Podcast: Play in new window | Download
Subscribe: Spotify | Email | RSS | More
Jason Barnard at Digital Olympus 2019
Jason Barnard talks about The Google Cookbook – Making Content Easy to Digest
Google’s food is information. It needs to identify, collect, chew, swallow and digest in order to be able to give the answers to users. We look at leveraging Schema.org markup, Dom extraction, semantic triples, tables and lists to prepare Google’s food. We meander through a lot of questions and come up with some interesting explanations. Schema markup is like recipes and food items (that analogy doesn’t fly for very long). When you rebuild your site, start with thinking about structured data, since that encourages us to better organise categories, pages, and even Fraggles. I realise that I have been saying that I am a double bass. Alex gives me a taste of my own medicine by asking a question I wasn’t ready for. I wriggle through by quoting Jono Alderson and Cindy Krum – chunks, blocks and Fraggles. I cite so many people, it is starting to feel that I don’t have much to say for myself. Conclusion is “Help the Google Beast / Pet” and it will help you.
Jason Barnard
SEO is AEO. Welcome to the show. Jason Barnard.
Alexandra Tachalova
Okay, so today I have a pleasure to interview Jason Barnard. And so, I’m very excited about that, and I’ll try my best. We’re going to talk today about The Google Cookbook – Making Content Easy to Digest, which is actually a very big problem because it’s very popular to have all those long-form content, but it’s really hard to … Just to read them. And so, I guess that’s a very hot topic nowadays.
Jason Barnard
Yep, it’s a very big topic too, and I’m going to talk about it tomorrow, so I’ve prepared it all, finished the slide deck this morning, and I’m ready to rock with this one.
What do you mean by digest?
Jason Barnard
I like the idea that Google is having trouble, not only collecting its food, which is information, but also swallowing it, and then digesting it, so that it becomes energy for Google. Isn’t that a lovely, lovely idea?
Alexandra Tachalova
Yeah, very, very, very good kind of comparison. Really.
Jason Barnard
I just made it up. I hadn’t thought about that one, which is really stupid of me. Because I wrote the questions. Yeah, so it needs to identify, collect, swallow, and digest all this information to become energy, to be able to give the answers to the users. And that’s a phenomenally big problem for Google.
Alexandra Tachalova
Okay, so it’s just more about understanding what’s going on, on particular pages and giving the right results to people. Based on this data.
Jason Barnard
Yeah.
Alexandra Tachalova
Okay. So, you talk about four main focal points. Let’s go through them one by one. Starting with structure data schema, which is very popular right now- tell me how it relates to digesting by Google.
Jason Barnard
Well, the structure data, as we all know, just confirms what’s already on the page, so Google would’ve swallowed rather the information on the page, even though it wasn’t structured. But it won’t be fully confident it has understood it. So, you put the schema markup, and then it becomes incredibly confident and that’s what I would call digesting
Alexandra Tachalova
So, ingredients really. So, “this is cucumber and it was organic”.
Jason Barnard
Yeah. Exactly. So yeah, you can give it all the information … Break your food down into an ingredients. I don’t know how far this is going to fly as an idea, but we can keep trying. But you break it down. It’s name value pairs, so it really knows what you’re talking about. And one thing I see is the people go, “Okay, great. I’ll use it.” And what they don’t realize and probably what they don’t do, is use it all over the place. You have somebody like Martha van Berkel who says, “Use it on every page.” Bill Slawski will tell you the same thing. Aaron Bradley will tell you the same thing, and they’re all right. It’s incredibly difficult to do, because it’s time consuming, and it’s not always easy.
Alexandra Tachalova
Does it raise any kind of conflict when you just go everywhere with structured data?
Jason Barnard
Well, you should be able to put it on every page, because you can explain almost everything with structured data. The only problem comes if you say in structured data something that isn’t on the page, which is a mistake that a lot of people make. That’s spam. We’re trying to cheat the engine again. So, obviously you want to stay honest, and Dave … I can’t remember his name… Ojeda, I think it was… Was saying, “When you’re building your new site, start thinking about structured data and build it from the structured data upwards, because if you do that, you will identify how your pages are structured. What’s in them and how the overall site is structured, so-“
Alexandra Tachalova
So, you need to start from structured data.
