English Español Deutsch Français Italiano Português (Brasil) Русский 中文 日本語

Schema Markup Essentials


Judith's slides >>>

Andrea's slides >>>



Peter Mead: It's Peter Mead, and hello and welcome to the Australian Search Marketing Academy. Today we have a really fun and exciting topic, and an important topic, schema markup essentials with Judith Lewis, Andrea Volpini, and co-hosted by Nik Ranger. 

I would love to introduce each person separately. I'm going to start with Nik Ranger. Nik is an SEO specialist at StudioHawk, overseeing digital strategy, data analysis, content strategy for large enterprises and local businesses. Nik focuses on data-driven results. And I've worked with Nik myself, and I've seen she does a mean technical SEO audit. And so welcome, Nik Ranger, how are you today? 

Judith Lewis is a renowned SEO and digital marketing consultant with over 20 years of experience, predating Google. Yes, I was there too. And working with some of the worlds largest brands. She's a regular speaker around the world on SEO and has been recognized by her peers as one of the most influential people in the UK digital industry. 

Andrea Volpini is a highly successful internet entrepreneur and the CEO of WordLift, and InSideOut10. Andrea also has 20 plus years of world-class experience in online strategies, digital media, and SEO. He's also been a co-founder of another company, RedLink. 

Without any further ado, Judith, are you ready to start showing us what you have for us with the slides? 

Being Deliberate and Properly Planning Your Schema Implementation

Judith Lewis: Absolutely. This is me. If you want to look me up on the search engine of your choice, please do: I do stuff online. It's been a long time for me, but schema is something that's quite new. What isn't new about schema is the attempt that we all have of creating relationships between things. Understanding why something exists on a webpage, what it means. 

Search engines, back when I started, were crap at this. Now they're amazing at it. It's not just about understanding things, but it's understanding the relationships no matter how weird those relationships might be. Because sometimes you're like, "I don't understand why this is related to the other thing, but okay, Google. Whatever." 

But what we really want to be doing is planning, and scheming with our schema. I have been waiting too long to say that. We want to scheme with our schemas. What we want to do is we want to plan things out. And I'm a huge proponent of planning. Because if we fail to plan, we plan to fail as my dad would say. 

Some of us want to mark up all the things. And you go in, and you're like, "Oh my God, schema, this is amazing. It has all the markup I could possibly ever want." Don't do that. Just don't. Just telling you now, don't markup all the things. Markup the most important things, but not all the things. 

And there are some really, really smart things that you can do. There are really, really useful things that you could mark up because it's not just about what is possible, but what is practical and reasonable. Everybody is in love with FAQ schema right now. There's a number of different things that you can put on your page. You don't have to be restricted to just FAQ schema and breadcrumbs.

Schema.org itself gives you a hugely easy way to surf through. Now, I've put in here a search for a product, because E-commerce is one of the places that schema is often employed. It's not just on blogs for recipes, it's also products, and also a lot of other things.

On FAQs, especially for product pages, it's a great way to ensure that when somebody goes to Google, or to Alexa, that you can hopefully get your answer spoken out. But just like we see here, 60% of the time works every time, because it doesn't work every time. This is why you have to be really strategic and tactical with your schema implementation. 

Google’s Schema Autonomy: Losing Rich Results and Changing Features

It's not just about implementing it and thinking somehow magically it will work. First of all, it has to be implemented correctly. But second of all, it doesn't always stay. Here in the UK, this is an example of somebody that I'm monitoring. In the UK, the UK has lost all sorts of rich snippets on products in the run-up to Christmas.

Just because you mark stuff up, doesn't mean it will remain. But also, just because it's there today, does not mean it'll be there tomorrow. I don't know why Google decided to punish all the retailers out there. But hey, merry Christmas, guys. 

But that's not all that Google did to us, man. Because they like to make things just a little bit more interesting for retailers. What did they do? They introduced these fantastic, not just shopping results with your lovely stars, and your prices, and everything, they gave us a whole side panel where we can price comparison shop. 

Also, you can't click through to any of these reviews. Let's say I wanted to go to Overstock, or Zulily, or any of these, I can't because those are not clickable. I have to actually copy and paste.

Deciding What You Should Mark-Up with Structured Data

What can we do then? Because you think about it, you've marked up all this stuff, and then Google went and came along and went, "Hey, we're just going to take that away from you and make it more difficult." 

