How to Help Google Make Sense of a Chaotic, Unstructured Web
The internet is host to a phenomenal amount of information. But it is, for the vast majority unstructured and very difficult for Google, Microsoft, Amazon et al to digest. To rise to the top you need to structure your content and make it easier for machines to understand. But how? The Brand SERP Guy will tell you how to best leverage Schema, HTML5 tables, semantic HTML5, knowledge graphs (not just Google’s!)… and even share my secret “Hummingbird content recipe” ????
Recorded live at Digital Olympus Conference (2019)
[00:00:00] The Brand SERP Guy (Jason Barnard): Super duper. I noticed that Deepak couldn’t remember my name, so I’ve done a very bad job. He had to look at the screen to find out who I was. I’ll be working on that from now on.
[00:00:20] The Brand SERP Guy (Jason Barnard): As he said, I’m going to be talking about How to Help Google Make Sense of a Chaotic Unstructured Web. The hashtag is my podcast. A little bit of self-promotion there.
[00:00:28] The Brand SERP Guy (Jason Barnard): And what I think a lot of us fail to realise is that the web is incredibly chaotic. We kind of look at it and as human beings, we make sense of it pretty easily. And we don’t realise how much of a problem it is for Google, especially what Google is now trying to do.
[00:00:47] The Brand SERP Guy (Jason Barnard): I’ve just realised having stood up here that I don’t have the thing to change the slides. So, I’m currently feeling very stupid.
[00:00:55] The Brand SERP Guy (Jason Barnard): Great stuff and I don’t know where or how they go, that must be it. Right.
Jason’s Approach to SEO: Understanding, Credibility, Relevancy
[00:01:01] The Brand SERP Guy (Jason Barnard): A friend of mine told me, do everything in 3s. He’s a teacher and he teaches people in the North of England and he says, if everything is in 3s, it always comes across and they remember it. If you have to, you invent a 3rd one just to make sure that the point comes across. I have understanding and credibility, I added relevancy. So, what does this mean?
[00:01:21] The Brand SERP Guy (Jason Barnard): It’s saying that what we’re looking for is to be relevant for whatever the person is looking for. We need to be relevant to the searcher. We need Google to understand that we have the answer and then we need credibility to convince Google that we are the answer it wants to show to its user and that’s another thing we often forget is we’re not talking about any other people, we’re talking about Google’s users, Google’s customers, and we’re asking Google to recommend us and provide our answer as the answer.
Knowledge Graph And Entities are Key to Helping Google Understand
[00:01:56] The Brand SERP Guy (Jason Barnard): So, today we’re going to be concentrating on understanding. I kind of went a little bit off track there. Understanding. How do we make or encourage or help Google to understand what it is we’re offering? That little icon at the top is a bit of design that I did myself. You can probably tell. It’s supposed to be a Knowledge Graph and for anybody who doesn’t know what a Knowledge Graph is, don’t worry I’m about to explain it. This is the World Wide Web, helping Google understand. Why? Knowledge Graphs, they’re the hub of answer engines. We’ve moved away from search engines. We’re now looking at answer engines. Google wants to give the answer to the question. It doesn’t want to give a choice. It wants the user to get the answer as quickly as possible then move on and look for something else and use Google again and keep generating revenue for Google. They’ve become, in my opinion, the hub of answer engines. It is the single most important thing to think about today. And we’ll see Dave Davis, who isn’t the guitarist from The Kinks, who is actually an SEO expert. He’s a really, really, really, really smart guy. Here’s a simple example, not Dave Davis yet. That was a little teaser for you to get you excited. Kate knows Mary. Mary likes Pete. Mary age 32. Pete brother of Kate. Pete born on 27-03-1982. What is that? It is people with relationships to either other people or things, information about themselves, such as their date of birth or their age.
[00:03:29] The Brand SERP Guy (Jason Barnard): There’s a nice representation of it. Makes sense. All of a sudden what we’re seeing is what we call Graph Theory, nodes with relationships. In the Knowledge Graph, entities with relationships to other entities. What is an entity? We’re about to find out. Very, very complicated explanation. Don’t worry about it too much right now.
