If you follow search news at all, you have seen two numbers this year, and they are twenty-one points apart.
SparkToro published research in 2026 finding that 68.01% of U.S. Google searches ended without a click in the first four months of the year. That was up from 60.45% in 2024.
iPullRank published a separate analysis and put the figure at 46.96%. Theirs covered 10.2 billion searches between October 2024 and December 2025. The data came from Datos, which runs a clickstream panel, meaning a large group of people who have agreed to let their actual browsing be measured rather than being asked about it afterward.
Both studies are careful, and neither is wrong. They are measuring different populations. The SparkToro figure is U.S. only, and Rand Fishkin weighted mobile at two-thirds and desktop at one-third to match how people use Google. The iPullRank figure is a global average across every country in the panel. Once you know that, the gap mostly closes on its own: iPullRank’s own country breakdown puts the U.S. at 49% overall, and U.S. mobile alone at 66%.
Fishkin is direct about the limits of his own chart. The older comparison years come from different panels, including the now-defunct Jumpshot, and he calls the cross-year comparison “a bit of apples and oranges.” The direction is dependable. The exact slope between any two distant years is not.
The number is the part clients bring to me, and it is the least useful thing in either study.
The statistic that should change what you do
From the iPullRank analysis: only 14.0% of clicks go to a site the searcher named in the query. The other 86% go to a site the searcher did not ask for by name.
That runs against the way most small businesses think about search. If someone types your business name and clicks your website, that click was never in play for anyone else; it was navigation, and you had already won it before the search happened. The 86% is everything else. It is the person who typed a problem, or a service, or a question, and then chose among the options they were shown.
That is the traffic worth competing for, and it is also the traffic that a rising zero-click rate eats into first.
The clicks did not go to ads
This one surprised me, and it complicates the assumption that Google is converting organic traffic into ad revenue.
Over the fifteen-month window, iPullRank found that the rise in zero-click searches came almost entirely out of organic clicks. Organic click share fell about 2.8 points. Ad clicks showed no meaningful change. Users in the panel were roughly twenty-five times more likely to click an organic result than a paid one.
So the clicks that disappeared did not move to the ads. They stopped happening. The data shows the click was not made; it does not show why, and the obvious explanation is that the results page already carried the answer.
Where the surviving clicks land
When a click does happen, the six most common destinations in the iPullRank data are YouTube, Google’s own properties, Reddit, Facebook, Wikipedia, and ChatGPT.
None of those is an ordinary business website, and I want to be careful about what that does and doesn’t mean. A list of the six largest destinations on the internet tells you very little about the long tail where most businesses live. Reddit, Wikipedia, and YouTube absorb an enormous number of clicks in aggregate, but a plumber in Kelowna was never competing with Wikipedia. The list describes scale, not your market.
What it does say is that three of the six are places you can appear without owning the domain. Those three are YouTube, Reddit, and Facebook. If a share of the clicks that still happen is landing there, that is worth knowing before the whole budget goes into your own site.
The mobile split is the largest one in the data
In the U.S., 66% of mobile searches ended without a click, against 49% on desktop. The United Kingdom shows the same shape, at 65% and 46%. Indonesia shows a much narrower spread, 55% against 49%.
I think the mobile figure is the one to plan around. Most local intent arrives on a phone, and a phone is where the answer box, the map pack, and the AI summary take up the entire first screen. If your reporting still treats mobile and desktop as one line, you are averaging away a seventeen-point difference.
There is also an age pattern. Younger searchers zero-click more, and the gap between the youngest and oldest groups is about ten points. Older users click ads about seven times more often than people aged 18 to 24 do.
What I would do this quarter
I do not think the answer to any of this is to chase the zero-click number down. It is not a number you control.
I would start by separating branded from non-branded queries in Search Console and reporting them separately. If 14% of all clicks are navigational, then a site whose click total is holding up on the strength of its own brand name is not growing; it is being visited by people who already knew about it. That distinction is invisible in a single sitewide clicks line, and it changes what you do next.
Then I would look at impressions rather than clicks on informational pages. A page that gains impressions while losing clicks may not be failing at all; its answer may be getting read on the results page instead. I would measure those pages separately from the ones meant to produce a phone call or a booking, because the two are being asked to do different jobs.
The third thing is the one I would spend money on. Since the clicks that remain are overwhelmingly discovery clicks, the pages that earn them are the ones that answer a question a person asked without naming anybody. Those are service pages, comparison pages, and the specific local questions your customers ask on the phone. A page that only ranks when someone types your name is not doing this work.
I would not restructure a site around any of these percentages. The direction of travel is well established across every panel that has measured it, but the numbers themselves move by twenty points depending on who is counting and where.