You can actually look at what's relevant for you to markup. Go in and look at the competition, what's going on. And oh, wait, that price is a lot higher than other people. And hmm, is two and a half stars out of five really helpful? I mean, do you really want to be showing that in search results? Are you really confident enough in your brand, boardgamegeek.com, that you really want to be showing those two stars down there? 

At a certain point, I believe it's 3.4, they stop pulling it through, because they're just like, "You know what, this is embarrassing." We do have to make choices about what we implement, because if we don't make better choices, then it's just useless. 

FAQ schema, right. Everybody is hot on FAQ schema. Everybody is in love with FAQ schema, I see gambling using it a lot. I did this search from google.co.uk, in the UK, on a UK IP. And what do I care about whether US players can play real money on EU poker sites? Or if online poker is legal in the US? I'm in the UK. Come on, people. Think about this. 

This is the exact reason why we have to be careful about even our FAQ schema. If we aren't perceived as relevant, we're not going to get the click. And gambling, online gambling, is hugely competitive. And you screw that up, and you lose the click, that's well ranking, but we've just screwed up and lost the click. 

It means that your click through rate on a number one or number two position goes down because you're not relevant. It's a US site; not relevant for the UK. 

We can do better than this. We don't have to just stick with this bad implementation, because dev time costs money, and we really want to be adding value with schema, we don't want to just do it because we can, because let me tell you, we can do it. If you want us to implement schema as devs, we can do it, there's plugins, there's all sorts. But should we? Because dev time costs money. 

Showing the Practical Value of Schema

If you have to convince a boss, or a CEO, or a client about whether you should implement schema, because dev time costs money, then actually, you can show the practical value. 

And the steps that you can take to implement are super easy. I swear. The first thing to do is don't panic, because you're like, "Ooh, dev time. Oh dear." Well, we've got to do something. Don't panic about implementing stuff, just really focus on what's possible. 

There's lots of ways to generate your structured data, to check it with Google. You can generate anything you want. But there is, of course, there are no guarantees, right.  Just because Google says that you can do a thing, does not mean that they're guaranteeing that it will show up in the search results. 

Like Ron Jeremy says, "60% of the time, it works every time." Don't think that just because you're marking up something, that guarantees it'll be pulled through. It won't necessarily be pulled through. We do use markup to get things like featured snippets, because we're helping Google what it is on the page that they should be looking at.

We'll use jump to links, we'll use breadcrumb schema, we'll use all sorts of things on a website to help guide Google, and help it understand the relationship, and guide Bing, and guide Yandex, and all the others that have their own spiders. But there are never any guarantees that it'll be pulled through. 

Thank you very much. That's me. If you want to get me online afterwards, if you're too shy to ask now, you can get me afterwards. 

Panel Discussion on Schema

Peter Mead: Thank you so much. That was a load of information. Andrea, do you have anything to comment on, regards to what Judith has been saying so far? 

Andrea Volpini: So FAQs: yes, it's a very, very ...relevant topic I believe, because everyone has been playing, because it's so simple, and immediately appeared on SERPs for a few months. But the reality is that FAQ tells a search engine this is an FAQ page, which kills any other intent that is news related. 

My thing here is that we notice that on news and media specifically, when we add FAQ, then we lose traffic from, let's say, discover, and we lose traffic from news carousels, top stories. FAQ can be very valuable when you have an evergreen type of content that is high ranking, and you believe that you can add some value on the SERP with that content. Otherwise, it's a waste of time. 

We have to be very careful in the way in which we use it. That has to do with the way in which Google interprets the schema. We shouldn't use FAQ, because FAQ, it's meant for FAQ pages. It's not meant for an article. 

And we saw, I think, the product panel. The knowledge graph product panel, this is the biggest thing in SERP. Judith does show it’s one of the biggest hits on SERPs for anyone doing E-commerce. Google is getting all the information in one place, and that turns into a negative impact in terms of clicks for all the retailers. It can be very beneficial for the manufacturers, but it can be very bad for the retailers. 

Peter Mead: Nik, what's your take on this? 

Nik Ranger: Schema, in general, is a really, really powerful way to be able to show web crawlers the authority of the sites. I mean, we've got organizational, we've got local schema, we've got breadcrumbs. 