[00:03:51] The Brand SERP Guy (Jason Barnard): You’ll be able to see the slides afterwards. You can probably watch the video again afterwards and you can read that in your own good time and you can go and look up @fluffymaccoy, loved her name. Just for that, she’s worth following. Dave Davis, here we go. “Entities are the single most important concept to understand in SEO right now.”
[00:04:10] The Brand SERP Guy (Jason Barnard): Brilliant! Full stop. That’s it. You can forget the rest. Well, not quite, but you know what I mean? This is your top priority from today other than building your own brand. Sorry Deepak. So, you’re going to go away with, I think 7 or 8 top priorities today. This is the one you should remember. Knowledge Graphs, Entities (think about relationships). You sir with the beard, me, our relationship, my relationship to the town I was born in, my relationship to the town I live in, my relationship to Digital Olympus. I am Jason Barnard, speaks at Digital Olympus, Entity, me, Digital Olympus, another entity, relationship, speaks at.
[00:04:55] The Brand SERP Guy (Jason Barnard): Attributes. My age, my height, where I was born, whatever it might be, and the relationships between these things. What I like is we’ve got E-A-R and it’s really rubbish. I thought of that this morning, it’s completely rubbish. So, you can forget E-A-R, but you do remember Entities, Attributes, Relationships, right? We know that an Entity is defined by Google as a thing or concept that is singular unique, well-defined and distinguishable.
[00:05:22] The Brand SERP Guy (Jason Barnard): Me, I’m unique as my mother always told me. Well-defined distinguishable, distinguished as well. It’s important to understand that it’s not just people, it’s not just places. It’s things like color, it’s concepts, color, money, price, tax. So, don’t just think of a Knowledge Graph as people with relationships or places with relationships to people.
[00:05:52] The Brand SERP Guy (Jason Barnard): It’s also colors. My t-shirt is an entity. It has a color that is red which is another entity. You’ve all seen this. I’m sure. Strings to Things doesn’t really mean very much until you think strings . What were Google doing, they’re analyzing character strings. Now they’re thinking in terms of objects, entities. Saying I want to understand the world so you can take away from this first part was a little bit complicated maybe. Google wants to understand the world. It understands the world using a Knowledge Graph and Knowledge Graph is a kind of encyclopedia for machines. Make sense? Hope so. The problem is how to build it. Manual curation, sorry, crowdsourcing. Easy. It reads with Wikipedia, it gets facts. It looks in Wikidata, it gets facts. It looks in IMDB, which is the movie database, gets facts. It gets all this information, no problem. But human beings have had to build that and that isn’t scalable. Google cannot understand the world if it’s asking you and you and you and you to check this information. This lady here was working at Google. Now, she’s at Amazon, everybody’s at it, but she started this idea of extracting data from the unstructured web.
Building Scalable Facts So Google Understands In A Superhuman Manner
[00:07:16] The Brand SERP Guy (Jason Barnard): And we’re going to move forward reasonably quickly because I want to get to the handy hints that will help you move forward. If we look at this, the top right-hand corner, Head Knowledge like a head keyword. Head Knowledge is the stuff that we have that has been curated by humans, that human beings have checked. It’s been fact checked, very important. We have an enormous amount of gray stuff that we have no, or Google has no access to. It can’t get the information because it cannot scale this human validation. So, understanding that problem is key to helping them solve it. As with everything understanding what, where the problem lies, gives the key to the solution.
