Search Traffic Forked: SEO and GEO Are Different Games

For twenty years, getting found on the internet meant one thing: rank higher than the next result. You wrote the page, you earned the links, you climbed the list, and users clicked the blue link near the top. The whole discipline of SEO is built on that single mechanic — a race up an ordered list where the user makes the final call.

That mechanic is quietly being replaced for a growing share of searches, and if you're a solo founder standing up a product with an independent site as your one piece of owned ground, the replacement changes what "getting found" even means.

The click-through collapse

Start with the number that should reorder your priorities. Pew Research Center, in a July 2025 study of real browsing behavior, found that among users who encountered a Google AI Overview, only 1% clicked through to one of the sources the AI cited. Not 10%. One. The answer sat at the top of the page, synthesized and good enough, and the reader took it and moved on.

The knock-on effect on the classic list is just as sharp. Ahrefs analyzed 300,000 keywords and found that when an AI Overview appears, the click-through rate of the #1 organic result falls by 58%. The position everyone fought for still exists — it just gets far fewer clicks, because the AI answered the question before the reader ever reached the links.

And this isn't a rare event. Google's AI Overview now surfaces on an estimated 25% to 60% of searches depending on the query type and the study you trust. Similarweb tracked the aggregate effect: the zero-click search rate — searches that end without any click to any external site — rose from 56% to 69% in the year after AI Overview launched. More than two in three searches now resolve without a single visit to a website.

Put those together and the picture is stark. A rising majority of searches never produce a click, the AI answers in place, and the one organic slot worth winning gets cut nearly in half when it does.

Two paths, not one worse path

The easy mistake is to read this as "SEO is dying" and doom-scroll to the next thing. That's wrong, and it'll cost you. Search didn't shrink — it forked. There are now two distinct paths a query can take, and they reward completely different work.

The first path is the one you know. A user types a query, gets a ranked list of blue links, scans it, and clicks the result that looks best. This is still a huge amount of traffic, especially for navigational and transactional queries, and traditional SEO still governs it. Everyone competing is in the list; the user picks the winner.

The second path is new. A user asks a question — in Perplexity, in ChatGPT's search, in Google's AI Overview, in Gemini — and gets a written answer synthesized from multiple sources, with a handful of citations attached to the side. Most readers never open those citations. The AI already did the reading. Your job on this path is not to be picked from a list; it's to be one of the sources the AI drew from when it wrote the answer.

Traditional SEO is a race for rank — everyone's in the list, and the user decides who gets the click. GEO is a race to be the source, whether or not anyone clicks at all.

That's the whole distinction, and it's worth sitting with because the two paths pull in genuinely different directions. Ranking #3 for a keyword can be a triumph in SEO and completely irrelevant to GEO if the AI synthesizes its answer from three sources that all rank below you. Being invisible on page two of Google can still get you cited constantly by an AI engine that values what your page says over where it sits. The two systems overlap but don't coincide.

Why the old scoreboard stops working

SEO gave you a clean scoreboard: rank, traffic, click-through rate, all measurable, all visible in an analytics dashboard. GEO breaks that scoreboard in a specific way — the win is often invisible on your own metrics.

When an AI engine cites your page in an answer and the reader absorbs it without clicking, you got the outcome you wanted — your product name, your framing, your data point landed in front of exactly the right person — and your traffic numbers show nothing. No session, no pageview, no referral. The value arrived and the analytics stayed flat.

This is why founders who judge GEO by their traffic graph conclude it isn't working and quit. They're measuring the old game's scoreboard while playing the new game. The right instruments are different: is your brand name showing up in AI answers when someone asks about your category? Are AI engines pulling your specific claims and numbers? Those are the questions that matter now, and none of them live in a pageview count.

The mechanics under the hood are different too, and that's what the rest of this series is about. A ranked list is decided largely by links and on-page signals. An AI-synthesized answer is decided by something closer to reputation — what the broader internet says about you, how clearly your content states verifiable facts, whether your site reads as one coherent knowledge node or a pile of orphaned pages. Optimizing for the second path is not "SEO but harder." It's a separate craft with its own counterintuitive rules.

What this means before you write another page

If you're building solo, your independent site is the one asset no platform can take from you — but it's only an asset if it gets read, and increasingly the most important reader is a machine writing an answer for someone else. A few things worth internalizing before you touch your next piece of content:

• Stop treating traffic as the only proof of visibility. A citation with no click is still a win. Judge GEO by whether AI engines mention and quote you, not by your session count.
• Don't abandon SEO — it still owns a real path. Navigational and transactional queries still send clicks. The fork is additive: you now have two games to play, not one to replace the other.
• Accept that ranking and citation are different targets. Winning one doesn't guarantee the other. Some of your best-cited pages will never rank; some of your best-ranked pages will never get cited.
• Recognize the win is often invisible on your dashboard. Build the habit of checking AI engines directly — ask Perplexity or ChatGPT about your category and see whether you appear — because your analytics won't tell you.

The founders who get this early get a real edge, because most of the internet is still optimizing exclusively for a list that fewer and fewer people scroll. The attention has moved into the AI's answer, and the answer is written from sources — so the game is to become a source.

Over the next four pieces, this series gets practical about how. Why AI engines trust what others say about you far more than what you say about yourself, and the counterintuitive data behind it. Why Reddit specifically has become the single highest-leverage channel right now — and why that window won't stay open. The schema markup that actually moves citation rates, and the far larger caveat that it does nothing without real content behind it. And why a structured cluster of connected articles gets cited far more than the same words scattered across unrelated posts. The fork is the setup. The rest is execution.

---

Part of a 5-piece look at GEO — getting cited by AI search engines. Next: Why AI Engines Cite Reddit Over Your Homepage, then The Reddit Playbook for Solo Founders Who Hate Marketing, Ten Schema Tags That Get You Cited by AI, and Pillar Pages and Content Clusters for AI Citation. If you're still standing up the basics, start with getting a domain live in twenty minutes and shipping a static site with nothing but git push. More builder insights.