How structured data is paving the way for AI-generated search
Last week I attended the Google Search Central Live Zürich 2023, or #SCLZurich for short. The conference, held over a whole afternoon at the Google HQ in—you guessed it—Zürich, Switzerland 🇨🇭, marked its grand comeback after the pandemic years. I had a really great day meeting with fellow SEOs, chatting with friends from the Women in Tech SEO group, and learning a lot from the Google Search staff. The talks were very interesting and got me thinking about many topics and ways to improve my skills.
If you want a breakdown of all that was said, head over to this recap post of #SCLZurich by Olesia Korobka or this 7 talks - 7 take-aways post by Corina Burri.
One topic that kept coming back throughout the afternoon was structured data. Although it’s, of course, a very important technical topic, I was surprised to see it mentioned, if not extensively developed in almost all talks that day. So this got me thinking: why are we hearing so much about this today? How does structured data inform the future of SEO, and how does it shape the impending era of AI-generated search?
How structured data helps improve search results
Microdata, structured data, schema markup, and semantic markup* are not new to SEO at all. We’ve been using these snippets of code in our pages for ages, especially when it comes to e-commerce. Structured data is what informs Google Shopping of your products’ features for example, as Matthias Weismann, Software Engineer at Google Shopping, explained it detail at the conference..
By the way, the Product schema markup is undergoing a hefty upgrade, with new variants that will better help the user find the product they are searching for. Think certificates, labels, sustainability or dietary specificities. These are exciting times!
By meticulously organising information about our products and services using structured data, we're essentially handing search engines a roadmap to our content. By linking or nesting items to one another, we’re saying, "Here's what we offer, and here's how everything is interlinked.”
This carefully crafted organisation doesn't just enhance the appearance of our search listings with rich snippets but significantly betters the user experience by delivering more precise search results.
Structured Data, AI and search
So what does structured data have to do with AI? Well, when Google crawls a website, it gathers all the information a page has to offer, including structured data. Then the indexing algorithms come into play and try to make sense of the page. At this stage, Google uses a bunch of machine learning, AI and NLP tools to figure out what the content is about and if it’s worthwhile for users.
Imagine a scenario where two pages have the same quality of content and relevant information. On the first page, Google has to sift through the content, analyse headings and links to understand what the page is about. In the second page, Google has to do the exact same work, but with added information, structured in the same way across all webpages for the same type of item.
On the first page, algorithms have to put in more work to categorise the content in their database, whereas on the second page, they “only” have to check if the information provided in the structured data bits of code match and correlates to the content on the page.
Of course, I'm oversimplifying here, but you get the idea. Structured data makes Google’s work easier. It takes way less computing power for Google to understand content with a roadmap than having to figure out a path on itself.
“While we have the technology to find that structure in web page text automatically, those systems are not perfect. (…) When you tell us what's on your web page in a structured way, we can more accurately interpret the contents.”
— Ryan Levering, Software Engineer in the Structured Data Team at Google (source)
If we think of it with a very down-to-earth capitalistic approach, it’s just more cost efficient for search engines. The same applies to having a technically sound and fast loading website, it requires less energy to load and crawl, which means it costs less to index. (Yes, it’s also more environmentally friendly, of course, but let’s not be blind to the societal system we operate in.)
How taking full advantage of structured data can benefit both users and website owners
The ripple effect of fully harnessing structured data can be huge. Imagine having your website not only fully present in the results of a query, but also taking space on the result page by offering rich results to users. It’s not only about the ranking position, it’s about the experience you offer.
For example, as Maria White, Global SEO Lead at Kurt Geiger, explained in her presentation at SCL Zürich, the experience you offer to your user does not start on your website’s homepage, it starts on the SERP.
Going beyond the typical variants offered in a Schema item and actually going through the documentation to precisely optimise every relevant bit of information will help your website in the long run. Right now, there are open conversations happening on the Schema.org forums about how items should evolve through time.
For example, food product items could get an allergens variant in the near future, helping users take care of their health when researching and buying food online. And guess who is actively participating in this conversation? Yep, Google is. They are even at the origin of some of these evolutions. I don’t know about you, but to me, having an announcement at the next Google I/O that search results for purchasing food will now display allergens seems like a totally on brand thing for Google to do.
For search engines, it’s a clearer, more efficient route to quality results. For us, it’s about becoming the “cost-efficient” choice that demands less power to crawl and index, potentially earning a favourable nod from search algorithms in the long run. It's not merely about scaling the search product but elevating the quality of the user experience while reducing the digital carbon footprint of our online endeavours.
This proactive approach, as I gleaned from the conference, could be a game-changer in how we approach SEO, ensuring a win-win scenario for both search engines and website owners. Doing your research—dare I say, hiring an expert—on structured data, and not just “filling in the fields my CMS put up for me” could be one element that will tip the scale in your favour in the near future. Actively optimising your website with a collaboration mindset with search engines rather than a “quick win” approach is sure to be beneficial in the long run. If not for your rankings right now, then for your users, who crave information before making an informed purchase on your website.
Conclusion
The narrative around structured data is far from over. It's an unfolding chapter in the SEO playbook that holds promise for a more efficient, user-centric, and AI-compatible future. I invite you to delve deeper, explore the schema.org documentation, and start weaving structured data more intricately into your SEO strategy. It's about laying a solid foundation today for the AI-driven SEO landscape of tomorrow.
*Structured Data is a term used to describe data that is organised in a specific manner, making it easier for search engines to understand the content on web pages.
Schema Markup is a semantic vocabulary or a set of code tags you can add to HTML to improve search engines' understanding of your pages, essentially a form of structured data.
Microdata is a specific syntax used for embedding structured data in HTML documents, and it's one of the formats you can use to implement schema markup on your website.
Semantic Markup is a broader practice that refers to the use of HTML tags and other markings to denote not just the structure, but the meaning of the content, which includes practices like schema markup and others.