[00:07:59] The Brand SERP Guy (Jason Barnard): This is the really, really, really exciting bit where I give you handy hints, tips, and tricks. She talks about Google extracting unstructured data, i.e. data or content that Google cannot fact check for itself easily through TXT, DOM, TBL and ANO. Doesn’t mean anything to you I’m sure or to some of you, somebody that does, I know, looking at Hannah Thorpe. What she wants to do, what she was trying to do at Google, and what she’s now doing at Amazon is be able to extract this information, the gray areas, as validated facts without human intervention, so the computer, Google, the machine is going to extract data and verify and check whether it is true or not, and insert it into the Knowledge Graph so that Google can then work on the idea of strings, not things, i.e. looking at understanding the world to return the answer that you want to your query and not relying on comparing chains of characters. And over time, if it can do it successfully, we will build up and if you’re asking, if you’re wondering what the dark green, the light green is, it’s dark green is fact, light green is probable. And the idea is that we’ll build over time and get more and more facts in this gray area around it and that is scalable, so Google will be able to understand the world in a semi-human manner, perhaps even in a human manner, perhaps in a superhuman manner because a machine is always going to remember everything, whereas we always forget, or I do anyway.
Properly Using Schema Markup on Your Website
[00:09:39] The Brand SERP Guy (Jason Barnard): Right, so we’ve got 4 aspects and I’m sorry, my friend told me 3, and Xin Luna Dong has completely ruined the presentation by giving me 4, and I haven’t got a choice, and I’ll have to do the 4. How can we best leverage Schema? Who knows what Schema is? Oh, that’s a brilliant start. Okay. We can have 4 questions as well, right? Schema Markup.
[00:10:02] The Brand SERP Guy (Jason Barnard): Have you ever thought about it like that? Who are you and what do you do? This is going to help Deepak, so we’re helping each other out here, mate. Using Schema to confer what Google has probably already understood on your site. You’ve explained who you are and what you do I would imagine. What you then do is use Schema Markup to confirm to Google that it has correctly understood.
[00:10:25] The Brand SERP Guy (Jason Barnard): So, I like to say, I mean, this is rubbish science, but it’s going from 40% confidence in what it’s understood to 60% because you’ve confirmed it. If you then go out and get other people to confirm it, it’s like NAPs in local search. Who knows NAPs in local search? Name, address, phone number. Thank you very much.
[00:10:44] The Brand SERP Guy (Jason Barnard): Consistency, confirmation, if you can get that going, you will end up with something that Google is 80%, 90% sure it’s understood, perhaps a 100% in which case Google will be terribly confident in presenting that information to its users as a fact, which is what it’s aiming to do. And if you’ve seen the Featured Snippets at the top of the results, for Google, in its mind, those are facts. The things it can rely on. They come from the Knowledge Graph. That’s very important to remember if you’re looking to get Featured Snippets, which Fernando was talking about earlier on. Little example, one of my clients, bit of advertising there for them, lucky them, very simple.
[00:11:21] The Brand SERP Guy (Jason Barnard): We all start with the About Us page and we say, this is who I am, this is what I do. This is in French, by the way. But in fact, every single page on your website could contain Schema Markup, should contain Schema Markup. Dave Ojeda was on a webinar series with me last year. He said, when you rebuild your website, start with the Schema Markup because it will help you understand what each page is communicating, what you’re trying to say and will help you with the categorisation, the structure of your website and it will mean that you are presenting pre-digested information to Google so that Google can say, okay, great, got this page content. Think I’ve understood it. That’s just confirmed that I’m away. That’s a fact or I’m confident that it’s true.
[00:12:11] The Brand SERP Guy (Jason Barnard): If you look up, the #SEOisAEO pretty much any subjects that I mentioned in this talk, you add that behind it, you will come up with a result like that. It will be a podcast episode or webinars with some of the people I am going to mention coming on because I’m going to give you the people you should be following if you’re interested in these topics. Please do take a picture of this screen if you’re interested in Schema Markup, these 3 are brilliant, well worth following, well worth listening to, very, very smart people, specialised in Schema Markup, which I’m not.
In HTML, DIVs Are For Design, Tables Are For Presenting Data
[00:12:48] The Brand SERP Guy (Jason Barnard): Tables and lists. We all think tables and lists are really simple but they’re not. How can we best leverage them?
[00:12:56] The Brand SERP Guy (Jason Barnard): This is very common. We’ve all seen this. What I love about it is that tables and lists are very old HTML. It’s stuff we were doing in 1998. I’m older than I look. Coming back into fashion now, Google needs this. It needs it because it’s structured. Because it can rely on the fact that, sorry, we’ve got columns, we’ve got rows, we’ve got a caption.