I've noticed that just simply just adding breadcrumb schema, I've seen web crawlers be able to have much, much better crawl rates through our pages. I've been able to make sure that users can be able to find things, and toggle through pages a lot more intuitively. 

Judith Lewis: I've seen a lot of points come through about recovering schema. It's not just those things that are being lost, it's that nobody gets them. Not just somebody lost them, nobody gets them. 

Nik Ranger: Yeah, what was the time period around that? Because I noticed when, famously, there was a little Google glitch, all of mine ... Yeah, a lot of the featured snippets that a lot of my clients are ranking for just completely disappeared. 

Judith Lewis: That graph is not that old. It goes from, I think, mid-August, until just the other day. It's only about 30 days worth of data. 

So SEMrush has a really great product where you can monitor things. One of the things you can do is monitor your search results for certain keywords. I think that's about 400 keywords. And we were looking at about 400 or 600 keywords for that. And across all those 400 or 600 keywords, all of those very few of them were still showing any sort of enhancements that are related to schema. 

Peter Mead: I mean, that's the power of Google, right? Here you go, giveth with one hand, take away with the other. This is part of our bread and butter, this is what we do. We want to be found in Google, so we play Google's game. And this is part of life. 

I want to keep moving because I want to get onto Andrea's and our presentation. There's a question here which may be a good lead-in, because one of our viewers, Rudo Makuyana, has asked, "Which plugins would you recommend?" Well, of course, there so many plugins. I mean, just jump online and start Googling for plugins for a schema. 

But of course, we have with us Andrea, who's the CEO of WordLift, which is a fantastic plugin. I would say, Rudo, definitely give that one a go. But anyway, that was a nice little plug for your tool. 

But there's so much you can do. Share your slides, let’s find out what people need to know. 

Different Ways to Show The ROI of Schema: Examples

Andrea Volpini: I want to talk a little bit about why we do this, why we decided to invest in metadata in the first place. There's this very beautiful concept by a philosopher that works at Oxford University, that described the semantic capital, which is what we do to provide meanings to the things in the world that we care about; to ‘semanticise’. 

That's exactly the point here. That relates to the content that you published, and the data that you published, because we are in the era of artificial intelligence. And data, it's really an important asset in your strategy. 

When you do approach schema markup, and there are plenty of solutions, so I'm not going to recommend you just to go to WordLift, but there are so many different ways in which you can embrace these vocabularies. You should focus on creating the semantic capital, creating the value in the data that you published. 

This value can be looked at in so many different ways. In this presentation, I just wanted to look a little bit at the different layers of the return of the investment. How can you measure the return of the investment on structured data, because of course as we heard from Judith's presentation, you don't have to markup everything. 

Where should you start? And so as we have been working on this since, well, basically 2011, 2012 we came up with the first prototype of WordLift, and so we started to use the schema vocabulary as the vocabulary was really being created. The ROI at that time was very hard to measure. 

Here we have six layers, six ways that you can experiment with in order to understand what is the value. I mean, when we talk about a local business, we might focus on one aspect only. When we talk about a large publisher, we might look at different aspects. 

Rich results, yeah, that's what we've been talking about until now. Now we have enough data in the search console that you then filter the clicks, and understand what has happened since I had a structured data markup. And so here we see a travel brand in Canada, and the improvement it was measured in terms of rankings, in terms of click-through rate increase, in comparison to the average clicks of the site. 

Looking at rich results is an easy way to measure your return of investment, because, at the end of the day, you want to understand the business value of this implementation. It's a client that is booking a villa, it's a service subscription. So what is the business value you're working for? It's a new client that is ringing at the door: it's a booking. 

And then what is the conversion rate that you have on the site? Because that allows you to scope, and say, "Hey, okay, that's the amount of money that we can generate if we bring more qualified traffic to it." Then, of course, you have to analyze what would it take? Should I get a framework like, I don't know, WordLift, or others,; a solution, in order to get a more comprehensive solution? It really depends. It really depends on the type of target.

But then you can do the math and get the ROI. And so you will understand that you can focus on specific areas of your site where there is an actual business value. And then, well, that's the basic formula. You take the total number of clicks, and then you multiply by the conversion rates, and then you multiply by the business value, and then you can subtract the cost. 

