Podcast: Play in new window | Download
Subscribe: Apple Podcasts | Spotify | Android | TuneIn | RSS
Jarno van DrielJarno van Driel is a true pioneer on the semantic web.
Even before you could add machine-readable semantic markup to webpages, he was discovering ways to help search engines understand what web pages were about. Much of that success grew out of his early focus on accessibility and usability.
When semantic markup was introduced, he was among the first cadre of experts on RDFa, Microdata, schema.org, and other semantic practices, and he is to this day one of the most respected practitioners of this craft.
We talked about:
- how he arrived at his work at the intersection of semantics, accessibility, and SEO
- his early introduction to the importance of accessibility in content work
- his surprise and curiosity about how his small primary school websites were outranking big commercial websites
- how now-common practices like on-page navigation helped his SEO efforts 20 years ago
- how he discovered semantic metadata by reading Drupal documentation
- the dearth of ontology guidance and syntax documentation in the early days of semantic markup
- how semantic markup started to take of with the introduction of Google’s knowledge graph
- the small early communities that formed around semantic search
- how the arrival of Google’s knowlege graph filled gaps the ability to disambiguate entities, especially in multilinqual contexts
- his take on the notion of the Semantic Web
- the evolution of his work over the past 10 year from getting rich search results to actually structuring meaningful websites
- how well-structured, properly marked-up webpages can deliver better results for a company, even when the page gets fewer visits
- the importance of focusing on messaging and content over technical markup if you want to be found on the web
Jarno’s bio
Jarno van Driel is an international Structured Data and technical SEO consultant. He started his career in 1998, during the early years of the web as a print and web designer, Flash and 3D artist, frontend developer, and accessibility engineer. Jarno has continuously embraced new and challenging roles throughout his career to become a well-rounded digital professional.
Jarno’s activities can be categorized and classified in many different taxonomies and ontologies. Yet labeling Jarno’s job title is a near impossible task because his activities overlap with a multitude of departments and specialists. An ambiguous state, because of which most simply know him as ‘just Jarno’, that structured-data fanboy from Rotterdam, the Netherlands.
Jarno became fascinated with the ‘semantic web’ when he discovered something called ‘semantic metadata’ (2008). Semantic annotations piqued Jarno’s interest because of his background in web accessibility, which is about expressing structure and meaning. Structured data markup takes things to a whole other level, making it the obvious path ahead.
As an early adopter of linked open data, his work was mentioned in several W3C discussion groups. Mentions for which Jarno is very grateful because it would lead him to start publicly participating in schema.org (2013). Through this, Jarno met and learned from many of the pioneers who are at the forefront of the semantic technologies being used today.
Connect with Jarno online
Video
Here’s the video version of our conversation:
Podcast intro transcript
This is the Content Strategy Insights podcast, episode number 168. We all want our content to be found on the web. It’s always been the case that that’s easier said than done. Jarno van Driel discovered early on that focusing on accessibility and usability would give his clients better visibility in search results. When semantic markup was introduced so you could add metadata to HTML pages, he found that he could do even more to help search engines understand web content – and to help his clients get better business results.
Interview transcript
Larry:
Hi, everyone. Welcome to episode number 168 of the Content Strategy Insights podcast. I’m really happy today to welcome to the show Jarno van Driel. Jarno is a multi poly accomplished guy. He goes back to the dawn of the interwebs. He’s been doing semantic stuff, accessibility stuff, SEO. He’s been called by our mutual friend Teodora Petkova a Schema Ninja. So welcome, Jarno, is that an accurate description of you?
Jarno:
Well, first off, thanks for having me and yeah, it sounds about right. It’s about as close as I will get to describing myself. So yeah. It’s a good starting point.