Jason Barnard
If you’re doing a new site, start with structured data. That was his advice. So, you’ve got not only your ontologies, categorizations, that idea of splitting your topic into silos, but then you also have structured data-
Alexandra Tachalova
But what about CMS systems, because a lot of people just don’t build a website from a scratch. They just use WordPress or Joomla, or something like that.
Jason Barnard
Yeah, okay. So, thinking about your new site, you’re obviously going to think about categorization. When you’re thinking about that, think about what schema you’re going to put on each category, and that will help you understand when you’re going wrong with your categorization, because it will help you spot the error. I used WordLift, which is Gennaro Cuofano and Andrea Volpini, who are brilliant guys. And I put it on my website, and they basically do this structured data and they’re trying to put everything as entities with relationships.
Alexandra Tachalova
Yeah.
Jason Barnard
And I realized how badly I explained everything. I wrote Jason Barnard, double bass. And that means I’m a double bass, and I’m not a double bass. I’m a double bass player. So, I’m using the thing for what I’m doing with that as it were. And that’s very, very confusing for Google. Jason Barnard is a double bass, which is incredibly stupid. Jason Barnard is a double bass player, is much more intelligent. But as human beings, we make that link between double bass and double bass player where it’s obvious what we’re talking about. And so, I went through my website-
Alexandra Tachalova
But it’s not for a machine
Jason Barnard
And realized how rubbish I was, and I should’ve started with schema. So, that’s the first one. That’s incredibly big, and I can’t stress it enough, but I think the people who are listening to this already know that. Next point.
What is DOM extraction, and why does HTML5 help?
Jason Barnard
I love HTML5. I think I’m the only fan in the world of Semantic HTML5. Maybe not. The biggest fan in the world of Semantic HTML5. The idea that Google can go into a page and immediately will identify which are the important parts so the content is completely ridiculous. Extracting from DOM’s is very difficult. The DOM structure becomes very quickly very complicated. HTML5 obviously identifies which is the central content, which is the menu, which is the side bar, which is the nav, so on and so forth. And you can say, “Look, come straight here.” And if you look at sites like Amazon or Yelp or Wikipedia, they’re structured incredibly always the same way. So, Google will take the time to figure out how they’re structured, because it needs that data, and those sites are so phenomenally big and so phenomenally important, that it has to do it. But on my little site, it obviously doesn’t.
Alexandra Tachalova
Mm-hmm (affirmative)-
Jason Barnard
So, I need to give it Semantic HTML5.
Alexandra Tachalova
Mm-hmm (affirmative)-
Jason Barnard
Cindy Krum also mentioned a really nice example of using AMP pages, because AMP is very simple and structured. So, you end up with a situation. If you have an AMP pages and a normal HTML page, Google can look at the two, compare them, and confirm to itself that it’s fully understood your HTML by using the AMP as a template… because AMP is structured, and because it’s hosted on Google, and because it’s very simple. Love the idea.
Alexandra Tachalova
Okay, so it makes sense.
Jason Barnard
Cindy Krum makes a lot of sense a lot of the time. All the time. Oops. Sorry, Cindy.
Alexandra Tachalova
So, I’m just wondering about categorization inside eCommerce websites. So, do you need to save the same navigation when it comes to just browsing via Google when you try to keep it structured? I mean, across big eCommerce websites.
Jason Barnard
I love the question. It wasn’t on the list, so you’re cheating here.
Alexandra Tachalova
A little bit.
Jason Barnard
A little bit. And I now know what it feels like when somebody asks you a question you’re not ready for. So yes. And in fact really interesting is you got Jono Alderson and Cindy Krum, so she keeps coming back … Jono Alderson says, “Blocks.” Cindy Krum says, “Fraggles.” And it’s the idea that we don’t look at webpages, categories, and so on. We look at blocks. The website contains blocks, which in your example are categories. Within those categories you have pages, which are blocks. Within those pages you have HTML5 Semantic blocks.
Alexandra Tachalova
Right.
Jason Barnard
And within those blocks you have Fraggles, which is Cindy Krum’s super duper favorite thing in the whole wide world, which is the headings, which then break that main content into blocks of headings (H2, H3) plus paragraph. So, whatever level you’re looking at it, you’re looking at it the same way, i.e. blocks. And that’s what Google wants to do, and it’s trying to do. And that’s where we’re going.
Alexandra Tachalova
To simplify everything.
