The new marketing buzzword is GEO – but is it real?

The new buzzword in marketing is GEO, Generative Engine Optimisation, but is it worth paying attention to? Or is it just more marketing spin for SEO?

The short answer? Yes, GEO is absolutely real. Search Engine Optimisation focuses on getting your website to rank higher for buyer-intent keywords in search results; GEO (Generative Engine Optimisation) is something entirely different.

It’s about increasing the likelihood that AI language models will recommend your business in buying situations. We’re looking at two distinct optimisation systems here.
Some call it ‘answer engine optimisation’, others ‘AI SEO’, but GEO seems to be sticking.

The name isn’t too important though. What matters is the end result: your brand being recommended by AI to buyers. Because, like it or not, AI is going to be building consideration sets for our customers. AI untangles the messy middle of search for users, so you need to get your head around how it works.

Keywords are dead, category entry points are king

If you’re familiar with Ehrenberg-Bass’s work, you’ll be aware of category entry points, the emotional and situational triggers that prompt consumers to enter the market to buy. Well, AI-powered search is fuelled by them.

Keywords have always represented impoverished search. We used them because they were the best way to communicate with search engines. But LLMs? They’re different.
Instead of users having to figure out the best keywords to find products and services, you can just tell ChatGPT and Google’s AI Mode your problem, and they do the searching for you.

They slice and dice your prompt and check all the sites for you, verifying them in advance, like having a personal assistant in your pocket. Essentially, AI analyses a user’s category entry point and develops a search strategy to identify the most suitable solutions. Your job as a brand is to give the AI everything it needs to increase the likelihood you’ll be recommended.

What your brand needs to give AI

We used to say backlinks were the currency of the web. Well, Mutual Information (MI) is the currency of AI. But what is it?

Mutual Information (MI) is the mathematical backbone of relevance. In short, natural language processing systems hate ambiguity, and mutual information helps them to increase certainty.

At a sentence level, if we say, “Russell loves his new Apple”, a machine doesn’t know if you mean a rosy red apple or a new iPhone. Just adding the word iPhone supplies more information for AI so it now knows what you’re talking about.

‘Profound implications for the industry’: Mastercard’s CMO on marketing’s agentic AI futureThis is the backbone of optimising for AI. You are increasing the supply of contextually relevant information about your brand to increase its ‘AI availability’.

That’s information across the web, in all the places that AI looks to find it, in many ways GEO is a bit like a twisted digital game of hide and seek. Except you want your brand to be easy to find in all the obvious places AI looks.

It’s how easily available you are to AI that matters, and how certain it is about your brand being the right one to recommend.

How brands grow in AI

As you’re a Marketing Week reader, I’m sure you know all about Byron Sharp’s ‘How Brands Grow’. But, for a quick recap, brands grow by increasing sales, and sales are more likely to occur when you increase both mental and physical availability.

Mental availability = the likelihood of a brand being thought of in a buying situation.

Physical availability = the ease and convenience you can be purchase from.

And now I’ll throw in a new model: AI Availability. The likelihood of a large language model (AI) recommending a brand in a buying situation.

It’s the amount of mutual information that supports your brand’s positioning which dictates how likely your brand will be recommended. And, yes, this puts words back in the driving seat of marketing.

Words are now the most valuable asset in marketing

Words have always been part of the SEO industry, in fact, if hired a content writer, they generally charge by the word… or they did do.

Generative AI changed all that, making words free. The result of which has been job losses galore. Well, writers are about to have the last laugh.

You see, words are what convey the ‘mutual information’ we discussed earlier. If a single word can help a machine understand if you’re talking about an iPhone or the one a day to keep the doctor away, imagine what millions of words about your brand does.

Which is why ‘word-based assets’ are going to be so important. And I’m not talking about churning out SEO blogs like the industry did for the last decade.

‘No magic pixie dust’: How B2B marketers are approaching the rise of agentic AIWord-based assets are the things that provide and build that mutual information for AI to consume. AI search is there to serve its masters and help them to solve their problem, so you need to supply information that shows not only how you solve the problem, but how you’re different from anyone else and why you’re the best.

You supply this through web copy, PR and other marketing communications. Your consumers supply this in the form of digital conversations around your brand – and if all of this sounds like brand marketing, well, it is.

It’s just that certain types of mutual information matter for certain types of brands -which is part of the fun of GEO – figuring out which levers to pull.

And you need to figure this out, and quickly, because search will radically change in the coming months…maybe less.

The death of traffic (as we know it)

Let me be provocative. The SEO industry as we know it will die. Not gradually but in a single day: the day Google makes AI Mode their default search experience. And there’s every reason to believe they will.

In recent weeks, Google has eliminated SEO rank trackers by modifying how they access data, making it prohibitively expensive to track rankings beyond the first page for a keyword.

That’s an entire industry’s primary measurement of performance gone overnight. It’s a little like throwing dirt into the eyes of all SEOs, making them blind. There are rumours of why they’ve done this; many are just conspiracy theories about stopping the way OpenAI uses their search engine.

But it makes SEO a bit like falling asleep at the start of the train journey and waking up when you arrive at your destination. Welcome to page 1. You know the destination, but you won’t get to see the journey anymore.

The thing is there’s going to be many more Google changes as we approach ‘G-Day’, the day they switch to AI mode as the default and all rankings become obsolete, making being generated in AI results the new online gold.

However, they have a sticky issue that’s tough to solve first.

The advertising dilemma

Google faces a fundamental challenge: how do you create an advertising model that works with LLMs?

When an AI’s primary job is to provide the best quality information, paid recommendations feel contradictory. ChatGPT and other platforms currently operate without ads, monetising through subscriptions instead.

Google’s cash cow is AdWords. You see, SEO was primarily based on user laziness. The old joke that the best place to hide a body is page 2 of Google was true. Humans were too lazy to scroll to page 2, and even if the brand’s ranking weren’t the best, they were ‘good enough’.

Paid search was even lazier still. If you can’t be bothered to scroll, here are some people who paid us to be at the top. LLMs deliver the best results for users, so any advertised brand in those results will either stand out like a sore thumb, or simply not be as ideal as what AI recommends.

AI is going to eat your search traffic – don’t let it eat your brandSure, you can place display ads, entice users with offers and discounts, and maybe even Google can offer a premium model, a little like ChatGPT for ad-free use. But it’s hard to see how this ends well for Google.

Even if they buck the trends of failed products like Stadia and Google Glass, as well as their attempt at a social network, Google+, it’s hard to see how AI mode doesn’t wreak economic chaos on their shareholders.

For you, though, what matters is that you start to plan today. When I was in the police we had a saying. Someone was either a known risk or an unknown one. There was never no risk. And for your brand, you need to find out your level of risk.

If you’re overleveraged on paid search, what happens if those costs rise dramatically? If you’re winning on organic search, what happens when it’s gone? And above all else, are you being recommended to buyers by AI now?

The changes in search will impact your brand. Just how much depends on your AI availability.