I shouldn’t do it. I know I shouldn’t do it.
But last night I ended up on Twitter again, wandering through the latest GEO and SEO commentary like a drunk man walking through a burning circus. And there it was. Another consultant, another thread, another grand revelation from the mountaintop: they were trying to figure out Gemini’s ranking factors. Or re-ranking factors. Or some other pseudo-scientific bollocks dressed up as insight.
At this point, I feel like Ricky Gervais at the Golden Globes. If you are going to write absurd things in public, expect someone to screenshot them and point out that they are absurd.
Because the nonsense has to stop.
The SEO industry is not being killed by AI. It is not being killed by Google. It is not being killed by Gemini, ChatGPT, Perplexity, Claude or whatever shiny interface gets released next week.
The SEO industry is being killed by the SEO industry.
The SEO Industry Has Found a New Thing to Misunderstand
This all started with the first wave of denial.
When generative engine optimisation started becoming a conversation, a large chunk of the SEO community lined up to say the same thing: “GEO is just SEO.”
Of course it is. How convenient.
Then, after spending a year saying GEO didn’t exist, many of the same people suddenly started posting about it every day. Some even managed to get themselves onto listicles of “top GEO experts to follow”, which is a marvellous achievement when you consider they had spent the previous twelve months insisting the thing was imaginary.
Now we have moved into the next stage of the farce: AI rankings.
Apparently, we are now supposed to believe that consultants are going to figure out Gemini’s ranking factors. They are going to uncover the re-ranking signals. They are going to decode the secret recipe. They are going to find the new algorithmic lever and pull it for clients.
No, they are not.
They are doing what the SEO industry often does when faced with change. They are taking a new reality and desperately forcing it into an old commercial model.
The old model was simple. Sell certainty. Sell rankings. Sell repeatable tactics. Sell the idea that somewhere behind the curtain is a hidden mechanism that can be reverse-engineered by clever people with spreadsheets and browser extensions.
That model worked when search looked like ten blue links and the industry could point to keywords, positions, links, metadata, traffic and conversions.
But AI search is not that world.
And pretending it is does not make you a strategist. It makes you a liability.
Google Finally Got the World It Always Wanted
Here is the uncomfortable bit.
Google has never loved the SEO industry. It has tolerated it.
For years, Google has said some version of the same thing: create helpful, reliable, people-first content. Its own documentation still frames SEO as useful only when it helps search engines discover and understand people-first content, not when it becomes search-engine-first manipulation.
That was always the official line.
Google did not want you figuring out links. It did not want you reverse-engineering algorithms. It did not want you building little content factories to chase every keyword variation under the sun. It wanted brands to make useful things, publish useful information, be reputable online and, when they wanted more control, pay Google.
That is the world we are now entering.
Google has pushed AI Overviews and AI Mode because search is moving from information retrieval to answer synthesis. Google says AI Overviews are designed to help people get the gist of complicated questions more quickly, while AI Mode offers a more end-to-end AI search experience with follow-up questions, reasoning and links to the web.
In plainer English: Google is taking more of the search journey into Google.
It is stealing the work of the web, compressing it into answers and keeping users inside its own experience for longer. Google will use more AI because it has to. ChatGPT, Perplexity, Meta AI, Claude and everything else with a prompt box are not minor annoyances. They are attacks on Google’s habitual monopoly over user behaviour.
So Google is responding by making AI the default direction of travel.
That changes everything.
From Keywords to Prompts
The old search economy was built around keywords.
Someone typed “boxing gloves”. That was a broad keyword. Inside it were dozens of possible intentions. They might want to buy boxing gloves. They might want to learn about boxing gloves. They might want kids’ boxing gloves. They might want 14oz gloves. They might want sparring gloves. They might want to know whether velcro or lace-up gloves are better.
SEO turned that ambiguity into a business model.
You found the keyword. You built the page. You got the links. You tried to rank. You measured your position. You reported the traffic.
AI search breaks that neat little machine.
The user no longer needs to type “boxing gloves” and sift through the mess. They can ask: “What are the best 14oz velcro boxing gloves for beginner sparring if I have weak wrists and don’t want to spend more than £80?”
That is not a keyword. It is a problem. It is context. It is intent, constraint, use case and preference rolled into one.
Now multiply that by millions of users asking millions of differently worded questions in different platforms, with different contexts, different histories, different locations and different levels of detail.
You are not going to see enough prompt data to optimise against it.
You are not going to build a page for every meaningful variation.