Jason Barnard
It’s indexing in blocks or Fraggles, as Cindy calls them. It’s indexing in blocks, so we need to think in blocks. And all of the things that we’ve done. The DOM, HTML5, the schema markup, is helping Google see these blocks and understand these blocks as individual chucks of content. I say chunks by the way, which is rubbish. I know. So, I said, “Chunks.” And Jono said, “No, blocks is better.”
Alexandra Tachalova
Okay, he’s like a pro.
Jason Barnard
Yeah. Jono’s a good guy. So, yeah. Brilliant
In those blocks, we have tables and lists. How do they help?
Jason Barnard
Yes.
Alexandra Tachalova
So, what about them?
Jason Barnard
Tables are brilliantly, super duper interesting if you’re into that kind of thing. Arnout Hellemans was talking about it. I think it was 95% of all tables on the web are used for design and not for data. So, Google has a phenomenal problem identifying which tables are design and which ones are data. And it had to do that with machine learning. It can identify which ones do contain data, and it actually has an experimental search system, within its table database.
Alexandra Tachalova
Yeah, yeah, yeah, yeah.
Jason Barnard
Arnout Hellemans mentioned on the SEMrush webinar series
Alexandra Tachalova
Yeah, it simplifies user’s life, what I think. Because when you search something, it’s just …
Jason Barnard
The other really interesting … Sorry, with tables is, we’re coming back to these very old HTML elements, tables and lists, which we all kind of threw away with the bath water 15 years ago. Sorry, that’s baby with the bath water. But it’s all old school HTML that’s now coming back, which is incredibly, well not… Interesting … Fun, silly, ridiculous, but cool.
Jason Barnard
And another thing. I can’t remember who told me this. It might’ve been Dawn Anderson. Might not. I can’t remember … Is, when you got a table, it’s really good to describe underneath what’s in the table for two reasons. One of which is some people don’t like reading tables. They prefer reading text. A bit like videos with transcripts. The other, is that it confirms to Google what it’s understood, and it allows you to give co-occurrence, and context around the table that makes that table all the more powerful in Google’s index. Brilliant stuff.
Alexandra Tachalova
So, you just add description after the table.
Jason Barnard
Yeah, you describe your own table. It sounds a bit silly, but actually when you see a page with that happening on it, makes sense to human beings. Anyone who doesn’t want to read the description, because they’ve read the table and vice versa, just skips. Google gets double confirmation.
Alexandra Tachalova
Okay.
Jason Barnard
And it creates a phenomenal co-occurrence and context around your table that makes it more powerful in Google’s algorithms. So, you will tend to come up more. I can’t see it can be any other way.
Alexandra Tachalova
Brilliant. I’ll do that.
Jason Barnard
Brilliant. Brilliant.
Alexandra Tachalova
Brilliant advice. I’ll do that definitely, because we have a few tables, and we also use lists.
Jason Barnard
Yeah, and the other thing is, get rid of all the tables in your website that are doing design. Replace them with divs, because divs do the job perfectly well. Then you’re sending a very clear signal, in my opinion, to Google.”I use table properly, and every table on my website is for data” and Google will take that as a very positive sign.
Jason Barnard
Help the beast feed itself. Help the beast digest and the beast will help you. Sorry, I’m calling Google a beast. I’ve just said it was very nice, sorry.
Alexandra Tachalova
He’s a lovely pet.
Jason Barnard
Oh, yeah. Google. The Google pet. Can we call it a pet?
Alexandra Tachalova
Yeah.
Jason Barnard
Not the beast.
How about using Semantic triples in free text?
Jason Barnard
Yeah, Semantic triples. Love them.
Alexandra Tachalova
Because I have no idea what is Semantic triples.
Jason Barnard
Okay. Entity, relationship, entity.
Alexandra Tachalova
Ah.
Jason Barnard
It’s a sentence. It’s English grammar.
Alexandra Tachalova
Ah, Semantic triple.
Jason Barnard
Yeah, it’s really simple. I was confused about Semantic triples for ages, and then …
Alexandra Tachalova
But it sounds so good.
Jason Barnard
Well, Bill Slawski explained it to me when I said something really stupid about it. He explained it very kindly in this, “Yeah, don’t worry, Jason. You haven’t understoodd. But here’s what it is man.” Which is great. I love it. And the idea of a Semantic triple is very powerful, because if you think about knowledge graphs, you’ve got entities related to each other with relationships, with attributes, and you can explain those to Google (and in fact, to your users), using the idea of entity relationship entity.
