@tjrobertson52

Is ranking in large language models like ChatGPT the same as ranking in traditional search engines? In other words, is GEO just SEO in a fancy costume? I have multiple videos talking about why I think GEO is in fact different than SEO.

However, I've now upset a lot of SEOs who say it's the same thing, but I wanna Steelman their argument and then I wanna provide a helpful metaphor that might clear up the misunderstanding.

As many SEOs will point out, when you do a search in a large language model like ChatGPT, it will then perform its own search in a traditional search engine. So the results it returns are the same results you would get if you went to that traditional search engine. Therefore, it seems entirely reasonable to say there's no difference.

Just do SEO, you'll rank in the traditional search engines, the AI will return those exact same results, and therefore you will rank in the AI search results.

However, the difference in GEO or AI SEO is not a difference in how the search engine works, it's a difference in search behaviour. Large language models use traditional search engines much differently than humans use traditional search engines, and that does change the strategy. Although it doesn't seem to matter how many times I say this, I keep getting misunderstood by other SEOs, so I have a metaphor that I think will help make this clear.

Let's say you're a company like General Mills that makes food products that will be sold at grocery stores. You have an idea for a new product. It's a gluten free cereal that's also high in fiber.

However, there's a limited amount of shelf space in each grocery store. They might have a section for gluten free cereals, but probably not enough room for your gluten free, high fibre cereal. Now, I don't know enough about cereal to know if this analogy actually tracks, but let's just say your product is too specific to make it onto grocery store shelves, and therefore this is a bad strategy.

Now, let's say something changes overnight, and all of a sudden all the grocery stores have not one, but 10 cereal aisles. 10 aisles devoted specifically to cereal and one entire aisle just to gluten free cereal. This changes the strategy.

It doesn't change how you make cereal or how you sell your product, but it does change which products make sense to develop in the first place.

That's the shift we're seeing in large language models. When a human does a search in Google, they search for one broad competitive term and they look through one, maybe three websites. In other words, the available shelf space that you're competing over is very small.

In that situation, the best strategy is to find terms that are getting high search volume where you can rank in the top three spaces.

On the other hand, when you prompt a large language model, it's gonna do a search in the same search engine, but the search is gonna be way more specific, and it's probably gonna do up to a dozen different searches. And then it's not just gonna look at the three top results, it's gonna look through multiple pages of results. In fact, large language models typically look at about 100 times as many pages as a human.

This opens up a ton of virtual shelf space.

So while the underlying search engine is the same, the strategy changes dramatically. Suddenly, it's worth making hyper specific content. Even if you can't rank on page one of Google, it makes sense to get your company mentioned on third party websites that might not rank on page one of Google and might not link back to you.

AI is exploring dark corners of the internet that no one cared about with traditional search engines before.

So I still think there's more that we don't know than we do know, but things are definitely changing. Yes, SEO still works. SEO is still important.

You do have to show up in the search engines, but things are changing quickly, and your strategy should change along with it.