Search Is Being Rewritten by People Who Don't Click Links

A while back, a friend of mine who works in marketing came to me with a complaint.

He said his company's website traffic over the past year or two felt like someone had quietly turned down the tap. Everything that was supposed to be done for SEO had been done. Keyword rankings hadn't slipped. And yet the clicks coming in were somehow a big chunk lighter than before.

I asked him, "Have you considered that those 'missing clicks' didn't actually disappear — it's just that nobody's clicking through to you anymore?"

He stopped short, thrown off.

I shared a number with him. Earlier this year, Google was processing more than 14 billion searches per day. At the same time, ChatGPT — the thing everyone talks about every day — was at 37 million.

14 billion versus 37 million. A 373x gap.

You might say, "Well then Google's the clear winner, right? If I just optimize for Google, I'm fine."

Yes and no.

A Crowd of People Who Don't Click Links Is Quietly Growing

Think about it — what does it feel like when you use a search engine these days?

Up pops a screen full of ads, shopping links on the left, some kind of card on the right, and a few "People also ask" blocks shoved underneath. By the time you actually want to click through to see what's going on, a dozen widgets have already worn away your patience.

The old-school SEO crowd probably still remembers Google's signature move from the early days — the move that put it on the map. It would proudly announce on the page: this search took 0.3 seconds. Clean. Crisp. Fast.

And now? Screen after screen of widgets and ads has turned Google into the Yahoo of twenty years ago.

And that's exactly where the window of opportunity opens.

Generative AI offered the reverse answer: you don't have to flip through pages, I'll just write the answer out for you. ChatGPT, Perplexity, and Google's own AI Overviews all walk this path. You ask a question, it gives you a summary, and you don't need to click anymore.

And so a new group of people appeared: those who only want the answer and have no interest in clicking links.

For now, they're still hundreds of times fewer than Google's users. But the habit they represent, once formed, doesn't go back.

Enter GEO: A New Business Built on "Being Cited"

So a group of people started working on a new word. GEO — Generative Engine Optimization. Some people call it AEO, Answer Engine Optimization. Different names, same idea.

What is GEO?

SEO is about getting people to click on you in the search results. GEO is about getting AI to cite you when it generates an answer.

These two things look similar, but they're actually two completely different playbooks.

When you do SEO, you think: a user searches "which coffee maker is good," I push my coffee maker page into the top three of the first page, and they click in. It's one-shot. Catch a query, trade for a click.

Doing GEO, you're thinking about something else entirely.

When generative AI gives an answer, it doesn't copy a paragraph from a single webpage. Inside its mind, it mashes together every entity related to the concept of "coffee maker" — brands, models, reputation, price, user reviews — and then outputs a paragraph. Whether it mentions your brand depends on how clearly it recognizes you as an "entity" in its mind.

What's an entity?

An entity is a "thing" that an AI can talk about without hesitation. A person, a company, a product, a place. When an AI says "what this company does, what makes it distinctive," and says it accurately, consistently, and the same way across platforms — that's when your entity has been "remembered."

The battlefield of SEO is position on the results page. The battlefield of GEO is recognition inside the AI's mind.

Position can be fought for. Recognition can't be seized — it can only be cultivated slowly.

So Is Google About to Die?

Short answer: no.

With that 373x number sitting right there, you know it's way too early to be writing Google's obituary.

But here's the interesting part: the exact same way Google took down Yahoo back in the day is now turning around and biting Google itself.

Twenty years ago, Yahoo's homepage was cluttered and chaotic, jamming everything in. Google offered a search box so clean it was almost stark. Users voted with their feet and just left.

Today, Google's own search results page is stuffed with more ads and widgets than Yahoo had back then. Generative tools swooped in and handed users that clean answer they'd been craving. History doesn't repeat itself, but it often rhymes.

So the right move isn't to pick a side. "Do I bet on SEO or bet on GEO?" — that binary choice is itself a false dilemma.

The reality is: you still need to keep eating Google's traffic, but you have to start playing the new game of GEO right now. Because recognition is the kind of thing that rewards slow, careful work. By the time everyone else is already being mentioned in AI's answers and you only then remember to start cultivating it, you'll have missed the boat.

What's Genuinely Counterintuitive Is How You Do the Math on This

At this point, a lot of people get nervous: "If GEO brings in so much less traffic, isn't my ROI going to be a loss?"

That's when you need to get the math straight.

There's data showing that traffic coming from AI search is down more than ninety percent compared to traditional search; from chatbots, it's down nearly 96%. Sounds terrifying.

But that small slice of people who do click through — they're pure gold.

Why?

Because they've already done all their homework on the AI side. Compared what needed comparing, learned what needed learning — the AI has already done the user-education step for you. By the time they actually click out of the answer, they're not browsing. They're ready to pull out their wallet.

So the "traffic" KPI needs to be redefined in the GEO era.

Just looking at unique visitors doesn't make sense anymore. You have to look at the value behind each UV. A hundred thousand casual clicks may not be worth a thousand clicks from people ready to buy.

But the Hardest Part Is That GEO Can't Be Measured Right Now

The nice thing about SEO is that the toolchain is thoroughly mature. Which keyword ranks where, which page's traffic went up or down, how many backlinks — all of it, right there at a glance.

GEO? Right now it's a fog.

There's no unified platform, no unified metric. It's very hard to know precisely whether AI actually mentioned you when it generated an answer, how many times it mentioned you, and whether what it said was good or bad.

So what do you do?

Start moving, don't wait. There's actually quite a lot you can do:

Keep a close eye on how the major models talk about your brand — ask them a few questions regularly and track the changes. Pay attention to the search trend on your own brand terms — that's a lagging indicator of AI citing you. Shift your KPIs away from "traffic" and toward "engagement" and "leads" — stop chasing raw UVs.

New tools are popping up too, with companies big and small all scrambling to fill this gap. A year from now, this will all be much clearer. But during this year, the people who move first will stake out their position inside the AI's mind first.

Coming Back to That Friend

After we talked, he asked me one thing: "So should I turn my SEO team into a GEO team?"

I said, absolutely not.

You still need to do SEO — Google's 14 billion is still sitting right there. But within your team, you need to carve out a group of people dedicated specifically to figuring out "how to make AI remember us." This is a whole new craft — it's not something you can pull off by dusting off the SEO playbook and tweaking it.

He went quiet for a moment after hearing that, then said, "This feels like 2011."

I didn't say anything. But I suspected what he actually meant was 2011, when Google suddenly cut off keyword referral data. The whole SEO industry was blindsided — they thought the sky was falling. And then? The sky didn't fall. The industry just found its footing and kept moving forward.

Traditional search gives you a list of results. Generative search hands you a single answer.

The users behind these two paradigms have completely different intents. One wants to pick for themselves. The other wants to be handed the answer.

What you need to do is make sure the people who want to "pick for themselves" can find you — and make sure the people who want to "be handed the answer" see you inside the answer AI gives them.

Both crowds. You can't lose either one.

As for that crowd of "people who don't click links," don't write them off as lost traffic. They're actually a harder-to-win, but more valuable group. What they want isn't a click. What they want is for AI to put in a good word for you in front of them.

How many clicks is that one sentence worth? You do the math.

I don't know the answer either. But that question is worth every brand owner thinking about tonight.