Alexandra Tachalova
Okay.
Jason Barnard
And so, if you look at your text, you can think of it as a knowledge graph, which is once again what WordLift does.
Alexandra Tachalova
Knowledge graph is cool.
Jason Barnard
I love knowledge graphs. It’s my top favorite subject. Don’t know why I’m talking about content and “feeding the beast”. I should be talking about-
Jason Barnard
Oh no, because I did knowledge graphs on Digital Olympus in October. That’s why. And the other thing is people write sentences. I’m going to try and make one up, but it’s actually not that easy. Hang on. What was it? “Clarks are the nicest company in the whole wide world, and they sell the most incredibly beautiful, all sorts and types of shoes”.
Jason Barnard
And you have the entity, which is Clarks. The entity, which is shoes. And the relationship, which is sells … Separated by lots of words. For a human being that’s okay. We can probably remember what it is that was said at the beginning of the sentence, when we get to the end. But for a machine that’s incredibly difficult to do.
Alexandra Tachalova
Yeah.
Jason Barnard
So, what you should really say, “The most wonderful company in the entire universe, which is Clarks sells shoes for men, women and children that are incredibly beautiful.” Or split it up in several sentences. So, you’re getting those-
Alexandra Tachalova
Yeah, simplifying it.
Jason Barnard
Yeah, and you get it close together, and you can put all the fluff and the poetry at the beginning and the end. Don’t put it in between the entity, the relationship, and the other entity.
Alexandra Tachalova
Okay, so it’s just like breaking the connections.
Jason Barnard
Yeah, and a lot of people will say, “Write short sentences.” You don’t need to write short sentences. It does help, but you can write these long sentences with all these wonderful words in them and make it very poetic and very charming to the user, and still keep it very understandable and digestible for the beast… oh no! The pet. Sorry, shit.
Alexandra Tachalova
We lost a little bit in our just internal terminology!
Jason Barnard
And then the other one is co-occurrence and context clouds, which are Dawn Anderson’s and Bill Slawski’s favorite things. And the idea that you use words that tend to occur together. She uses the example of seashore, beach, shell, sea, ship, cliff, beach towel, beach ball, sand castle. Whatever-
Alexandra Tachalova
So, it’ all related to-
Jason Barnard
Co-occurrence, and it’s using those words around it to create what Bill would call a context cloud. Which is saying, this context cloud, this particular topic has this cloud of context around it, which is great by the words of co-occurrence.
Alexandra Tachalova
Creating a picture.
Jason Barnard
Brilliant!
Alexandra Tachalova
You’re just like, you see a picture …
Jason Barnard
Yeah.
Alexandra Tachalova
It’s like sending you some emotions. You just can’t feel it. It’s …
Jason Barnard
Yeah. And I can end this little bit on that point, is once again, with tables coming back, with lists coming back, it’s all this old stuff… I would say link building is now PR. That’s really nice. You probably disagree.
Alexandra Tachalova
No, no. I agree.
Jason Barnard
Oh, good. That’s really good. Everyone agrees. How lovely. And then, real writing is coming back. We’re not key word stuffing. We’re not having these awful sentences that don’t read properly.
Alexandra Tachalova
No, no. Don’t do that. Your flow logic should be really easy to follow.
Jason Barnard
Yeah. So, we get real writing back. Real copywriting.
Alexandra Tachalova
Yeah. It’s very important.
Jason Barnard
Brilliant. I’m so, so, so happy about that.
Alexandra Tachalova
Yeah. So, let’s sum up things. So, we need to use structured in data, and all those beautiful terms.
Jason Barnard
Yep.
Alexandra Tachalova
In order to make Google understand better what we do.
Jason Barnard
Yep.
Alexandra Tachalova
But Jason will sum up in better using all his terminologies.
Jason Barnard
I thought you were doing really well, there. That was it – We need to use structured data. We need structure data, HTML5 Semantic triples, tables, lists, schema markup. And Bob’s your uncle.
Alexandra Tachalova
Yep.
Jason Barnard
Bingo.
Alexandra Tachalova
That’s it.
Jason Barnard
SEO is AEO. Thank you Jason. Isn’t it weird that I’m singing to myself? Thank you, Alex, for interviewing me. That was really good fun.
Alexandra Tachalova
Thank you.
