Entity optimization, a crucial and evolving aspect of SEO, can help your brand dominate search visibility. From structured data to knowledge graphs, our guests will share insights and strategies to optimize your content for better results in 2024 and beyond.
Joining David Bain to share their key SEO tactics for Entity Optimisation is Jarno Van Driel, Andrea Volpini, Ulrika Viberg and Jason Barnard.
Watch On-Demand
Listen to the Podcast
Transcript
David Bain
What are the key entity optimization tactics for SEO? Hello and welcome to the November 2024 edition of the Majestic SEO panel, where we will be debating the key entity optimization tactics or SEO. So I’m your host.David Bain and joining me today are four wonderful guests. Let’s meet them, starting off with Ulrika.
Ulrika Viberg
Hello everyone. I’m Ulrika Viberg, and I am the CEO of the Sweden based SEO agency called Unikorn.
David Bain
Thank you so much for joining us, Ulrika. Also with us today is Jarno.
Jarno Van Driel
Hi, I’m Jarno Van Driel, an independent consultant, just representing myself.
David Bain
Thank you so much for joining us, indeed, and CyberAndy, let’s go with that. Pleased to have you joining us as well.
Andrea Volpini
Thanks. Thanks for inviting me. I’m I’m Andrea or CyberAndy. I am the CEO of Wordlift.
David Bain
Superb, thank you Andrea, and Jason is joining us as well.
Jason Barnard
Hi, thank you for inviting me, David. And Lovely to meet everybody for the third, fourth, fifth, sixth, seventh, 20th, 100th time. I’m Jason Barnard, CEO of Kalicube, and we specialize in personal brands and corporate brands, as represented by Google today and the AI tomorrow.
David Bain
Good to have you here, Jason. Well, let’s stick with Jason for a second, and Jason, I’ll just ask you the first question. That is, what is entity optimization?
Jason Barnard
Oh, a great question to start. Entity optimization is ensuring, or, in the context of SEO, is ensuring that Google, Bing and the AI, understand who you are, what you do, and who you serve, and the next extension of that is that you are a credible solution for the subset of their users who are your audience.
David Bain
Ulrika, you’re nodding away there, you agreed with everything that Jason said, I presume?
Ulrika Viberg
Yes it was smoothly said and rehearsed. Could you please say that again? Jason, so that we actually hear it once again, so we really hear what that means?
Jason Barnard
Yeah, optimization is ensuring that Google, Bing and AI understand who you are, what you do, who you serve, and the extension of that is EEAT credibility. Are you a credible solution for the subset of their users who are your audience?
Ulrika Viberg
I would also add to that entity optimization today and in the future, would also mean secure your place as an entity in the LLMs. And I guess that was also one of the things that we are going to talk about today, how you do that, but that is like the end goal and results of everything that we do today, and securing the place in the LLMs, and also in the Knowledge Graph.
Jason Barnard
I would add one thing to that, which is, actually I stated, what entity optimization is for a brand, a personal brand, a corporation, potentially a product, a music group, a film, all the same thing understand, they need to understand, or you need to make sure they understand who you are, what you do, which audience you serve, and that you’re credible.
But entity optimization goes much, much, much deeper than that, and much, much, much, broad. We talk a lot with that, about that with Jarno about the depth of the work you need to do to fully optimize entities. And the idea of EEAT, credibility, we call it NEAT credibility, which includes notability and transparency, takes it broader.
But then there’s also the question of, can the engines deliver you to the subset of their users who your audience? So we call it a Kalicube understandability, credibility, deliverability. So for me, it goes much further than just SEO entity optimization.
David Bain
And Jason, you mentioned Jarno there as well. Would you like to give your thoughts on a definition?
Jarno Van Driel
I’m not one very much for definitions, but that’s because I mainly move around in the world of unstructured data. So I try to make sense of those things that don’t make sense. And in regard from that, Jason, I align a lot with the things we do, the methodologies we use only focuses on the human side of things, whereas I mostly focus on the machine readable side of things.
