Search Engine Land (2025) — Schema & AI Overviews visibility test

Summary:

This article shared a simple test of whether structured data, also called schema, might help a page show up in Google AI Overviews. The authors wanted to move past theory and run a direct comparison.

It was published by Search Engine Land, a search marketing media outlet owned by Semrush.

That means it sits close to the SEO industry and its business interests, but this piece reports on an experiment by two contributors rather than a formal journal study. Their goal was to see whether schema quality matters, not just whether schema is present.

They built three very similar one-page sites. One had strong schema, one had broken or incomplete schema, and one had no schema at all. They tried to keep other things the same, including the basic site setup and the keyword difficulty. They used three chosen search queries, submitted all pages for indexing on the same day, and then watched what Google did.

The biggest result was simple: only the page with well-done schema appeared in an AI Overview.

That same page also got the best normal Google ranking, reaching position 3.

The page with poor schema ranked for more keywords overall, 10 versus 6, but only got as high as position 8 and never appeared in an AI Overview.

The page with no schema was crawled quickly but was not indexed, so it could not rank or appear in AI Overviews at all.

The authors say this points to schema quality being more important than just having schema markup on the page.

They also spell out what they counted as strong schema, including complete Article markup, FAQ markup, breadcrumb markup, proper dates, and clear author and publisher details.

There are important catches. This was a very small experiment with just three pages and three target queries, so the findings are shaky and could be a fluke.

The pages were not perfectly identical because they targeted different keywords, and the no-schema page also had a slightly lower word count.