In the latest episode of hack my growth, we’re talking about taxonomies and knowledge graphs and how we can use them to build a better site structure that serves both our users, as well as search engines.
Table of Contents
Video Transcript:
So in this video, we’re going to take a look at taxonomy and knowledge graphs and how improving our side structure can actually improve our SEO.
What is Website Taxonomy?
Well, taxonomies are used to classify things. They define relationships and similarities among a number of different terms and labels. Within a website, a taxonomy is a classification system that makes it easier for users to navigate and find the content they’re looking for. This is possible because a taxonomy uses labeled structures and it allows us to add some structure to otherwise unstructured content.
So in short, it makes it easier to organize, arrange, rearrange, search and distribute content various ways. Now this is taken from SEOptimer and you can read a lot more about taxonomies on that link here on screen.
What is a Knowledge Graph?
So what’s a knowledge graph, a knowledge graph, as we’ve talked about on this channel, it represents a collection of interlink descriptions of entities or objects, events, or concepts. Now knowledge graphs put data into context via linking as well as semantic data. And this provides a framework for data integration, unification, analytics, and sharing.
So as you can see, there’s a little bit of commonality between what we would call a taxonomy and then having a full-blown knowledge graph.
How do Taxonomy and Knowledge Graph Fit Together?
So how do these two things fit together? Well, taxonomies and knowledge graphs, they are complimentary and they are important parts of data ecosystems. The reality is our website is a database it’s full of information and data that needs to be structured for both our users, as well as our search engines.
Now a knowledge graph provides that structure to an enterprise domain and the taxonomy is what’s going to give us the hierarchal structure to key terms or relevant concepts within our businesses. So this is a lot of heady stuff and you can read all about this stuff and the links that we’re going to provide.
The important thing to understand is we can put these two things together to really improve the way that crawlers view our site and index our content as well as improve the overall experience for our users.
Creating a Knowledge Graph and Taxonomy
Okay, so now it’s time to put the theoretical aside and let’s talk about practical. How do we put this into practice? So if we want to create knowledge graphs and taxonomies, we need to do these seven things:
- Define the core topics your business is an expert in.
- Research the entities related to your core topics.
- Reconcile entities with Open Refine
- Create Knowledge Graph on your site and link all data.
- Use core topics as main site categories for your taxonomy (WordPress Categories & Tags)
- Tag related page with topics.
- Track Results and expand as needed.
First, we need to define the core topics that our business is an expert in, what are we known for? What do we do? Where is our expertise?
Then we need to research the entities that are related to those core topics. We may have an entity that’s related to multiple topics, we may have multiple entities within a topic, we need to do that research.
Then we need to go out and we need to reconcile those entities. And typically this is done using a tool like Open Refine, where you can search the linked open web, and you can pull all of that semantic information into either a CSV or into a database or whatever tool you’re going to use to really make sense of the data.
And then you want to create that knowledge graph with this data and make sure that it’s all linked in our website and in using all the other appropriate linked open data references we need to have, we need to make sure that we have triples and we need to make sure that we’re adding the right citations. And we need to make sure that we’re connecting all the different same as is and the linked open data.
And then we make sure that it’s available in a format that the search engines or the crawlers are going to understand. So whether this is going to be like Json LD data, or whether it’s going to be R DFA, and as you see this starts to be pretty complex. Now, thankfully, we use a tool called WordLift that does all of this hard work for us.
Now, the next thing we want to do after we’ve created the knowledge graph, we started building it on our website. We can use core topics as main site categories for our taxonomy. So WordPress is awesome because it allows us to do this for right out of the box, right? We can have categories and we can have tags.
Then we want to put these tags into play. So we can use these tags to tag related pages with our topics. Then we want to track the results and we expand as needed. Maybe we need to add more entities, maybe we need to add another level of tagging within our taxonomy to help describe a certain section more than another.
1. Define Your Core Topics
So let’s talk a little bit about how we can do this, and I’m going to actually walk you through how we would set this up using one of our sites.
So the first thing you have to do is define your core topics. As we talked about, you have to do this as your first step. You need to understand what you do and what you sell. And in this case, when you’re getting started, think broad and shoot for maybe three to four max, don’t try to push yourself into having 20, 30, or 40 different things that you do, because it’s going to get way out of scope really fast.
And you’re going to stumble and trust me, I know this because this is what happened for me. All right? So this is one of my websites and it’s a website that we’ve featured here a lot, right? This is the site that we put our courses on and in. And when we talk about this website, it’s focused on some very specific things. And so what I did to get started was build out the sheet and I said, okay, what are our core topics?
Well, we talk about schema.org. We talk about semantic SEO. We talk about structured data. Those are the three big topics that we talk about when we create videos. And these are what our courses are about. Now, what I could do is I could go ahead and use the semantic audit that we’ve created over here. And I could run the site through the semantic audit and start to understand the entities that we have.
So I’ve done a video on this and I will link to that video. And essentially when you get to the end of it, it’s going to give you some entity extraction using SpaCy. Now, one of the things that you might know is that when you get to this visualization part of SpaCy, you get things like orgs and persons in Cardinal and GEP all of this like really high level stuff, but it might not always be 100% related or connected.
There’s a lot of sifting that you have to do to either. There’s a lot of sifting you have to do to make sense of this data.
2. Research the Entities Related to Your Core Topics
So one of things that WordLift has recently done that I think is really stinking cool, is taking this a step further and actually pulls us here out of Collab and allows us to do some of this research, right within a Google sheet. So over here, you’ll see, I’ve got this little WordLift button here. Now, WordLift, what it lets me do is highlight one of my topics. And then it allows me to run a semantic audit of that topic within the search results. Now I’ve already done this, because it takes a few minutes to run, but at the end, it’s going to spit this information out. Now I can do a few things I can put in a URL or I can just put in a keyword.