The two of them meet somewhere in a messy middle, but those are two sides of the same coin. And Jason actually tells the story from the human side, and whereas I tried to translate it into a machine side of the story,
Jason Barnard
I really love that we’ve got the human side and the machine side and the messy middle is actually what entity optimization is. The messy middle is making sure that messy middle becomes clear for the engines and also for your audience.
Andrea Volpini
I think it’s good. I think it’s good that we’re coming out to looking at these two audience, the machine audience, the human audience, and we’re looking at this idea of concept. I think it’s foundational for building a good conversational experience to focus on the concepts, and that’s what entity optimization really is, is to look at concept and then to do it for both the mechanical audience, the robots and the human audience. In the LLM kind of sits in between, because in knowledge representation, an LLM would would use primarily words, whether you know, for example, a traditional crawler would rely more heavily on on machine friendly metadata.
I think that this is also the transition that we are experiencing from search to conversational search is that we’re moving, you know, from structured data to, you know, different level of unstructured data that still has to become as current as possible for sustaining training of language models. And that’s, that’s the beauty of doing this show now, because we have direct experience.
David Bain
I think everyone’s mentioned LLMs, Large Language Models. Does that impact where your entity home is, or it doesn’t actually change where your entity home is and how you go about optimizing that?
Jason Barnard
Sorry, the entity home is the concept of where the entity lives online, and all of these machines are looking for where the entity is providing information about itself to get their version of the facts, and your role as an entity optimizer is to create a place online, a page or a website that states the facts very clearly for the machines and indeed, for your audience.
And if they’ve got that, then they can do what John Mueller from Google calls reconciliation. I take all the fragmented information about you around the web, put it all together in the puzzle that should match the puzzle, or the finished puzzle that you’ve put on your entity home, and that is the key.
We’ve been talking about the hub and spoke wheel model. The hub is the entity home. Your website is the web page that represents the entity controlled by the entity that’s the initial version of the facts you link out using the spokes on the wheel to all the different parts of the web that confirm what you’re saying on trustworthy authoritable resources make sure all of that is consistent, and you have a puzzle that the machine can put back together, understand and be confident it’s understood. At that point, you’ve won the game.
David Bain
And these spokes, is it essential that the other sites are linking back to you as well, or is it merely the fact that they’re mentioning your entity and it’s obvious that it’s you they’re talking about, sufficient to drive that authority?
Jason Barnard
Well, initially they won’t understand who you are, so once they’ve understood who you are, then they will understand the mentions and realize it’s about you, but you always should be pointing to the corroborative sources that confirm what you’re saying, and if possible, they should link back.
But in truth, once your entity is understood, the idea of linkless links comes into play. Google understands you, Bing understands you, the LLMs understand you, and they can understand that a mention is about you, as long as the context is clear for them, and the LLMs can analyze that context and extract the entity for them.
David Bain
Ulrika, do you have any thoughts on the entity home page and how that’s structured and optimized to begin with?
Ulrika Viberg
I also want to add something else before I go into that, because I feel that SEOs usually jump in, right into the optimization of the entities, and we kind of forget what is an entity, and how do you decide what entities you want to optimize for, etc, whereas as SEOs, we usually do like the keyword research before we do the thing, the things that we want to optimize for.
But in this case, especially for me as an agency owner and a consultant, I need to push my clients to decide the brand values and the propositions and the place that they want to or who do they want to compete with? What kind how do they want the Google and the other being and everyone on the search engines? How do they do you want them to define you? What entities do you want to connect, be connected to your brand? And I think that is a huge part of the entity optimization as well as I guess that’s where it all starts. And also decide what other entities you would want to connect with, your primary entities. How do you want to build your entity home? So to say how do you optimize for it?
I would usually say that you have to start with the messages and everything, and they have to be consistent, and you have to have your tone of voice and that that should be consistent through your entity home, make sure that you have the basics in place. I guess that is a very basic advice. But here it is also very important to have the trustworthiness and everything in place and get that those mentions in places that you want to strengthen your brand. And I mean, that is the old SEO as well. So we are used to doing it.
