Correlation / schema claims

If you've been following GEO advice on how to rank an AI search, you definitely heard about the importance of structured data or schema markup. I think it's safe to say this is the most commonly repeated piece of advice for ranking an AI.

If you want AI to notice you, you need to have structured data or schema markup on your pages. Most of the time, people making these LinkedIn posts or TikTok videos had no data to support this claim, but occasionally they would point to the fact that pages cited by AI are much more likely to have structured data on them.

And that seems like proof, right? Well, recently Ahrefs called BS on that.

They just published, as far as I know, the first large scale study to figure out if structured data has any impact on AI citation rate. Ahrefs understood something that these GEO experts on LinkedIn didn't.

Correlation does not necessarily equal causation. Pages that have structured data on them are just more likely to be optimized and maintained, and therefore also more likely to be cited by AI search...

Ahrefs' blog post provides a lot more data and information, including why they think there might be a slight negative correlation in AI overviews. Their methodology is really impressive, but for those of us that have actually been doing the work and running the experiments, this was no surprise.

The truth is, if structured data actually made AI cite your pages more often, it would be really easy to prove, and there are a lot of incentives for people selling these services to prove it. Of course, that data has yet to emerge, and unless something changes in how these AI systems work, it never will.

Hopefully this is the final nail in the coffin. There are, of course, many other things that will help you get cited and recommended more often by these AI systems, and that's what I talk about in like half of my videos.