Then they’re going to pull the rank. So the positions here, what sites ranking and positions one, two, and three. What are the different entities that they have? What type of entity that is, what the entity label is and then some of the linked open data resources here.
So I ran this keyword structured data and what WordLift has now done. It’s given me the entities most related to structured data within search results. So right now I know I need to add JSON LD. I know I need to look at semi-structured data because this is pulling up when we talk about structured data, I need to talk about schema and unstructured data and R DFA and data warehouses and HTML Text. All of these have a very high confidence level here with the topic that I’m talking about. Now just because they’re in here doesn’t mean I have to have all of them.
I want to organize this, right? So the page in position, 1 R DFA, JSON, HTML tag, structured data, all important stuff. Google search is showing up in here. These are all things I could do. Search engines are showing up here. Shopify. That’s interesting. So like I said, not all of these I have to have on my site, but some of them I do. Now here’s the other cool thing, when I use this, I’m using my WordLift key. So if my site has this entity, it’ll actually show my entity ID here, coming from word lift. So I’ll know that, okay, I already have that entity on my site. I don’t need to have, but if I’m just starting out, I could easily take these ones that make the most sense. I know that each of these are extremely important to me and I want to make sure that I talk about them.
3. Reconcile Entities with Open Refine
So I could start to say, here’s the entities that I need to create and I can go ahead and paste them down in here. So I’m starting to build my knowledge graph. I only have this information and I could go take this information and pull it into OpenRefine, and start to do all that data analysis, or I could go to WordLift and start to create these entities. So there’s a number of ways you can create new entities. You can either just add new entity here and start filling the information out yourself, or you can do it right within another piece of content that may talk about the entity you’re creating.
So I could go ahead, and I could fill this out and I could add new entity and just do it like a WordPress post. And then I’ve got all of the same as data that I can pull in here, which makes it super easy to do. And there’s even ways you can upload it, but I can also go into my content and I can look at something like this piece right here, and I can start tagging it and actually using this content to help me build the entity out. So I know one of the things that we were looking at was JSON, but this post may not have anything about JSON LD. So let’s look at one of the other ones that we wanted to create, maybe structured data, right?
4. Create Knowledge Graph on Your Site and Link All Data
So I could go here and I could see this term and I can highlight it and I can add structured data, see how they just did that. Now it’s going to give me a number of different things that I can can look at. And as I can see here, this one, this entity right here, structured data is the one I’m talking about. So I go ahead and click it. And there we go. I’ve just added a new entity. Not only annotating it to this piece of content, but I also added it to my knowledge graph.
5. Use Core Topics as Main Site Categories for Your Taxonomy
But now if I go back to my vocabulary, you’re going to notice something new there. I have something called structured data now. And here it is, and I can go in here and I can edit this.
And I can see that here’s structured data. It’s used to create links between information and post. So here’s a definition for what it means. And I have this tied into my website and it’s now something that can be tagged, but it’s also a linked open data reference point. There’s an endpoint now to this URL, which sends that structured dated at Google. So what does that actually look like instead of me just saying all these funky words, if I go to my blog article and I want to see what are the entities I have tagged here and I can run the URL in here, it’ll actually show me what I have underneath. So here we have the article, right? All the information. But under this article, you notice that we’ve got things now associated with it. We’ve got things like search engine technology. We’ve got structured data, right?
And not only do we have this vocabulary endpoint, which now we have within our database, we also have all the different mentions. And we have all of this structured information that the search engines are using. And there’s also this ID, which is the data in ID that pushes back to Google, that they can read about it as well. We’ve got a number of these things, knowledge, natural language processing is another one that we talk about here. Here’s all of that information that we pulled in here and built the relationship around. And if we go back to this little WordLift widget here, you’ll notice it’s very similar to all these links that they’ve pulled. So I could go in there and build JSON. I could put all of these links within that reference point and I can start to map it out from there and create all of these different entities.
6. Tag Related Page with Topics
WordLift is helping me build my knowledge graph. I have all this information underneath, that now I can use to tag or annotate my content like you just saw me do. I clicked on it and added it. And that’s going to help Google better understand what that piece of content is about. Now, if I want to build a taxonomy structure, I can go over to my posts and I can look at things like categories and tags. Now you have to decide how you want to use these. You can use categories, you can use tags, you can use both of them. Maybe you can use categories for more of your broad topics and tags for more specifics under there, or maybe use one or the other. There’s a number of ways to do it. So in my instance, I would probably go to categories and I don’t need article or uncategorized.
I probably should have thought through that a little bit more, but in this case, I want to use it based on the knowledge graph we’re building. So I would go here and I would say my categories would be schema.org, semantic SEO, and structured data. And now I can start to tag all of my content around these different categories and create logical groups of content that are going to help the search engines better understand what it is that I’m talking about. And I can use these categories to group my content for my users so that they can go, okay, this piece of content is about schema or this piece of content is about semantic SEO, or this piece of content is about structured data. So as you can see, we can build this structure and it’s going to make the user experience better, but it’s also going to help the search engines better understand how our content is put together.
If you got any more questions on taxonomies and knowledge graphs, and maybe how to practically do this, please comment below. I’d love to continue that conversation with you. And as you know, we’ve got a number of courses on both an overview of semantic SEO, as well as how to leverage structured data and really how to implement that on your site to get better results. And we’re also making a lot of updates to that semantic SEO course, to make it way more practical, to give you more insider tools, to do a lot of this work and to automate a lot of this work. But if you have any questions on what we talked about today, please comment below. And until next time happy marketing.