Larry:
I think anybody who can’t… It’s hard to Ascribe a label to. You’re always the most interesting people, which is why I wanted to have you on the show. We’ve talked about a lot of stuff, but one of the most interesting things to me is the way you’VE stitched together and we’ll explore this over the next half hour, these kind of worlds of the Semantic Web and linked open data and that whole semantic stuff with SEO and with accessibility and accessibility is really intriguing to me ’cause I don’t have any trouble seeing the link between those, but I don’t think it’s a real common link. So can you talk a little bit about how you ended up with this bundle of really interesting interests?
Jarno:
First of all, it started off when I built my first website in 1998, and that experience made me aware that I wanted to do something with the web. I had no idea what, like the majority of people in those days. So I became a web designer and for about the first five years of that, I think it around with just about anything. I installed servers, I installed email environments, but also played around with 3D software, with Flash, with Shockwave, all kinds of stuff.
Jarno:
And then around five years in things started to become more serious. I took on a job at a typical 2000s IT company and one of the tasks they had been appointed with was maintaining the websites of roughly 500 primary schools in the Netherlands. So for the first time, I had a dedicated front end job. I had one specific task to accomplish which turned out to be anything but very clear because six months in the Dutch government decided, you know what? We subsidized the crap out of this stuff. Everything that’s subsidized for us needs to become accessible. So suddenly I also had to become an accessibility engineer to make sure that all those primary schools started to produce accessible websites.
Jarno:
Interesting phase, implied a lot of reading of W3C documentation. That’s actually my first introduction to W3C documentation. And yeah, interesting. It led to a lot of experimentation and also having to, surprisingly enough, as a developer a lot with content because suddenly nobody was aware of accessibility guidelines or writing for accessibility. So I ended up reading tons of stuff about accessibility guidelines, especially in regards to writing and how you actually write very unambiguous pieces of creative works that are very clear, very well-structured because you have to keep in mind you’re dealing with people with limited cognitive abilities or you’re dealing with people who have other kinds of issues like ADHD, try to keep people focused on a single topic. So you need to learn to write around all those different forms of disabilities.
Larry:
That’s so interesting. I’m just going to say if I’m doing my math right, that’s 2003, like five years after ’98. That’s 20 years ago. You’re doing-
Jarno:
Yeah, exactly.
Larry:
Yeah. A lot of people are just now discovering the… I mean not discovering, I mean, good writers have always known about the importance of clear, accessible, simple writing, but you were actually embodying it in big web, like this giant Dutch school system thing 20 years ago. And I can also see the way you talked about the importance of structured meaningful content in the context of that and that kind of leading to the writing part of it. But it grew out of the accessibility project, right?
Jarno:
Yeah, far more because I think it was around three years in when they started noticing certain websites of primary schools out ranking high commercial websites for certain terms. And that was something I never paid attention to. I was a front-end developer and accessibility engineer, but I did don’t do anything with search. So that got me curious and I started to look into things for a couple of months and realized, “Hey, you know what? A lot of the disabilities people are suffering from Google seems to suffer of those as well.” Cognitive bias, navigational issues, motorical issues, having difficulties navigating through a site.
Jarno:
So when I start to follow that pattern for a few months, I came to the conclusion, “You know what? Accessibility is seriously helping out search results,” which made sense because it helped search engines. It became easier to parse website, it became easier to extract the information in there. If you’re writing according to accessibility guidelines, you also add all kinds of relevant links with good descriptions. So yeah, that also helps search a lot. And after having toyed around with that for a few years, I was like, “Okay, you know what? The accessibility has been fun for five years, let’s dive into the world of SEO for a few years and see what I can come up with.” So it all was sort of a very natural progression over time.
Larry:
No, that story makes complete sense in that transition. I’m really curious about the… So to go from a school district that’s outranking these commercial people who are probably paying $10,000 a month for SEO, first of all, it must’ve just been irritating them to no end, to have this school district-
Jarno:
Probably.