I guess that Jarno will come in and speak about how you optimize it with schema and structure data, because that is, as I have seen, super important. I’ve gained loads of results using structured data for my clients, when I optimize them for the entities that I would want them to be seen as. Yeah, absolutely.
David Bain
I was gonna ask Jarno to come in next.
Jarno Van Driel
I normally come in when, when there are tons of unstructured data. So one of the first things I generally tend to do for projects, especially around entity home pages, is that I try to create a template of the way I want the information to be on the web page. So I end up with a JSON-LD file normally, that already gets thrown down the development chain pretty early on in the project, but those templates often only cover, or most of the time, cover more than there actually is present in their back end systems. So while the developers are working, I’m also having discussions with people in management, or people actually writing the content, so that we take that same blueprint I’ve made for the developers and actually turn that into information throughout the entire business which gets collected on the entity home page.
So I use my markup as a guideline to actually start producing the homepage information we need. And that’s not just on entity home pages. Like Jason also said, I can also count for product pages or places, services, you name it, everything can have its own entity home page. What’s important once you’ve established the foundational information that needs to be on every type of page is that you then start interlinking things, and that’s where it crosses over to the conceptional world of regular SEO only.
Semantic SEO is just that bit different, because the regular SEO is we want the keyword we want to rank for. We stuff that into the link, and we make the bridge to another web page. Whereas I look at links and I look at them as if they’re the prejudiced statement in the triple statement, so you have subject prejudiced object, and I look at links and I treat links like they’re prejudiced. So I may, not necessarily, put in the keyword we are looking for to rank nine to 10 times, I stuff the keywords in there that express the relation I want to build between those pages.
So that’s a using old fashioned HTML and natural language actually, to express the markup we have created in a much simpler way. But the markup itself isn’t really all that powerful. It’s the fact that we’ve turned that power, that markup into something humans can consume easily and can follow those links to get truly relevant and additional information. That’s how you build that hub and spoke wheel model Jason was talking about. So for me, markup is something, yes, I absolutely create in-depth templates, but the justification for those templates is this is the information we need to have on page.
Beyond that, if you start looking out on the web, it’s also curation of the information about an entity out there on the web. But once you have that blueprint, you can actually say, okay, do all those pages mentioning us at least contain the correct information, so that you can align everything. And once everything is finally aligned and you have your baseline entity information done throughout the web, then you can start expanding on that. My methodology for that is mostly to use markup. The majority of that markup is not recognized by the search engines, but that doesn’t matter.
Schema.org provides a beautiful blueprint of the types of information you could find. You can also look at if it’s a highly specific industry like metal machining, there are ontologies out there with entire vocabularies of information about ontology about metal machining. So you can use those ontologies as well as inspiration for the content you’re producing on page. But also the information, the keywords you want to be associated with, and which entities belong to that you can actually easily find that other vocabularies and schema there.
David Bain
Why would you want to use markup that isn’t recognized by search engines?
Jarno Van Driel
It gives you an information template. So one of the trickiest things with entity homepage is always, what does it need to contain? And that differs highly per industry. What are you describing a business, a person, entity or a location? So you need to have some kind of blueprint in your back pocket that covers the majority of information you need every entity page type to contain.
Why look schema.org? Well, schema doc started with a group of people from the different search engines that imagined together what would be valuable information for the future. That doesn’t mean all of schema.org is being used, but the preconception of those terms is this might be valuable information for search.
So even though not everything is being used, it gives you a form of guideline. And if you get into that, into the vocabulary, you actually start to imagine, okay, this is the type of information the search engine would like to present to its users. And if you have a good picture of that, then you actually can start to extrapolate that and build upon that by making new types of relations that aren’t necessarily in schema.org, but that align, or you could consider sub properties of those properties already existing in schema.org So you’re using schema dork as a guideline.
David Bain
Talking about sharing valuable information. Andrea has got a slide that he’s going to share with us. Andrea, if you could possibly just talk us through it, so it’s valuable and relevant for the audio listener as well.