You are not going to track stable “positions” in the old way.
You are not going to reverse-engineer this with a keyword tool and a smug LinkedIn post.
The unit of search has changed. And when the unit of search changes, the discipline built around the old unit has to change too.
You Are Not Going to Figure Out Gemini’s Re-Ranking Factors
Let’s make this brutally clear.
You do not know the original prompt.
You do not know how the prompt was rewritten.
You do not know what retrieval set was used.
You do not know which sources were considered.
You do not know which passages were extracted.
You do not know how the answer was grounded.
You do not know what user context affected the response.
You do not know what personalisation was involved.
You do not know which model version generated it.
You do not know what evaluation layer sat on top of it.
You do not know the re-ranking process.
And, most importantly, you do not have stable, complete, visible data at a scale that would allow you to optimise against it with any serious confidence.
So when someone says they are figuring out Gemini ranking factors, what they usually mean is: “I ran some prompts, looked at some outputs, made a few assumptions and now I need a commercial angle.”
That is not analysis. That is theatre.
Even the companies building these systems often struggle to explain why a specific answer appeared in a specific form for a specific user at a specific moment. But apparently, ‘Dave’ from SEO Twitter has cracked it after a few screenshots and a spreadsheet.
Do me a favour.
AI search is probabilistic, contextual and unstable. Traditional rankings were at least visible enough to create the illusion of measurement. AI answers are generated, synthesised and shaped through layers you cannot see.
There are no rankings in AI search in the way SEOs mean rankings.
There are citations. There are mentions. There are recommendations. There are source selections. There are brand inclusions. There are answer patterns. There are visibility signals.
But there is no clean, universal, trackable, controllable ranking position sitting there waiting for you to optimise.
Stop selling that nonsense.
GEO Is Real. It Just Isn’t SEO in a Fake Moustache
Here is where the SEO crowd gets confused.
Saying there are no rankings in AI search does not mean GEO is fake.
GEO is real.
But it is not SEO with a shiny new acronym. It is not “technical SEO plus prompts”. It is not blog content churn with a different label. It is not query fanout cosplay. It is not citation counting dressed up as strategy.
GEO is digital brand marketing under AI conditions.
It is about reputation signalling. It is about being known for the thing you want to be chosen for. It is about making your proposition clear, visible and repeated across the internet. It is about getting other people to talk about you, recommend you, cite you, review you, compare you and associate you with a category.
In other words, it is marketing.
Proper marketing.
The kind that makes SEOs uncomfortable because it cannot be reduced to a checklist.
If AI systems are consuming the web, interpreting language, grounding answers and recommending brands, then the job is not to “rank number one” for a prompt you cannot see. The job is to make sure the internet contains enough clear, consistent, credible evidence about your brand for machines and people to understand why you matter.
That means strong copy.
Clear positioning.
Useful product and service pages.
Third-party mentions.
Digital PR.
Reviews.
Category associations.
Influencer mentions.
Publisher coverage.
Social proof.
Brand search.
Direct demand.
Consistent language around what you do and who you do it for.
That is not a ranking factor list. That is a visibility system.
E-E-A-T Is Not a Strategy Either
And while we are here, let’s deal with another comfort blanket.
EEAT is not a strategy.
Expertise, experience, authoritativeness and trustworthiness are useful ideas. They are not a positioning strategy. They are not a marketing plan. They are not a reason for anyone to choose you.
Saying “we need more EEAT” is like saying “we need to be better”. Fine. At what? For whom? Against which competitor? In which category? With what proof? Through which channels? With what budget? To create which commercial outcome?
This is where a lot of SEO thinking collapses.
It loves concepts that sound strategic but avoid the hard work of strategy.
Strategy means choices.
Who are we targeting?
What are we trying to be known for?
What do we want the market to believe about us?
Where do we need to be visible?
Who needs to talk about us?
What proof do we need?
What are competitors already associated with?
Where are we absent?
What activity will actually move the needle?
That is GEO.
Not another acronym. Not another checklist. Not another “ultimate guide” written by someone who discovered the topic six weeks ago.
Read These Two Books Before Posting Another Thread
If you actually want to understand where this is going, stop reading SEO influencer threads and read things that explain search, language and networks.
Start with Jun Wu’s The Beauty of Mathematics in Computer Science. It explains the mathematical foundations behind information technology, including natural language processing, statistical language modelling, information measurement, graph theory, web crawling and PageRank.
