You did the work. You picked the keyword, wrote the better page, earned the links, and watched it climb. And one morning there it is — position one, the spot everyone’s been chasing for fifteen years.
Then you open the traffic report, and the line is flat.
I’ve had that exact morning, and so has every SEO person I know. It’s a strange kind of hollow win — you won the thing you were told to win, and somehow it didn’t pay. The first instinct is to assume you measured wrong, or that it’s seasonal, or that the click-through just needs a week to catch up.
But more and more, that’s not it. The page is fine. The ranking is real. The clicks just aren’t there anymore, because something changed about what a Google results page even is.
So let’s look at what the numbers actually say — because a few people went and measured it.
Why Ranking #1 Doesn’t Grow Traffic Like It Used To
The clearest read comes from Rand Fishkin’s team at SparkToro, who published a 2026 zero-click study built on Similarweb’s panel data of real US Google searches.
A “zero-click” search is just what it sounds like: you type a query, you look at the page, and you click nothing. You got your answer right there on Google and moved on. No link, no visit, no traffic for anybody.
So here’s what they found, for searches between January and April 2026:
- 68% of US Google searches ended in zero clicks — more than two out of three.
- AI Overviews (the AI-written summary that sits at the top of results) showed up in over 20% of searches.
- When that AI summary appears, the click-through rate on the links below it drops by roughly 60%.
Sit with the shape of that for a second. Your rank is a measure of order — who sits above whom in the list of blue links. But if two-thirds of searches never produce a click, then winning the order is winning a smaller and smaller pie. You can be first in line for a train that fewer people are boarding.
And the AI Overview makes it sharper. When Google answers the question itself, the links underneath don’t just slip a little — they lose more than half their clicks. So the better Google gets at answering, the less your #1 spot is worth in actual visits. The ranking is intact. The payoff leaked out from under it.
This isn’t only SparkToro saying so, either. The Pew Research Center tracked the same effect from the user side and found that clicks roughly halve when an AI summary is in the results. Two different teams, two different methods, the same direction. When the AI answers, the clicks go quiet.
But Don’t Tear Everything Down In A Panic
Now, before you delete your SEO playbook and rebuild around chatbots, the same SparkToro study has a brake pedal in it, and it’s worth pumping.
During that same period, the searches that actually moved over to a conversational “AI Mode” — the full chat-style search where you talk to Google like you’d talk to ChatGPT — were just 0.34% of the total. Not 34%. Zero-point-three-four. The overwhelming majority of people are still using the classic results page, scanning the same list of links they always did. The main stage hasn’t moved.
So if anyone’s telling you the search results page is dead, come on — the data says it’s still where almost everyone is standing.
But here’s the part that keeps it from being a comfortable “carry on, nothing’s changed.” At its developer event (Google I/O), Google said AI Mode had passed 1 billion monthly users, and that the volume of queries it handles is more than doubling every quarter. A thing can be 0.34% of searches today and still be the fastest-moving floor in the building.
So I think the honest stance is two-eyed. One eye on reality’s main stage, which is still the classic results page where the traffic lives right now. One eye on the center of gravity, which is sliding toward the AI side a little more every quarter. Not “switch.” Not “ignore it.” Watch both, and notice which way the floor is tilting.
The Real Shift: From Ranking To Getting Cited
Here’s the move underneath all of this.
The old SEO game was a game about order — get above the next site, climb to position one, win the click that rank was supposed to deliver. That whole game assumes the click happens. With 68% zero-click, the assumption breaks. You can win the order and still never get the visit, because the visit isn’t on the table.
So what’s left to win? Getting cited. Even when nobody clicks, being named and quoted inside the AI’s answer still does the job a click used to do — it puts your brand in front of the person, in a context they trust, at the exact moment they’re deciding. The exposure didn’t disappear when the clicks dried up. It moved — out of the ranked list of links and into the body of the AI’s answer.
That’s the whole shift in one line: the battleground moved from where you rank in the list to whether you’re in the answer at all.
And it’s not just Google’s results page where this is playing out. The pressure is coming from a few directions at once, and the primary data all points the same way:
- Cloudflare reports a widening “crawl-to-click gap” — AI bots crawl sites heavily to build their answers, but send back far fewer human referrals than they take. The model reads you; it just doesn’t always pass the visitor along.
- Research from London Business School and UCLA found that people who started using ChatGPT cut their traditional search by more than 20%. That’s demand quietly walking off the search page entirely.
Put those next to the zero-click and AI Overview numbers, and the picture stops looking like one weird quarter. People are getting more of their answers from AI, and clicking through to sites less while they do it. Rank still describes your place in a list. It just describes less and less of where the exposure actually happens.
Conclusion: Measure Whether You’re Cited, Not Just Where You Rank
So what do you actually do with this on a Monday?
The one habit worth breaking is judging your content by rank alone. I think rank now tells you maybe half the story — it’s still a real signal, especially while 99.66% of searches sit on the classic page, but it goes blind exactly where the new exposure lives: inside the AI’s answer. A page can rank first and never get named by the AI, or rank fifth and get quoted in every summary. If rank is your only gauge, you can’t tell those two apart, and they’re not the same business outcome.
So two concrete moves, and only two. First, write content that’s easy for the AI to cite — clear claims, real numbers, named evidence, the stuff a model can lift cleanly into an answer. Vague, padded copy ranks badly and cites worse. Second, start tracking, as its own metric separate from rank, how often your brand actually shows up in AI answers. You can’t manage the new exposure if rank is the only number on your dashboard.
And the caveat that keeps this sane: AI Mode is still 0.34%, so most of your traffic genuinely does still come from classic search. So don’t ditch the SEO that’s working — that would be trading a real channel for an emerging one, which is the opposite of the point. The move is re-centering, not switching. Keep doing the SEO, and gradually grow a second muscle next to it: content designed to get cited. Maybe I’ve got the timing wrong and the AI side stays small for longer than I think. But the direction the floor is tilting seems pretty clear, and it costs almost nothing to start watching the second number now.
Sources
- [R] Rand Fishkin (SparkToro) (2026), “In 2026, Less than One Third of Google Searches Still Send a Click”, SparkToro, SparkToro
- [R] Athena Chapekis, Samuel Bestvater, Emma Remy, Gonzalo Rivero (Pew Research Center) (2025), “Google users are less likely to click on links when an AI summary appears in the results”, Pew Research Center
- [R] Cloudflare (2025), “The crawl-to-click gap: Cloudflare data on AI bots, training, and referrals”, Cloudflare
- [R] Lambrecht, A., Padilla, N., Lam, H.T., Hollenbeck, B. (London Business School / UCLA) (2026), “The Impact of LLM Adoption on Online User Behavior”, London Business School