Larry:
Yeah, which is a fine little fun thing about that. But as you’re talking about that, it’s like how were you conveying that to the search engines at the time? Was it purely in the content… ‘Cause I’m trying to remember when RDFa and Linked Open Data and all that stuff, that was about that time that stuff was introduced. Am I remembering that right, or?
Jarno:
I’m still talking about the phase prior to things like RDFa and microdata. So we actually still were using HTML links and strings.
Larry:
That’s what I mean. So it was just the on-page content that was doing that, getting that result.
Jarno:
Nothing more, nothing that. But especially in those times, nowadays, people are much more aware how HTML works and how you should structure things. Back then it was still, for a big part, the wild, wild west, a lot of people were just doing something. So if you’re one of those crazy guys that actually followed guidelines and studied what those had to say and you try to put them in practice, then it was quite easy to, nowadays we would call it manipulation, but back then it was simple innocently doing the right thing without any other intention. And there was still space on the web to accomplish things with that. So 20 years back, it was a lot easier than it is nowadays.
Larry:
Yeah. And Google has always claimed that they’ll reward you for doing the right thing. There’s not always evidence of that, but it sounds like there was back then.
Jarno:
In general, I still think that is their overall intention, whether that always works out. Like any company, you try things, you break things. So one thing is a success and the next is a failure, yet Google is no different. It only happens at a scale a thousand times bigger.
Larry:
Yeah. Now, I wonder if some of your success back then points to semantic issues, like you were talking about the structure of the document. Were you doing properly hierarchically organized pages around H1, H2, H3 headings and-
Jarno:
Headings, identifiers in their on-page navigation. Nowadays, everybody has a WordPress plugin to provide an on-page navigation. I was already doing that in 2003, 2004, only hand coding dose. So you actually made it easy for Google to jump through a document and end up at specific topics. Now, back then, they didn’t utilize those features yet in search. But looking back at it, I now think, yeah, they wouldn’t use it for features, but their algorithms sure we’re digesting it.
Larry:
Yeah. Well actually, and let’s jump forward a little bit now because I’m wondering how the Semantics evolve as the Semantic Web kind of annotations come into play and you’re doing different stuff at that point. Can you talk about that? Was that sort of an easy transition for you going like, “Oh, this is just another way to share,” or how did you… I think you were in the Drupal world or something then right when you discovered them?
Jarno:
Exactly. Yeah. By the time I had made the switch for working for a Dutch publisher of magazines and plain websites without any magazines, and we were busy in a project of creating new frontends for all those websites, roughly somewhere around 15, 20 websites, and we were working on Drupal 6 templates when I was researching how the heck is Drupal 6 supposed to work, because Drupal had quite a steep learning curve. So it involved a lot of reading, and I started to run into documentation mentioning “machine-readable metadata.”
Jarno:
I was like, “Machine-readable metadata for Semantics? Semantics that’s my cup of tea.” ‘Cause I’ve been working with an non-Semantics for five years only in a different format. So that made me curious and that opened up a world of RDFa. So I’ve played around with it since Drupal 6. Then Drupal 7 announced they would integrate RDFa into their core. And the people I were working with back then already prepared new templates for all those websites. And the day Drupal 7 was released, we released new front ends for all those websites, all build up full of RDFa. And if I’m right in the first two or three years of the existence of schema.org, I dare to say I was at least amongst the top three publishers of RDFa triples with schema.org vocabulary.
Larry:
Well, you were a true pioneer then ’cause I think something like 40% of the world’s websites now have schema markup, and you were one of the first 10 or 12 probably I would imagine.
Jarno:
No kidding. I was probably amongst the first 25 websites out there because I discovered schema.org before its official release.
Larry:
Oh, interesting. Yeah. And let’s talk a little bit about the relationship ’cause in order for something like RDFa… Oh, and we should probably explain that RDFa is like a W3C standard format for annotating pages. There’s also microdata… There’s other ways you can do that. There’s other ways you can tell Google and other search engines what’s on the page. Like JSON-LD now is probably more common, but can you walk a little bit through just that technical foundation of how semantic markup is used in webpage building?