Andrea Volpini
What I’m sharing here, it’s the extraction of the of the structured data from Jarno’s personal web page and basically, I’m using our agent for doing three tasks to analyze the entities that are present within the text of the Jarno personal web page. I’m extracting the embedded HTML metadata. These include the JSON-LD, the microdata, the RDF, the Open Graph, whatever it’s available there and then, I’m asking the agent to to build a mind map, because I think that visually, we can more clearly understand that is an important relationship. Information is is transferred by building relationship.
When we deal with entities, whether we do these with the textual mention that would be grabbed by a search engine or by the training of a language model, or it’s embedded within the structured data, it’s by creating this relationship that we can get to know more about a person or a thing.
So in this case, we do have a lot of information from this chart that has been designed and extracted from the structured data. So point being, relationship matters, whether they are present within the text or they are represented within the structured data.
What I was representing there is the structured data, because that’s how Jarno understand and design the content and I think that when we optimize entities, we have to start by looking at relationships.
Ulrika Viberg
I totally agree, and the relationships between the entities are where you actually find the secret sauce, how you color your own entity or the entity that you’re optimizing with other entities, and the the syntax and semantics of that entity and how they they interplay with each other.
I think that is still super exciting to see when it happens, when I see how things are, the relationships that builds between things that have some, some kind of connection or relationship, and then you see where, okay, now they understood there’s a relationship between this and how the entity that I’m optimizing, how that sort of changes in its semantics. I really like to see that.
Jason Barnard
It comes right back to what Jarno was saying, that if we perceive or look at links as relationships, it’s really helpful. Relationships are what’s important. The links bring power to each of the two entities, the subject and the object, and in this case, the relationship is bringing that power. So the relationships, as you rightly said, focus on the relationships between entities that the machine has understood. Very importantly, because if it hasn’t understood one of the entities, there isn’t a relationship to be understood. So that understanding of entities is vitally important.
And I’d just like to kind of add one thing about schema markup with. Talking a lot with Jarno about this, schema markup isn’t as powerful and helpful as a lot of people think, and it’s losing its effectiveness as Bing and Google are not developing further their support for schema markup and relying much more now on LLMs and SLMs to understand what’s in the page.
And a lot of that comes down to the power of the LLMs and the SLMs and their ability to extract entities and relationships between entities with precision. And part of it is to do with schema markup isn’t as implemented as we think, as broadly implemented as we may maybe think within our industry, because we see so much of it, it’s badly implemented, and there’s a lot of spamming going on. But like the keywords meta field in the head, I would imagine that at some point in the not too distant future, Google and Bing are going to be using it significantly less. It won’t be as bad as keywords, but that’s a good analogy.
Jarno Van Driel
I think I had some chats with several people, both from Bing and Google, and in their mindset they need to be able to make sense of an Internet without markup, because the majority of the Internet still contains its information without any supportive markup.
Having said that, there are some very specific benefits to markup when it becomes when it comes down to enriching the search results, even for AI overviews, even though we’re not seeing, as far as I’ve seen, we’ve not seen star ratings and things like that show up yet based on structured data. But I think that’s just because of the current state generative AI is in that all those companies are working really hard to develop new feature that’s a matter of time.
The big advantage for markup in the search results, regarding the enrichment of search results, is that it provides search engines with a very specific data set that allows them to populate very specific features. It costs them very little processing power to achieve that, and they don’t have to do any post processing either. It’s like, here’s the data set, populate your search result. You don’t even need to take it into account for your algorithms, you can do it even at the moment of rendering your search results.
So from that point of view, for a search engine, it’s a very cheap method of enriching the search results, and LLMs are way too expensive to be taking over that part of the search results,
Jarno Van Driel
Enriching search results is moving us off the topic, which is entity optimization. And sorry, Jarno, to interrupt you there, but the carrot, as Andy once said, it’s the carrot to attract us towards schema markup, those snippets, but entity optimization is not involved in that particular part that’s getting the lovely stars that Google needs to be able to understand more deeply.
Andrea Volpini
I can share a couple of beats here, because I think that this is a very interesting discussion, especially for for people like like us. I mean, most of the people today are heavily invested in structuring information and entity optimization. So we’ve been coming a long way before and understanding the actual support of search engine for language like schema.
