Why I think GEO might be a thing. And why I don't

If you've followed me on LinkedIn much of late, I've posted a few times about how GEO - that is 'Generative Engine Optimisation' - is not a thing.

(I also posted to the same effect on this newsletter a few weeks back).

My position here is that I generally work with large traffic sites where Google is now by far the most dominant source. Social media has all but withered away, leaving search taking a bigger slice of the pie.

GEO has entered as a potential traffic source. But the referral data is comparatively tiny and messy. So I still think it's an incredibly small thing to be concerned about for website owners.

Many counter comments have stated about its potential. I agree chatbots will grow in usage over the next few years, but they will probably evolve into general purpose applications that stretch way beyond information retrieval (see Gemini 3, which I have nowhere near had enough time to review just yet).

And with that... I don't know, it's kind of foggy even 2 years ahead, but I still don't think chatbots will replace transactional search in a significant way over 2 years. In my view, ChatGPT and Claude won’t turn into 'major' referrers at say 20% (combined) traffic across big websites. The reason being is Google has a very capable AI, and a largely independent AI ecosystem (as last paragraph). So if anything is going to eat Google search, it's Google, which will occur quite subtly.

So GEO isn't a meaningful thing in terms of web metrics right now, and I don't think attempting to appear on ChatGPT and co is a meaningful endeavour right now either. I'd even say that about Google's AI Overviews.

The reason being is all of these points of distribution are fairly unstable - in that the 'results' are transient. You're in AIO one query, not in it the next. All the recommendation based articles I see about GEO align to SEO fundamentals.

There are some experts, like Mike King, who go into extraordinary detail about GEO and make the boundary with SEO clear. I applaud the research, but this detail doesn't make it any more a thing, as the contrast is still not that high, and the metrics, as stated, are low.

Over the next year, Roku predicts that 100% of the streaming audience will see ads. For growth marketers in 2026, CTV will remain an important “safe space” as AI creates widespread disruption in the search and social channels. Plus, easier access to self-serve CTV ad buying tools and targeting options will lead to a surge in locally-targeted streaming campaigns.

Read our guide to find out why growth marketers should make sure CTV is part of their 2026 media mix.

Learn more.

But I had an actual physical experience this last week that might be the beginning of changing my mind. It wasn’t out of body or anything. Physical. Just involving my hands and general body.

It went like this: I moved out of a rental house and after 2 years there was a fair amount of wear and tear that was probably beyond our contractual obligations.

Thus there were various risks associated with not fixing them. There were about 5 in total. Odd jobs, all completely different, and ultimately difficult for one person to do on a call out. So I went to ChatGPT and one by one, through taking photos of the issues and chatting back and forth, it recommended solutions with associated products.

One, called Barkeeper's Friend, was an astonishingly effective cleaning agent that I'd never heard of before. Another was a type of floor resin that got rid of all the scratches on a wooden floor. As tedious jobs go, it was really quite satisfying seeing numerous physical dominos falling as I used AI to reach a solution. 

The thing was, I think I only clicked through from ChatGPT once, through an Amazon link. The rest of the time I accepted Google’s recommendation and Googled the product. Then I largely bought via Amazon, or went to a physical store, if I needed something quickly. 

So in this scenario, GEO probably could be a thing, but more for product creators rather than any third party advisory content or product curators (like stores). The AI essentially recommends how the product could be used in a hyper personalised scenario.

The funny thing in the pre AI web, the product creator website was largely redundant. It's doubtful I would have come across Barkeeper's Friend via a Google search for 'multi purpose cleaners' or 'how to clean stubborn wire wool scratches'... and then read the information on the product domain. I would've surfed all over the place, getting wiped out by the endless choice on offer. Contrastingly I want my AI to tell me exactly what to do, and it did.

This triggered something about GEO and the role of product content. If the product creator website (or content package) can recommend all the various scenarios where it could be useful in a well written and authoritative way, then this will be a good way to appear on the many prompts and queries that will be entered into chatbots to find solutions in the next few years.

This is GEO in action because the queries are hyper personalised, often using images, and then you can be returned scenario based recommendations. 

But let's not pretend this is new. It's essentially what product testing sites, like Which? etc, do a lot. They review products and test them, which creates the diversity.

And I also think, isn't the creation of such content just good SEO? And I realise I've just gone in a circle of saying, yes, it is.