Whether it's the SEO that you engage with for implementing the schema, or it's the tool that you're buying, or whatever cost you have. But that's the calculation that you can use for understanding what's the business value. 

Here we see, for instance, the impact on news articles. We are, in this case, in the context of a news publisher. And this is in Europe. As soon as we plugged in the proper markup for a news article, which is not just about adding the markup of news articles, but it's a lot of other small parameters that WordLift uses...we could immediately see an uptake in terms of click-through. 

How long does it take? If the site is large, like in this case, it takes a couple of weeks and you start seeing the results. It can take maybe four weeks. Sometimes it can take more, depending on the traffic volume and the search volume that you're dealing with. 

This is a very small photo blog. This is a guy that I met at a conference that started to use WordLift, and then he showed me this graph. And he said, "You know, when I applied the solution, I started to have a lot more keywords, and ranking for them." 

And that's another impact, because of course, we mark up a lot of content with our tool, and then this makes the content more visible, and so we generally see an uptake on the impressions, which doesn't mean that we're bringing directly more clicks, but we get more visibility. There are more keywords that we can work on. 

There is a lot more than actually schema enables. This is the way, the mechanism, that Google uses for providing results that take into account multiple keywords for that same intent. Structured data is mentioned as a way for Google to expand the query, to understand the query, and to trigger an intended query that better answers the user request. 

Structured data, it's not only a way for getting the stars, or the clicks, or a better click-through rate. It's a way for creating data that AI like Google can use for talking about, or interacting about your brand or your person. 

And then of course, here, we start to see that this knowledge panel that are composed no longer by Wikipedia, but from the open web. And so we become a source of information for Google, and for other semantic search engines with our own websites.

Having said so, is schema a ranking factor? No, it's not. It's not a ranking factor. But it does improve the way in which the information is presented to the user. That can stabilize your rankings. 

Voice search; this is another area, a complete different area of ROI. I mean, there are a lot of companies that want to make the first move into voice search. Well, structured data is a simple and easy way for you to get into the Google Assistant. Whether you're publishing news articles, or recipes, or how to's, or FAQ as we discussed in the previous presentation, you can immediately get into the voice search results of the Google Assistant. 

Speakable structured data, this is very limited. US-only. Content has to be written in English. But it works because it provides a way for a search engine like Google to understand this content is what I expect to be speakable from this specific article. 

Content structuring. This is more for the mid-size businesses, and people that have a lot of content. We can use schema as a base for a content model. And that means that we can use the same language that Google uses for understanding the content we write about to organize and structure the content, and therefore, improve the reuse of this content. 

Here you see, for instance, this is a Flyjet website that has been working for many years now with WordLift. And here we have, at the center, the event for South by Southwest. And then you see the connection with the airports, and the connection with the things to do in Austin, and then the reason why you should fly with us. 

This creates a topical cluster around a specific intent, which is, "I want to get to South by Southwest with my private charter." It can be wrapped up by creating content using schema as a content model. 

This, of course, creates a lot more pages, and so this is an example of a design blog that has been adding a new index to all the entity pages that have been created by WordLift. And then all of a sudden, updated the plugin and removed the new index, and so it got a lot of traffic, because of course, there is a lot of content when we start organizing and structuring stuff around schema. 

This is another example, we can plug schema markup into your Google Analytics so you can go out and say, "Trends by semantic SEO." That means that we start looking at content not in terms of pages, but in terms of entities. That has a tremendous advantage, for instance, if you are selling advertising, because you can start saying, "Hey, I have this amount of traffic for semantic SEO, and this amount of traffic for artificial intelligence." Much like a social network would do. 

With this same data that you are creating, we are starting to use also widgets in content recommendations. On the site, we can recommend better content, because schema is also used by our plugin for creating content recommendation. 

This is just to tell you that schema is not just for Google. Schema is a way for creating this semantic capital that you can reuse for improving your business. And that's pretty much all I have to say. 

The Value of Expertise for an Effective Schema Strategy

Peter Mead: Thank you so much. What have you got, Nik? 

Nik Ranger: Yeah, I'm just blown away a little bit about basically just showing the data, the change over time, and the CTR, and the difference, and what that could actually mean.

What you were saying, it's almost like a philosopher trying to talk to an everyday layman about how it works, and what it means to mark it up as an entity, and how those entities are categorized, and how they will relate to each other rather than just, "But the content is on the page, so therefore, Google should understand it." But it goes a little bit further than that. 