I want to expand here based on my latest research that would be soon published with the new web almanac that also Jarno is contributing to, because it’s interesting, when you look at millions of websites, how structured data is being used, regardless of the support of the search engine, evolving over time. And it’s also interesting to note that there has been a lot of updates from Google itself about attributes that should be added in structured data.
So while on one side, we all agree that it is definitely less relevant than most of SEO would think, at least in terms of relationship between the information extraction and the usage of the mock up. We have to consider two factors, one, that Google is still heavily investing on it. And for example, if we look at e-commerce and product entities that are on the rise. I mean, of course, as Google transitions more into becoming a marketplace, there is a heavy, heavy investment on product entities, and we could see that there is a lot more support for additional data for products in the developer guides of Google, because Google still needs that data to be structured. So whether, on one side, of course, SLMs are representing a way of aggregating some entity attributes.
On the other side, structured data still remains extremely relevant for some use cases, and here is where I want to get. For example, in Jarno’s vision, structured data is a way of building the scaffolding for content. In my way, structured data right now, it’s about creating a conversational experience on my client’s websites. It’s about building a knowledge graph that I can use for content generation. It is less about triggering the search engine rich result, which still pays back. But again, it’s a carrot and we don’t want to depend on the Googles. We want to build efficiency for our audience, we want to optimize the experience for our clients.
So schema, it’s really an ontology. It’s really a vocabulary that we can use with different use cases. And so on one side, the usage of it, it’s growing in complexity, so we have way more refined markup this year than we used to have last year. And on the search engine front, especially if we look at e-commerce, there is a lot more attributes. We now have loyalty, membership points. We now have, you know, certification. We now have the GS1 Digital Link.
All of this is in schema, and it’s supported by Google and requested by Google, and it wasn’t there a few months ago. So it is transitioning, and some things are less relevant, of course, for the search engine overall, for us, it’s the skeleton of how we build information.
Jarno Van Driel
What I think is becoming interesting right now is that because of the Digital Markets Act in the EU, we now also have the digital product passport specifications in Europe, version 1.0 somewhere in next year, we’ll have the digital product passport for version 2.0 for apparel, clothing.
Why is that important? While the EU is creating those vocabularies because they want to force manufacturers and brands to publish that information digitally, so that, for example, they can get rid of that toilet paper roll label within each t-shirt. But because the EU is starting to mandate demand this information digitally from manufacturers.
There’s a growing interest from search as well. So the search engines, whether or not they’re going to be doing it through markup or through natural language processing, doesn’t really matter in the fact this is that they’re going to be taking those new vocabularies into account one way or another, simply because legislation is forcing them to.
David Bain
Ulrika, you talk about the linguistic nature of entities, what do you mean by that?
Ulrika Viberg
I actually studied linguistics, that was my university degree, and then I become an SEO, and finally, I sort of tie the loop together and get to use my knowledge in SEO and in linguistics, and then I really like that.
So for me as a linguistic entities are not just the keywords or the nouns or how something is described as things in the real world that exist, because it’s so much more, it has a meaning. Imagine a bucket of meaning, and in that bucket you have variations of this word, of the semantical meaning of this entity, but it can also be like local meaning, but it can also be variations of the grammar, and they can also be the variations of the relationships you have with other entities and relationships on the internet and it changes this bucket of meaning.
I think it’s better to understand it like this, especially when you are using it for brand optimization or digital brand optimization, because what it means is that your entity is going to be had this, the semantical meaning of your entity is going to change dependent on what kind of relationships, as we talked about earlier, the relationships you have with other entities, like values and things that you like to tie to your brand, that you want to be connected with.
And why is that important for SEO? You might say, because we haven’t been investing tons of effort into brand optimization in the past, but now this is the new. Anything for us, it’s because you want your brand to be connected with the searches that your your customers or your audience is making, whether it’s on Google or in on a AI driven system, or where you want to want it to be happening. You want your brand to be connected with those things that actually taste good and the clients are searching for because they might not search for your brand initially.