Jarno:
When I started out, and it was around 2008, it involved a lot of confusement because the only documentation I could find were academic papers or W3C documentation and neither of which was very clear on how you were supposed to use it or what type of limitations you should apply or which ontologies you had to use. So I played around with it for roughly four years until I became so frustrated. I was almost done with it. I was almost ready to throw in the towel because I really had the impression, this is not going anywhere. It’s way too ambiguous for normal people. How are normal people going to figure out which of those 5,000 ontologies out there they’re supposed to use without any guidelines on how to use them. And oh, by the way, there’s also no information about how to use the syntax.
Jarno:
So yeah, I became highly frustrated in those first four years roughly. And then one day I discovered a publication on a W3C mailing list talking about a website I had worked on, and that was Ian Hickson, one of the guys who came up with schema.org, and he had some things to say about our implementation, which were rightfully criticizing, but it felt hurtful for at that moment in time because I was a noob. So when you see some high-level person talking crap about your stuff, you’re like, “Uh-oh, what did I do?” So I lurked around for those mailing lists for some time, and I think it was two months prior to schema.org being released, I understood that this thing was really going to happen. So we already prepared all the templates for those websites to switch to using schema.org. And again, when Drupal the exact same day, Drupal 7 was released, we published that information and I think that was within a week after schema.org released sort-of-ish.
Larry:
Wow. So before schema.org, the markup format existed, but there wasn’t a-
Jarno:
The syntax existed.
Larry:
The syntax existed, but there wasn’t this kind of Google approved, canonical upper ontology to attach things to. Is that-
Jarno:
Yeah, the only way I knew that it was picked up is because back then there was a search engine called Sindice, and Sindice was an experimental search engine that actually was trying to find structured data out there on the web and offer a new type of search experience. That search engine never made it, and it died pretty soon. But that’s how I got my numbers and how I found other websites that were using it because, yeah, again, Google still wasn’t using it, and I think it took at least a six months to a year before Google actually started showing sites that were using schema markup in their search results if you were looking for them. So I was heavily dependent on Sindice in finding other resources to find out how people were actually doing this stuff.
Larry:
Interesting. It’s so great that there… ‘Cause I remember doing some of the same explorations that you made in that same era, and frankly, giving up because I had no tolerance for impatient know-it-alls and obscure academic references, and I just went on to do other stuff. So thank you for sticking with that. But one of the things that unfolded not too far after that was Google’s introduction of the knowledge graph, and that’s kind of where this stuff goes, right?
Jarno:
Yeah. That’s where it started to take off. Although the impact at first from the knowledge scope was probably bigger amongst the people already involved in the Semantic Web than the general search community out there. Right now we have entities, we still need keywords for that. So in the minds of most people, not much change actually. And around the same era, it was that I got introduced to Aaron Bradley. He was running a Google+ community around schema.org. Back then it had like eight members, and over the years we built that out to the biggest community surrounding structured data that as far as I know has ever been on the web. So that’s something I still miss, and that’s where a place where a lot of people ended up nowadays who are more or less known figures, at least in the SEO space, who do a lot of things with entities. But those are people that go back in my history 10, 15 years already.
Larry:
Yeah, we were hanging out with some of those folks last weekend. But to that, regardless of the emergence of the knowledge graph and stuff, that evolution of the importance of schema, and I’m thinking especially from the perspective of a content person. ‘Cause I think many, if not most of us, I don’t know exactly how many content folks were in that era, just like I felt like I was just kind of grasping around for, you know, you’re under pressure from whoever to get better search position in Google, and you’re just trying to figure that stuff out. You were approaching it sounds like, in a structured way and more patiently maybe. But anyhow… So I’d love to hear a little bit about that. But one of the key things I think, and you just alluded to it a minute ago, was shortly after Google introduced the knowledge graph or about the time they introduced it, they talked about that notion of things, not strings, like, entities, not their labels. Can you talk a little bit about how that whole different understanding of the web emerged?