[00:13:21] The Brand SERP Guy (Jason Barnard): It can rely that that information plays a specific role, and we’ll talk about playing roles or the roles that content plays within the content later on. The other thing, very important to know here, is one of the biggest problems Google has had is 95% of tables are used for design by people and not for data.
[00:13:40] The Brand SERP Guy (Jason Barnard): So, it had to develop a system to throw the design to one side because it wasn’t structured data. It was simply design elements. So, if you have tables on your site, which are for design, change them to divs. Divs are there for design. Tables are there for presenting data. Another little handy hint, if you’ve got a table, describe it in text underneath. Why? 2 reasons: some people don’t like reading tables. They like to be given the conclusions of that table. The other is it confirms to Google that it is understood and it also gives it a choice of the way it wants to present that data to its users. So potentially, you can have a table in one set of results and the text in another.
[00:14:25] The Brand SERP Guy (Jason Barnard): This is really cool too. It’s actually exploring data research by Google. It’s an experimental table search engine and you can type in a term and it will pull out the tabular data that it has for that term. I gave an example here, a conference I’ll be doing in a few weeks. As a site owner, you can then check that Google has digested, understood and is confident it has understood what you have in your tables and of course your design based tables will not get in here.
[00:14:59] The Brand SERP Guy (Jason Barnard): 2 people to follow for that: Bill Slawski. If anybody knows him, if you don’t follow him, he’s super, super smart and he knows about all this stuff. He gets a couple of mentions, but obviously I can’t put him on every slide.
DOM-Digestion Gives Google Precise Information On What Role Each Piece Of Content Is Playing
[00:15:16] The Brand SERP Guy (Jason Barnard): DOM extraction. You all know what a DOM is? I hope so. Put your hand up if you don’t. Great. One person doesn’t. It’s the HTML, it’s the tags, the divs, and all that stuff. Good?
[00:15:31] The Brand SERP Guy (Jason Barnard): DOM-digestion. I was talking to Alex yesterday about this. I like the idea of Google going out, getting its food, collecting the food when food is supposed to be data content. Gets that food, takes it in, swallows it, digests it.
[00:15:49] The Brand SERP Guy (Jason Barnard): So, we’re looking for DOM-digestion which is a kind of cool phrase I thought I would share with you today. Consistency is the most important thing. What is Google’s biggest problem. It’s rubbish, HTML. It’s people who write different structured HTML pages for different parts of their site or they hand code it, which is what I do.
[00:16:09] The Brand SERP Guy (Jason Barnard): And it all goes horribly wrong because I’m not consistent. Human beings are not consistent. We’re rubbish at it and Google is very, very reliant on it. Google is very reliant on consistency. Why am I showing you this? Because Wikipedia is incredibly structured.
[00:16:32] The Brand SERP Guy (Jason Barnard): Every single page is structured the same way, so Google can read it and it can rely on the fact that it’s understood what role each piece of content is playing. Amazon. It crawls Amazon and it has had to understand how Amazon structured it pages because it needs that information. Yelp is the same thing.
[00:16:50] The Brand SERP Guy (Jason Barnard): You aren’t in that position. Your site is tiny. Google isn’t going to spend any effort at all trying to figure out what your rubbish page structure is. So, you need Semantic HTML 5. I won’t explain this in detail, but basically all we’re doing is saying, this is the header, navigation, at the bottom we’ve got the footer, not very important. On this side, we’ve got an aside, which means it’s not central to the content and in the middle, we’ve got the article, that’s the important content. The role that that plays in this page is the principal content that you should be concentrating on. The role that the footer plays is being the footer, the navigation has a role.
[00:17:33] The Brand SERP Guy (Jason Barnard): So, the idea is to say this part of the page has this role, so you can treat it in a specific way. So, I would advise there are articles all over the place here on this topic. I wrote one. This is another piece of design that I did. I’m not very good and I’ve got a pretty bad taste in color. Oh, sorry. One very important thing here talking about color is you’ll notice that the orange squares do not correspond to the Semantic HTML 5.