That is not how search works, really. So that’s why I start talking about entities and try to describe the value of having done your homework as a brand or as a company, not just put on content on your page that is randomly, sort of loosely connected to what you do, or what you think your audience want are looking for, which is, you know how we did it in the past, but now, instead creating content that enriches your brand and your the perception of your brand and the things that you want to be connected to how your audience how you want your audience to see you, and so you sort of dress yourself with nice things that you want to shine with and the things that you want to be connected to.
David Bain
We’ve got a question for Jarno from Jeannie Hill: How do you see these new data vocabularies being more implementable in 2025 and the role of the GS1 Global forum.
Jarno Van Driel
No, the vocabulary is currently being built within Europe. We won’t see those in search. One simple reason being, those will cover things manufacturers will need to communicate to each other and within the EU, what will likely happen is that it’s those vocabularies are being built and as an extension to schema.org sort of multiple vocabularies are being taken into account in their designs.
But in practice, if you look at GS1, it has its own vocabulary, and that’s a one on one extension of schema.org so you can enter into mix those, and that will probably happen over time as well with the vocabulary, the vocabularies the EU is developing right now, and if it really becomes important for the search engines, well, the pragmatic solution always has been, okay, that works, we will integrate it into schema.org. So for example, good relations vocabulary was integrated, and because of that, we can describe products and services in schema.org but there are also sometimes parts of the GS1 vocabulary where we then take over the entire vocabulary in schema.org.
Now we adopt certain properties the GS1 already worked out. Those are added to schema.org and they’re mapped to each other, so that the search engine knows whether you’re using the GS1 vocabulary or you’re using the schema.org one. They’re linked to each other so we know what it means.
Schema.org is turning into a sort centralized hub of vocabularies. It doesn’t necessarily describe everything itself, but it’s being built upon and being built against, so that things become an extension of schema.org and whether that’s being used by search or not, is not really the reason why it’s being developed, because all those extensions have their own use cases, but they come together around schema.org because that’s now become a public standard.
David Bain
Can we talk a little bit about why we’re marking up data and why we’re optimizing entities? Obviously, it’s so that search engines like Google can better understand who our brand is, what it represents, but it’s so that information is displayed on the SERP and in other places. So where are the key areas that we need to determine that we have optimized successfully for, and how do we track success?
Jason Barnard
For the whole kind of product area, I realize listening to Andy and Jarno, how little I know about schema markup for products and products in general, because we optimize the content creator and the content publisher entities.
And I’m talking now in SEO three tiers, content level, which we’ve always been doing and now we need to optimize the content creator, which Google now recognizes explicitly. It can track them, it can apply credibility signals, whatever they may be and the website publisher and that trust in the website, publisher and the content creator are phenomenally important. If you’re not doing that entity optimization today, you’re missing out on two of the big tiers of the three tiers of SEO.
So at Kalicube, we just focus on corporate brands, personal brands, how do the machines understand and represent you, and we’ve now got a new tracking system where we track assistive engines, Chat GPT, Gemini, Perplexity over time, and we see how their perception of you and their description of you has changed over time. And what we find is when we optimize over time, they become less vague and more and more precise and more and more accurate in what they’re saying. They’ll go from entrepreneur to business strategist to business strategist in the aviation industry.
And for me, that’s super exciting, and that shows that we’re really getting the machine to understand. We’re educating it like a child, and it’s understanding, and it can represent us, and we can direct it and educate it so that it correctly understands us and can represent, as Ulrika was saying earlier on, the brand narrative that we defined. And I loved when you said, we all jump in feet first with entity optimization, and we actually have to define what we’re optimizing and what we’re trying to communicate, and if you haven’t done that, don’t even start with the entity optimization for the website, publish or the content creator, personal brand and brand, you need to define it to start and Ulrika, that’s a great point.
Jarno Van Driel
The only thing I think for that’s really changed for enterprises is that they can no longer treat the internet as a trash bucket. Everything out there needs to be optimized, historically wise as well. So if you got documentation about your business of 10 years ago, it needs to be curated.