Jarno:
For me, it actually was the first time, the whole introduction of Google’s knowledge graph was nothing else as a logical next step in my mind. But that’s again, coming back from a accessibility background where when you’re writing content, you’re doing your best to explain entities and relations in a very unambiguous and very transparent manner. So your focus is explaining meaning and relations. That’s very important in writing for accessibility. So when I discovered the possibility of creating semantic annotations, it was like, that makes sense because this removes the limitations I currently have. Especially in multilingual environments, it’s very difficult to be unambiguous about entities in different languages. And normally we used to resolve that through link networks that in the SEO space that has a negative vibe around it, but for disambiguity and information retrieval, connecting the right resources together it is highly useful. So in that regard, it made sense for me.
Jarno:
It filled in gaps I was not able to close with the old ways of accessibility. Going into the whole knowledge graph thing. Yeah, I was an early adopter, so the people I was hanging around with were all of a like-mindedness as I was, and not necessarily interested in it from a commercial point of view or an SEO point of view, but really more from a semantic point of view. And as things evolved over at Google and they started to introduce more of rich results over time, and people started to see, “Hey, I can drive traffic through those special results,” it’s nature started to move away from the idealistic Semantic Web to a much more commercial happening where all kinds of people started to showing up with all kinds of other interests than explaining meaning.
Larry:
That’s right. That’s sort of an ongoing tragedy. I think we’re seeing it now with the arrival of a lot of the AI tooling, but that aside, this notion that something born of really altruistic and nobly motivated things like the original idea of the Semantic Web was that the web will be this one giant database. We’ll all be sharing everything. It’ll be great, and then it pretty quickly gets commercialized. Did you have any success? I mean, I think to this day you still exemplify that old original intent of the Semantic Web, but is it harder to express that commercially these days… I mean, you kind of have to express it commercially to make a living, but can we still have the old Semantic Web? I think a lot of people would like to see that.
Jarno:
I think there’s a difference, depending on your definition of the Semantic Web. If you go to HTML, HTML offers a lot more features than the search engines tend to use for their indexation purposes. Why? Because most people make a mess out there. So if they use every single option out there to improve their search results, there is going to be a huge amount of fall off because people not understanding the technology correctly. So search engines need to make sense of a less-perfect world than we imagine it to be. Looking at it from that point of view. So working on the structured data markup itself has a limited return on the investment, probably the majority of your return of investment, you’ll find in those rich results they generate because of it, because those generate additional clicks.
Jarno:
What I am happy about now is that we’re starting to move away from just that carrot on a stick we’re holding in front of everybody’s nose, and I’m starting come across a lot more companies that are actually interested in creating meaningful data sets, but beyond that also are much more aware about their messaging and its influences on their marketing, on their business. So by now, if I compare my work now to that of 10 years ago, 10 years ago, I was just getting everybody rich results, and by now I’m actually structuring websites, structuring content on webpages, and the structured data markup in the end is just a result of that process, but it’s not an isolated process. And for me, that’s where the true value lies. If you really use the methodologies behind Semantic optimization, and there are a lot of them out there, I don’t know all of them, but there are many different methodologies you can use to come to the same end result.
Jarno:
In the end, that’s where the true value lies. You’re producing highly structured, valuable and meaningful content, which offers different entry points, which offers different methods for interacting with it. And that’s for me, where the value lies. And markup is just one little piece that an entire puzzle, but people often just look at one single piece because it’s hip and it’s the flavor of the moment, but you need to zoom out, you need to look at the overall picture, and then you understand the place where markup has its purpose, but people should not focus on the markup alone. It’s the entire methodology surrounding it.
Larry:
Interesting. So the markup is just the expression of all the deep semantic work that’s gone into structuring the content correctly, establishing the meaning, expressing it that way, and then you just reinforce it with the markup, with the structured data on the front end.