[00:18:04] The Brand SERP Guy (Jason Barnard): That is because the orange squares are design and this is Semantic, giving a role to each piece of content. So, every time you look at your page and the design and the Semantic HTML 5 are more or less the same or are the same, you’ve probably got it wrong because it is very rare that design and semantic roles that a piece of content plays in a page are the same.
[00:18:29] The Brand SERP Guy (Jason Barnard): AMP and WordPress. Why is that important? AMP is structured. It’s very simple. It’s very simplified HTML. Cindy Krum came up with an idea the other day that I love. It’s worth building an AMP site, not just for speed, but also because Google hosts it. It is very structured, very simple, and it can map your HTML normal desktop content to the AMP content and be confident it’s understood what content plays what role.
[00:18:59] The Brand SERP Guy (Jason Barnard): The other reason I’ve put WordPress up here is because Google is now employing WordPress developers to help with the development of WordPress with the aim of making responsive design and PWAs. So, what they’re doing is they’re saying we’re going to start structuring WordPress in much the same way that we have here. So, anybody who’s got a WordPress site, let’s say 3 years time, 4 years time, if you stick to best practices as set out by Google WordPress developers, you will be in a situation where it will be able to read your content, understand your content in a similar manner to this without you having to do that. So, do that today. If you’ve got a WordPress site, in 3 years time, you probably won’t need to do it anymore. Cindy Krum, Joost de Valk: both of them brilliant. Cindy Krum on AMP and Fraggles which is HTML 5. She’s a big fan of all that. And Yoast de Valk obviously big, big, big expert on WordPress.
In SEO Copywriting, Avoid Ambiguity By Using Words Of Relatedness And Co-occurence
[00:20:05] The Brand SERP Guy (Jason Barnard): Last little section. We’re coming into the final stretch, Free Texts. These are great words. Who knows what they all mean? Whoa, I didn’t 6 months ago. And then I talked to Dawn Anderson. She explained it all to me. Writing copy. Here we go. Does anyone remember that from earlier on? Entity, relationship, entity.
[00:20:32] The Brand SERP Guy (Jason Barnard): That’s what we do. Apple makes iPhones, bananas are yellow. It’s simple grammar, very simple grammar. It’s saying entity-relationship-entity. I keep saying it but I can’t say it enough. It’s incredibly important. I’ll let you take that photo. Not good. Why not? We have a phenomenal problem for a machine is that the entity “Sarenza”, the relationship “sells” and the entity “shoes” are terribly, terribly separated. There’s lots of fluff in between. There’s lots of useless words that the data appeal to me as a human being but the computer, the Google finds it difficult to understand these parts of the text when your texts are set out like that.
[00:21:22] The Brand SERP Guy (Jason Barnard): This is a really good example. I’ve said the same thing. I haven’t thrown away my wonderful copywriting skills, as you can see. I’ve got the same information but I’ve put my entity, my relationship, and my second entity right next to each other. Google will be able to pull that out and be very confident it has understood that “Sarenza sells shoes”.
[00:21:52] The Brand SERP Guy (Jason Barnard): Relatedness, co-occurrence, context clouds. The idea here if anybody’s using synonyms today, stop. Synonyms are old hat. The idea here is to create a context cloud using co-occurrence and words that are related to each other within the context that you’re using. For example, on the beach, these are words that tend to appear together.
[00:22:18] The Brand SERP Guy (Jason Barnard): So, Google would expect to see them in a page about a beach. If you then look at bay for example, in this context, it’s clear that it’s a bay of water, but it could be a bay in a delivery situation where somebody is delivering with a Laurie at Lowry Bay. So, the context words around that will be very different and refractory. It will be Laurie. It will be drive. It will be unload. So, the context gives meaning to the word “bay” and it knows what kind of bay we’re talking about. And you would be very surprised or you probably wouldn’t be very surprised because you’re all terribly clever, how many words have multiple meanings and how confusing it can be and how much if you just look at this cloud, the context becomes incredibly clear and all of those words mean something within that context.