A lot of organizations looked at marketing as this year we’re doing this jump throw it over the fence. Next year, we change strategy, throw it over the fence. And organizations, and especially corporations and enterprises, are now faced with the fact it’s time to clean up your mess, and if you’ve done that long enough, it’s going to be a costly and highly annoying job. Yeah, that’s how you were as a business as well. You were causing people a lot of work. You were highly annoying with your messaging, and now you’re forced to clean that mess up.
Jason Barnard
Which is lovely, because when we’ve got the Kalicube Pro platform I built. It’s all designed to clean up the digital footprint. And the way I say clean up your digital footprint, you say clean up your mess.
David Bain
Jason, what’s the ROI of cleaning up your mess?
Jason Barnard
Unfortunately for SEOs in general, it’s kind of medium term to long term, and the brand needs to understand the importance of the clarity of the brand narrative across the entire digital ecosystem, the importance of their brand and the value of their brand. And if they’re looking for ROI from website visits and direct sales that they can identify from those website visits from Google, then they’re not going to invest in this particular area.
But what we find at kalicube is that when you optimize an entity for SEO, spread it out towards the credibility, and then move towards deliverability, get it out there, you end up naturally standing where your audience is looking, LinkedIn, YouTube, Forbes.com, search engine, and wherever it is, you’re publishing the content you’re demonstrating to that audience, when they’re looking at your credibility as a trustworthy source of a solution for their problem, and you’re inviting them down the funnel, and Google and the AI see this, and they will replicate it.
Andy once said to me, the web is just another representation of the real world, a powerful representation. But what Google and the AI engines are trying to do is replicate human behavior, and if you can demonstrate you’re walking the walk, standing in the right places where your audience is looking, demonstrating credibility, inviting them down the funnel with a real solution to their real problem that brings real value, you’ll win the game.
David Bain
So we’ve talked about Google and the AI as being opportunities for entity optimization, where your brand will appear. Let’s just break down AI just very slightly and just just do a very quick one stop piece of advice, maybe from each of our experts, where, from your perspective, is the key AI opportunity is that particular LLM is an AI powered search engine that you can recommend, and how do you go about optimizing for that?
Jason Barnard
I would say, if you’ve got a strategy designed for a particular engine or a particular platform, then you’re making a mistake. The strategy you have needs to work for all of the platforms, all of the technologies, or you’re going to get caught short by picking the wrong horse, or the technology changing over time, or potentially a penalty, whatever it might be.
I talked to somebody at a conference who said, Oh, I’ve got my strategy for energy optimization with Google. Now I’m going to figure out the one for Bing, then I’m going to figure out the one for Chat GPT, and my answer to him was, it should be the same strategy for all of them. They’re all serving the same client base with the same technology, the same data source, and the same aim, which is to bring their user to the solution to their problem as efficiently as possible.
Your strategy needs to make sure that the machines understand who you are, what you do, which audience you serve, that you’re a credible solution, and that you give them the material they need to deliver your solution to the subset of their users who are your audience. And if you do that on the web using the entity home, all of these machines will get it. It doesn’t matter who wins the upcoming search and assistive engine war.
Ulrika Viberg
I mean, also, like soon in the future, we will see new AI driven systems, right? That we don’t even know how they how they work yet.
Jason Barnard
Google learn. I’ve been playing with that. Andy showed it to me or pointed out on Twitter the other day. It’s so much fun.
David Bain
Okay. Andrea, do you have thoughts?
Andrea Volpini
I want to kind of get back on the return of the investment part, because I think that looking at entity optimization for AI, it has to be grounded into what is the return right? And I fully agree that it’s one process for old systems, I tend to disagree that this system have similar technologies, because as we’re seeing, the technology behind GPT search in regards to entity understanding, it’s way behind what Google has. We can see these also on transactional queries as well as in local type of queries.
However, in order for us to look at concrete return, my advice is always to build data you can also consume. I mean, when you’re building data or annotating a web page, or you are building a knowledge graph, or you are injecting structured data, or you are lexically designing the content around your core entities. You are not just doing that for you know the top of the funnel. You are doing that also because you want to drive a better customer experience within your properties.