Then read Albert-László Barabási’s Network Science. His work explains why networks are not random, why hubs emerge, why connected nodes become more connected and why visibility compounds. The online version of Network Science is openly available, and its chapters explain the architecture of real networks, hubs and preferential attachment.
This is far more useful than another post about “AI visibility hacks”.
Because AI search is not just about content. It is about how language systems interpret the world and how networks distribute authority, association and attention.
Preferential attachment matters.
Being mentioned by the already-visible matters.
Being connected to trusted entities matters.
Being repeatedly associated with a category matters.
Reducing ambiguity around what your brand does matters.
If the internet is a giant messy network of language, links, mentions, entities, publishers, platforms, creators and brands, then GEO is the work of making your brand easier to understand and harder to ignore inside that network.
That is the game.
Not “what is Gemini’s fifth re-ranking factor?”
SEOs Can Still Matter. But Not Like This.
This is not an argument that SEOs are useless.
Far from it.
A good SEO can become incredibly valuable in the GEO era. But only if they stop pretending they own the whole thing.
The future role is not “I found the ranking factors.”
The future role is “I can help the business understand where it is visible, where it is absent, what competitors are doing, what AI systems are surfacing, what content is being cited, what brand searches are changing, what pages are working and where marketing money is being wasted.”
That is a valuable role.
Someone needs to monitor brand search.
Someone needs to analyse web traffic.
Someone needs to understand how visibility is changing across Google, AI Overviews, Gemini, ChatGPT, Perplexity and the wider web.
Someone needs to look at competitor mentions.
Someone needs to track whether PR, content, social, reviews and partnerships are having an impact.
Someone needs to sit in the data and tell the marketing team what is happening.
That person could be an SEO.
But only if they become a search marketer in the broader sense.
Not an algorithm chaser. Not a ranking-factor fantasist. Not a blog-retainer merchant. A proper search marketer who understands that search now sits inside a wider system of brand, content, PR, social, reputation and demand.
GEO will not be a single channel.
It will be the connective tissue between channels.
That means collaboration. It means the SEO person does not get to ram SEO down everyone’s throat and declare themselves the leader because they learned a new acronym.
Humility would be useful.
There is a lot of humble pie going around. Some people should eat a second slice.
Brands Need More Marketing, Not More SEO Theatre
The timing of all this is absurd.
Marketing budgets are being cut. Teams are being reduced. Publishers are under pressure. Traffic is declining. Algorithms are strangling organic reach. AI systems are answering questions before users click anything. Everyone is fighting for attention while the platforms keep more of it for themselves.
And at exactly this moment, brands need more marketing, not less.
They need to be more visible.
They need stronger direct demand.
They need more people searching for them by name.
They need more publishers talking about them.
They need more influencers mentioning them.
They need more customers reviewing them.
They need clearer propositions.
They need better copy.
They need better category positioning.
They need more evidence across the network that they are worth recommending.
But instead of building consensus around that, parts of the SEO industry are creating more division by pretending the answer is another technical rabbit hole.
We have had citation counting.
We have had AI visibility scoring.
We have had query fanout.
Now we have Gemini re-ranking factors.
And the industry keeps chasing its tail while brands need actual help.
This is damaging because it sends budget in the wrong direction. It convinces clients there is another shortcut. Another hack. Another secret. Another thing an expert can uncover if they just pay the retainer.
There is no secret sauce.
There is no hidden lever.
There is no AI ranking throne waiting for you.
There is only the hard, expensive, necessary work of becoming visible, credible and chosen.
The Future Is Visibility
The brands that win in AI search will not be the ones obsessing over fictional ranking factors.
They will be the ones that become widely known for something specific.
They will have clear positioning. They will invest in copy. They will build useful pages. They will earn mentions. They will get talked about by credible people in credible places. They will grow brand search. They will create demand. They will show up across the network often enough, clearly enough and credibly enough that both humans and machines understand what they are about.
That is not “just SEO”.
That is not “just GEO”.
That is marketing in a digital ecosystem where search has become generative, fragmented and less controllable.
Google has got the world it always wanted. Create useful content. Be a good brand online. Accept what organic visibility you get. And if you want more control, pay.
The only way around that is not to crack Gemini.
It is to build a brand that the market, the web and the machines cannot easily ignore.
So no, there are no rankings in AI search. Not in the way the SEO industry wants there to be.
There are answers.
There are recommendations.
There are mentions.
There are associations.
There is visibility.
And there is a lot of nonsense being sold by people who should know better.
Stop buying it.
Andrew Holland
Originally published on LinkedIn on 31 May 2026.