[00:23:10] The Brand SERP Guy (Jason Barnard): June. I’ve just seen that one. I hadn’t seen it earlier on, it’s a film. In another context, you would have the actors, you would have the producer, you would have the film studio and so on and so forth.
[00:23:22] The Brand SERP Guy (Jason Barnard): So, it’s incredibly powerful if you can use words of co-occurrence and relatedness, you will give the context and you will help Google understand which type of bay, which type of June you are talking about. Once again, Bill Slawski. He reads patents if nobody, if anybody saw, who didn’t know that. He reads the patents, they’re really boring. And he writes articles that allows us to all just kind of read his articles and say, great, simple. Dawn Anderson. She was the one who told me about co-occurrence and relatedness, great people.
To help Google understand your copywriting, it is incredibly powerful if you use words of co-occurrence and relatedness since you’ll give context, and that helps Google immensely.
Jason Barnard, The Brand SERP Guy
Conclusion: Empathy For The Machine (Google)
[00:23:54] The Brand SERP Guy (Jason Barnard): So, the conclusion, the web as a database. If you think of the web as a database, it helps you understand how Google is approaching this problem.
[00:24:05] The Brand SERP Guy (Jason Barnard): It’s the biggest Knowledge Graph ever but it’s very badly organised and if you can help, if you can organise your little corner of the web, Google will thank you and when Google thanks you, it sends you traffic and that traffic presumably makes your money or you wouldn’t be here. What’s ironic here is I got this phrase, this screenshot from Diffbot and not from Google, but it’s a great way of presenting it. Google wants to turn the web into the biggest Knowledge Graph ever.
[00:24:36] The Brand SERP Guy (Jason Barnard): Last is Empathy for the Machine. I’m a big fan of the word empathy today. Obviously having empathy for a machine is not necessarily what we all want to do but understanding what they’re trying to do and what problems they’re coming up against and what problems this machine has to overcome to understand what it is you’re doing allows you to solve those problems for or help solve those problems for it, make it happy, so that it can send the traffic your way because it has understood that you are relevant, that you are credible, that you are the best answer, and you will satisfy Google’s users. Thank you.
[00:25:19] The Brand SERP Guy (Jason Barnard): If anyone’s got any questions, I’m happy to ask, answer. From my Knowledge Graph that’s up here in my head. No? I’m always disappointed when there aren’t any question.
[00:25:35] The Brand SERP Guy (Jason Barnard): Yeah. Ask me a question. Ask me a question.
[00:25:37] Audience: What about tools you would suggest for doing this entity, sir?
[00:25:44] The Brand SERP Guy (Jason Barnard): You are Italian. Yes. WordLift, great Italian company.
[00:25:48] The Brand SERP Guy (Jason Barnard): They’ve got a plugin for WordPress which identifies or tries to identify entities and relationships. And what they’re saying is build your internal Knowledge Graph and this is brilliant. I love these guys and Gennaro Cuofano, who was up there earlier on and a guy called Andrea Volpini are the people behind this. They’re saying build your own Knowledge Graph, entity-relationship-entity, and then you feed it to Google.
[00:26:10] The Brand SERP Guy (Jason Barnard): And they help you to do that using that plugin. Absolutely brilliant. I tried on my site, little story that I quite like, I wrote Jason Barnard full stop, double bass, full stop. How stupid is that? I was saying to Google, I’m a double bass and I meant to say Jason Barnard double bass player because I played the double bass.
[00:26:36] The Brand SERP Guy (Jason Barnard): And what you realised with this kind of tool, with WordLift in particular, is how badly we express ourselves. WordLift has trouble getting a handle on it. I spent hours going through it, trying to do it all, trying to link it all up, trying to link my entities with the relationships. Took me absolutely hours and I realised how rubbish my writing was, how silly some of this stuff sounded when you take it out of context. Brilliant tool. So, if you’ve got a WordPress site, you’re away with that. Thank you for the question. It makes me feel much happier. Thank you very much. Have a lovely afternoon.