So my advice, in order to maximize the return, is to work for yourself as if you were Google or GPT, because then you’re capitalizing on the way in which you’re organic organizing the data, and that brings a lot of value. Example for e-commerce, work on on your product data. It’s immediate. You can measure a return of 5x for structuring data around the product. And it’s super simple. And it’s not just about, you know, the top of the funnel, again, it’s also how I’m improving the internal search. How am I improving the ability to drive the user from page A to page B with relevancy? So that’s kind of my advice.
David Bain
This conversation certainly needs a part two. I think we can keep on going for another hour or so, at least, but in the meantime, I’d like to ask each of you to share what you think is your number one action will take away from today’s discussion and at the same time, tell the listener who you are and where they can find you. So let’s start off with Ulrika, what would be your number one takeaway from today and where can people find you?
Ulrika Viberg
I really want every business owner or organization that in the future is going to hire me or any other SEO to listen. Please have your brand platform in place. Please work on who you are, what you want to do, who is your who is your audience? Where do I start? Who where? What is your value to the world, and what is it that you want to say to them? Have that first, and then we can help you. It’s going to be so much easier.
I don’t I actually don’t use Twitter so much anymore, so just search me on LinkedIn or email me at ulrika@unikorn.se.
David Bain
Thanks so much for joining us Ulrika, Andreas?
Andrea Volpini
My main takeaway is to focus on building the best possible data. Because whether you’re big or small, to grow, you need to, you know, have the best data available.
You can find me on X as CyberAndy or Andrea Volpini on LinkedIn, or just drop an email to andrea@wordlift.io.
David Bain
Thank you so much for joining as well Andrea, Jarno?
Jarno Van Driel
I think my takeaway would be don’t stuff everything you find on schema.org in your markup. Be very specific in what you provide in markup. Plainly said, focus on Google’s feature guidelines. That’s probably where you’ll find your direct ROI.
But beyond that, take schema.org into account so that you become accustomed in thinking in data structures, because in the end, the thing you explain through machine readable markup needs to be translated to human content. So you need to learn at those machine readable formats to understand, okay, how do I actually need to make up my actual content so that NLP is able to extract that information from my content, and whether you then actually also provide the markup that doesn’t really matter. Use that markup as a means to optimize your content. That’s probably the biggest takeaway.
But don’t worry all too much about stuffing everything in there. And people can find me online, on Twitter/X, BlueSky, Mastodon, and LinkedIn simply as Jarno Van Driel.
David Bain
Wonderful. Thank you so much. Also with us today was Jason.
Jason Barnard
My big takeaway is that I need to focus more on relationships between the entities. I focus a lot on the hub and the wheel, and I don’t do enough focusing on the relationships, the links, implicit or explicit links, doesn’t matter, but that’s my takeaway that I’m going to take away.
I actually wrote a couple of notes down here of things I need to change in my approach. So thank you, Ulrika, Jarno and Andy, for pointing me in the right direction. And Ulrikas point about define who you are and what you’re trying to communicate before you try to start optimizing it, because if you do it all and then have to roll it all back, it’s double the cost. So start with a good idea of what you’re trying to communicate.
And if you want to find me, this is my fun thing. Google me, and if you don’t like Google, and you’re up for the future and you’re ready for Chat GPT, you can just ask Chat GPT.
David Bain
I’ve been your host, David Bain, you you’ve been listening to the Majestic SEO panel, if you want to join us live next time sign up at majestic.com/webinars and, of course, check out our other series at SEOin2024.com.
Previous Episodes
Follow our Twitter/X account @Majestic to hear about more upcoming episodes!
Or if you want to catch up with all of our previous episodes, check out the full list of our Majestic SEO Podcast episodes.
- Get onto Bluesky. A Guide for SEOs - November 18, 2024
- October 25th, 2024 – Majestic turns 20! - October 25, 2024
- Key SEO Tactics for Entity Optimisation (Live Podcast) - October 24, 2024