMB and some schema. So GMB handles the local stuff. Schema is interesting for this, because a LocalBusiness - which is the schema you want to be looking at - doesn’t explicitly mean just a bricks-and-mortar store that serves the high street. It just means a business that has an address that a consumer can walk up to and wander around your shop.

So you can say: this is a local business with this address and these opening hours, but also that the area served is this big geographic radius. You can use sameAs, as you mentioned, which is really nice - to say “a reference to a map of that area.” That map could contain a geographic radius. Maybe even use your GMB map. And so yeah - some kind of relationship between the two is nice. But don’t be scared about saying “this is a local business with an address” if you have a broader service area. That’s absolutely fine.

Paul Lovell: All right. That’s -

Jason Barnard: Sorry, go on.

Paul Lovell: No, go ahead.

Jason Barnard: No, I think that’s a great answer. Because it is something that we worry about. And I’ve got these Brand SERP courses, and one of the questions that kept coming up - that people were saying to me, “Can you please add this to your course?” - is: what do I do if I’m a national or an international business, but it keeps bringing up the Google My Business panel, making me look small?

And the answer is: you don’t have to be afraid of it. Because it’s a question of the intent of the user. And although Google isn’t always getting it right, it’s tending towards that.

And the idea of actually having a physical location is actually very important - that Google can identify where the brand is based. And interestingly enough, for the web - because you’ve got all these websites that can be anywhere - the physical location of your business is perhaps one of the more important aspects of your business that you need to communicate to Google.

Jono Alderson: Yeah. Legitimacy, I guess.

Paul Lovell: And how have you guys seen - within structured data, there’s a part which is, I think it’s “areaServed” - where you can call out a GeoCircle or you can call out GeoCoordinates to map out a specific area. Have you seen much of that used at all?

Jason Barnard: Well, I did it using Google My Business, in fact. Just to see how it works out. And I’ve tried the tiniest, like street-level, to town, to country. And then I did the world. And it starts to look a bit ridiculous with the world. Which is a bit of a pity.

I think the idea that Google should start looking at - yeah, I said “the world.” And as you were saying, subscriptions: a lot of us are doing subscriptions, a lot of us serve the entire world thanks to the internet. And it’s ironic that Google hasn’t really got to grips with it.

I actually saw - sorry, events, one of the things with events which is my favourite thing at the minute - was it needed a physical location. The virtual location has now come into play. But I actually saw a Google event where they’d cited it as being “taking place on the Earth.” And it actually had a link in the Knowledge Panel where you clicked on it and it showed you the Earth.

Paul Lovell / Jono Alderson: Nice. Nice.

Jason Barnard: So I think apparently Googlers have got a sense of humour, even if they’re not necessarily up to date with the schema markup.

Paul Lovell: That’s super. Well, that’s a wrap. Thank you to everyone that’s watching and people that have been commenting and asking questions. If you’ve got any more questions, you can reach out to all three of us via social media, as well as SEMrush as well. So if you want anything else answered, then feel free to pick up with us there afterwards.

Thank you guys for coming along. And thank you, Jason, for a great presentation. And thank you so much for letting us on. And we’ll see you again soon.

Jason Barnard: Yeah, thank you for having me. I’m really, really pleased. Thank you, Jono. Thank you, Paul. Thank you, SEMrush - where are you down there somewhere?

Paul Lovell: Good work, everyone. Good work. There you go.