The Search Session

The episode contains substantial insights about structured data mythology, the gap between schema.org theory and practical implementation, and realistic assessments of ROI for entity optimization. However, significant portions devolve into conversational meandering, personal anecdotes (summer weather, gaming preferences), and repetitive discussion of basic terminology rather than advancing novel claims. The guest does deliver contrarian takes (FAQ schema was unnecessary, most vocabulary goes unused, schema.org may need resetting) but these are not consistently packed throughout.

“90% of the markup out there only exists because Google actually had a feature for it”

“unless you got a lot of money to burn, stay away from it. Any form of agent optimization.”

Van Driel offers genuinely counterintuitive framings: structured data markup is irrelevant for LLMs (which parse natural language), entity optimization ROI peaked around 2016-17 due to NLP improvements, and most schema.org vocabulary will never be used. However, these insights are presented somewhat discursively and the core thesis (schema.org over-engineered, focus on pragmatic use cases only) while sensible, is not deeply fresh. The framing around knowledge graphs for internal business purposes is solid but not radically new.

“LLMs don't need that format. They're not created to work with those formats. LLMs are created to deal with natural language”

“schema.org was designed for a different era...I wouldn't be surprised if we see a reset the coming years”

Van Driel is a legitimate practitioner with deep pedigree: working in structured data since 1998, active volunteer contributor to schema.org governance, hands-on experience with enterprise knowledge graph implementations, and documented case studies (Panda recovery 2014, entity optimization evolution tracking). He speaks from 25+ years of real implementation work rather than theory. The only deduction is that he's a consultant/expert commentator rather than a current operator running a live business at scale.

“he started his career...in this field in the last century, because it was 1998”

“I'm one of the volunteers out of dozens...Everything in there ended up in there with a certain use case in mind”

The episode lacks concrete data and named examples. Van Driel references his own case studies vaguely (Panda recovery, entity optimization ROI decline around 2016-17) but provides no metrics, timelines, client names, or quantified results. He mentions general concepts (NLP reached 85% accuracy by 2017-18, knowledge graph error reduction to 80% with agents) without sources or specifics. Most claims remain abstract: 'Google has been aligning Merchant center specs for eight years,' 'product schema as biggest use case,' but without hard evidence or numbers.

“I started the discussion around product variants around 2017 and it took only like six or seven years to roll out into production”

“NLP had become so good that it already crossed the 85% benchmark in accuracy”

The host Gianluca asks reasonable setup questions and allows Van Driel space to develop thoughts, but the conversation lacks aggressive follow-up, pushback, or pressure-testing. The host often agrees or pivots to related tangents rather than drilling deeper. When Van Driel makes strong claims (FAQ markup was unnecessary, most schema never used), the host doesn't push back with counterexamples. The discussion meanders into weather, gaming, and extended anecdotes without redirection. Some questions are soft ('How is SEO treating you lately?'). There is one good moment where the host challenges Google's documentation clarity, to which Van Driel provides a nuanced rebuttal.

“Yeah. Yes. In fact, it's interesting how the this is something that I like to do every year”

“Yeah, I understand. Well, maybe the future. Let's see. One of the evolution of a technology”