Same website, two different games. This is the comparison I draw on a whiteboard in the first call — nine rows, each one a decision that genuinely differs, with the published data behind it.
The one-sentence version: search engine optimization competes for a slot in a list of ten links, and generative engine optimization competes to be one of two or three sources named inside a single paragraph. Everything below follows from that, including the uncomfortable part — there is no slot four in an answer.
| Dimension | SEO | GEO / AEO | Why it differs |
|---|---|---|---|
| What you win | A slot in a list of ten links | A named mention inside one generated paragraph | There is no slot four in an answer |
| Unit of measurement | Keyword ranking position | Citation share across a prompt set | Answers vary per user; a single rank has no meaning |
| Where the source sits | Positions 1–10 | ~90% of ChatGPT citations come from 21+ | Retrieval samples far deeper than a results page |
| Primary lever | Links, keyword coverage, crawl health | Entity consistency, extractable answers, corroboration | Models resolve who you are before deciding to name you |
| Content shape | Long-form built to hold a ranking | Direct answer in the first 100 words, then evidence | Retrievers lift passages, not pages |
| Structured data | Optional; earns rich results | Load-bearing; it is how the claim is machine-read | Schema is the least ambiguous statement on your site |
| Third-party pages | Useful for links | Frequently the cited source instead of you | Assistants prefer roundups and review platforms |
| Traffic profile | Higher volume, mixed intent | Lower volume, ~4.4× the value per visit | The buyer arrives already qualified by the assistant |
| Time to signal | 3–6 months for competitive terms | 6–10 weeks for entity fixes, a quarter for share | Entity corrections propagate faster than authority |
The claim that page one and the retrieval pool are different sets of pages is the load-bearing one, so it is worth showing it measured twice, independently.
Position one in Google buys you very little here. The assistant retrieves from a different pool than the one most sites have been optimising for.
Semrush, AI Search & SEO Traffic Study, 2025
Across 200,000 keywords. An average AI Overview carries eleven links and includes the #1 organic result just 46% of the time on desktop, 34% on mobile.
Semrush, AI Overviews Study, 2025
Across 300,000 keywords, March 2024 against March 2025. Your rankings can hold perfectly still while the traffic behind them drains.
Ahrefs, AI Overviews Reduce Clicks, 2025
AI-referred visitors arrive further down the funnel because the assistant already did the shortlisting. Low volume, high intent, and it leaks badly without a sequence behind it.
Semrush, AI Search & SEO Traffic Study, 2025
All four are reasonable-sounding, and all four cost a quarter.
The most expensive assumption in the category, and the one both Semrush studies contradict. Ranking is weak evidence of retrievability: the pools overlap at somewhere between a fifth and a quarter. Being on page one is genuinely useful and genuinely insufficient.
Schema matters far more here than it does in SEO — it is load-bearing rather than optional — but bolting it onto a page that never states a clear claim changes nothing. The markup describes the claim; it does not manufacture one.
Volume is the SEO reflex and it inverts here. A model retrieving your category term and finding forty near-identical pages has more to sift and nothing quotable to lift. In a well-ordered programme volume is the last lever, and usually never pulled.
A cited page appeared on about one day in three between its first and last appearance in the Ahrefs experiment. A monthly check therefore has roughly a one-in-three chance of describing your actual position, and no way of telling you which time you got.
SEO competes for a slot in a list of ten links. GEO competes to be one of two or three sources named inside a single generated paragraph. The practical difference is where effort goes: GEO weights entity consistency, answer-first page structure, structured data and third-party corroboration far more heavily than link volume or keyword coverage. The measurement differs too — a keyword has a rank, but an assistant's answer varies per user, so the unit becomes citation share across a fixed set of buyer questions.
No, and anyone saying so is selling something. Ahrefs measured a 34.5% clickthrough drop for top-ranking pages once an AI Overview appears, which is a serious cut but not an extinction. Classic rankings still deliver most commercial traffic for most businesses. What changed is that finishing SEO no longer finishes the job: the same ranking now returns roughly two-thirds of what it used to, and the paragraph above it is being answered by somebody.
Yes, and they are cheaper run together than separately. Technical health, crawlability and clean information architecture are shared inputs — a page an assistant cannot parse is almost always a page a crawler handles badly too. The mistake is not doing SEO; it is assuming the two programmes are the same programme with a different name on the invoice.
Less than you would expect. Semrush found roughly 90% of ChatGPT search citations point at pages ranking 21 or lower, and its separate study of 200,000 keywords found AI Overview links overlap the organic top ten only 20 to 26% of the time. Two different engines, two different methodologies, the same conclusion: the retrieval pool and page one are largely different sets of pages.
Technical health, crawlability, site speed, internal linking and any genuinely useful long-form content all carry over and are worth having. What does not transfer is anything built purely to hold a ranking — keyword-stuffed pages, thin location variants and content written to a word count. Those tend to be actively unhelpful, because they give a retriever more of your site to sift and nothing quotable to lift.
The scorecard takes two minutes. Or send your site and I will run an automated read of it and email you a ranked fix list — free, no call.