Jarno:
For me, markup has two purposes. First of all, it helps clarify those things that natural language and processing is not able to digest. Sometimes we write in such granular details that it becomes difficult for automated systems to extract the exact detail you were talking about and then that markup can give that final push. No, I’m not talking about this, I’m talking about this specifically, and that can really help. I’ve seen websites in the past that lost half their traffic for certain pages. Yet if you looked at the goals for those webpages, so their conversion targets either quadrupled or more, even though they were attracting a lot less visitors, but they were attracting the correct visitors. They were actually… Google was able to match the right person to the correct content, and that was enforced through that additional markup, but again, not in isolation, also because the content on that page resembled that same quality and that same depth of information.
Larry:
That’s so interesting. And that’s more about the markup, expressing the actual intent of the page. How much of that… You mentioned earlier keywords and you were still using keywords to identify these entities that we’re dealing with, but I just wonder how much of that is of the benefit of that that you just talked about, like, how you can essentially quality over quantity. Again, does that go back to the amount of work that goes into both the… Again, it’s kind of the same question I asked a minute ago now I think about it, you know, the work that goes into the content to give it that quality, and then it’s expressed and Google goes, “Oh yeah, this is exactly what these people need.”
Jarno:
Yeah, it’s very easy to match the actual search intent of certain queries. Especially in the SEO field, a lot of people tend to oversimplify intent. You can’t categorize intent in just five categories and that there are thousands of different variables to the intent while you’re searching for something. So to be able to match that exactly to the searcher’s query and then the end goal you’re sending them to, I still find that an amazing resolve nowadays. The fact that search engines are able to pick up on that type of granularity and translate that in search results, I’m still amazed by it. It looks simple on the outside, but if you really dive a little bit in the technology stacks needed for this, it’s amazing that we actually got it off the ground, is quite surprising if you think of it.
Larry:
Right, it’s crazy. And that sounds like what you’re describing is the mechanism by which enterprises would love to be doing personalization is to have their own little Google intent-discernment machine and all the right content that might match that intent and just magically… In fact, Philippe and I were talking about that at your little gathering last week.
Jarno:
Yeah. Cool.
Larry:
I can’t believe it, Jarno. We’re already coming up close to time. I mean, we could literally talk about this stuff forever. But I always do this, before we wrap up, is there anything last, anything that’s come up in the conversation or just that you want to make sure we share about semantics, accessibility, SEO?
Jarno:
I think one of the things that I touched on just a little bit ago is the interweave-ed-ness, by lack of a better word, in my vocabulary between content and markup. I see a lot of people looking at markup as if it’s a type of fairy dust they can throw on top of a website, and that automatically will generate beautiful results for you. And I wish the world would focus a little bit less on the technology they used and a little bit more about the message they’re putting in their writing or painting or in their video or in their podcast. It’s about the substance. It’s not the frame you put it in. And I see a lot of people over worrying about their framework, and people should worry a bit more about what they’re putting inside that framework a whole lot more actually.
Larry:
Yeah, I heard… Karen McGrane did a great presentation years ago about, it’s not the box, it’s the gift inside the box.
Jarno:
Exactly.
Larry:
Yeah. Yeah, that’s great. Well, thanks so much, Jarno. Hey, one very last thing. If people want to follow you online or get in touch, what’s the best way to connect?
Jarno:
I’m on Twitter, Mastodon and LinkedIn. Sorry, I still call it Twitter, and as of today, I’m also on Bluesky, all under the same handle, that JarnoVanDriel.
Larry:
JarnoVanDriel. Great. Cool. Well, I’ll link to that in the show notes as well. Thanks so much, Jarno. Really good, fun conversation.
Jarno:
Oh, thank you very much for having me. Like you said, we can do this for hours, so…
Larry:
Well, we’ll definitely have you back.
Jarno:
Thank you.
Leave a Reply