In terms of entity categorization, what a lot of WordLift does is associating with the who, the what, the where, and the who, to be able to create organized databases of those entities obviously to try and help the user experience. How easy is that being able to be able to see that as an SEO to help inform what to do, and maybe what even to do next with content? 

Andrea Volpini: Yeah, I think that as an SEO, I mean, if I would work with Judith, I would immediately ask Judith support for building the knowledge graph, because you need the expertise of a content domain expert in order to understand, "Okay, this is going to be turning to an entity." 

Whether it's a product, whether it's a local business. You need the content expertise to make the decisions that are required for saying, "Hey, this is an entity of this class." That's the beginning. 

Judith Lewis: Yeah, absolutely. You need the expertise of the subject in order to be able to establish what the relationships are. And I have already seen one person asking a question saying, "Oh, I've tried to implement this, but this doesn't relate to this product. What do I do?"

It can be very confusing for people when they're first implementing schema as well. Take it slow and easy. Don't start with FAQ maybe. That's possibly a little bit more complex. Maybe just tackle something like breadcrumbs. Get your feet wet. 

Andrea Volpini: Yeah, breadcrumbs is a good example, because you get an immediate impact on the SERP, but you have to think content, you have to think user, you have to think about the navigation of your site. Again, I mean, let's not look at schema as a way to get on the SERP, because that's what we are here for, but the reality is that schema is a way for organizing your website, and breadcrumbs is a great example for that.

Peter Mead: There's a question we got submitted earlier on from a person on LinkedIn, and I believe you've already started answering this question, Andrea. It's from Varun Mittal. And he asks, "Google Search Console is showing errors. Either offers, review, or aggregate ratings should be specified." So it's product schema, okay. And we see this quite a lot with these kinds of errors coming through. But how are we going to handle these kinds of errors? What's your take?

Andrea Volpini: The real answer is that you don't need to fake data. Google relies on the guides. The guides are way more important than the way that the tools are reacting. You want to make sure that you read Google guides before testing errors on the rich results testing tool, or the good old structured data testing tool. 

You don't want to fake data, you don't want to create fake data just to remove the errors, because it's not going to work. 

Judith Lewis: If you don't have an aggregate rating, if you don't have legitimate, genuine reviews, do not mark them up, don't fake it in order to pass a test. It's a lie, it's a criminal offense in the UK, I think it also is in the EU. Don't fake that stuff up just to pass a Google test. 

Quickfire Tips from Our Panelists for Using Schema

Peter Mead: Well, we've just got five minutes left. Perhaps what I might do is go around and just ask for an impromptu, off the cuff one thing that you would do. Nik Ranger, one thing that you would do with schema? Just a one minute grab. 

Nik Ranger: Well, I think I've already said it. breadcrumbs is incredibly important. I think just from being able to implement that, again, it really, really helps with crawl, crawlers going through your website, being able to pass and understand it. 

And it's a really, really good way of being able to show authority around your site. Plus, it also really, really helps users to be able to go back, click around, and be able to toggle through pages, product pages, and the like.

Peter Mead: Great. Judith, do you have the one thing that our viewers can go away and do with schema?

Judith Lewis: For me, it's planning. As Nik said, breadcrumbs are so important, because they help you understand your website. If your website isn't actually organized in a way that you can markup for schema your breadcrumbs, then you're in trouble. 

Plan everything, sit down, and don't be overwhelmed by schema.org. The world is your oyster. But sit down and look at your website, and look at the competition, and see what they've got, and what is possible for you first, and plan on only those elements first. If no one in the SERPs have FAQs, don't do that first: do ratings, do reviews, do something else, calories in your recipe if you're a big recipe blogger. Look at what everybody else is doing first, and then go the next step. 

Peter Mead: Andrea, what would be your one thing that you would say to people to do with schema right now?

Andrea Volpini: Think of schema as your sitemap on steroids. It's a new way for creating a sitemap for a search engine and its machines. But also, don't stop experimenting. Be your own Google, be your own Google. Think about using this data yourself in the first place. 

Peter Mead: What a great webinar. I think this is really great information for everyone. And without further ado, I will hand over to the next bit of information on the web that's not us. Thank you all and goodbye.

All levels