An interesting discussion emerged on Twitter inspired by an article written by Jono Alderson. The article proposes thinking about Schema.org structured data markup as a way for emerging AI technologies to better understand and surface published Internet content.
Schema.org Structured Data Markup
The content on a website is called unstructured data because there is no formal organized structure to it that labels each part of the content in a machine readable way. Structured data on the other hand is the exact same content but organized with labels that identify images, authors, and content so that a machine can immediately understand it.
Schema.org structured data markup is generally seen by publishers and the SEO community as something to use in order to make a web page eligible for rich results features in Google. That way of thinking is manifested in the many SEO and Schema.org WordPress plugins that are limited to outputting structured data that Google may use for surfacing rich results.
New AI technologies that can use structured data are here, requiring search marketers to consider a new approach to how structured data is deployed. What Jono encouraged in the article is to think of structured data as a way to create a “data-first foundation” that is ready for the near future.
The article proposes thinking of Schema.org markup as a way to communicate what a web page is about and how it relates to everything else on the website. Jono writes:
“But don’t shy away from building a connected graph of broader, “descriptive”” schema just because Google’s not showing an immediate return. These “descriptive” types and relationships might end up being the lifeline between your content and the AI models of the future.”
Jono tweeted about his article on X (formerly Twitter) and Martha van Berkel, founder of SchemaApp, agreed with Jono’s article that the role of Schema structured data markup is shifting.
She tweeted:
“I agree with you that the role of schema markup is changing. Building a knowledge graph to manage how your website/content is understood with schema, and then asking it questions will be more important than optimizing for Rich Results or for Google.”
Ammon Johns tweeted:
“The biggest issue with Schema is that it is largely just self-declaration, no different in essence to META content, and we know how reliable Google decided that stuff was. So Google will use it, but they are unlikely to fully trust it.”
Ammon is right of course that structured data can’t be blindly trusted. One way to solve that problem is to use a smaller index of high quality websites the wa Perplexity AI does.
Gagan Ghotra tweeted how they sometimes would like to expand their use of structured data but are limited by what the SEO and structured data tools offer.
Read Jono Alderson’s X discussion here.
Read Jono’s article:
What if Schema.org is just… Labels?
Featured Image by Shutterstock/PureSolution
