Search Share Fell To 12% For North American Publishers: Chartbeat
Search now sends North American publishers only 12% of their traffic, so AI answers are clearly taking clicks that used to reach their pages.
On this page
A falling search share at publishers is a warning, not yet proof of lost traffic
Search Engine Journal reported on Chartbeat's third-quarter 2026 Audience Report. It puts search engines at 12% of pageviews at North American publishers in Chartbeat's network, against 21% in the equivalent 2025 report. The decline has been steady. The fourth-quarter 2025 report showed 16%, and the second-quarter 2026 report showed 13%. Nearly every region fell. Central Asia dropped from 53 percent to 33 percent. Africa was the exception, rising from 41 percent to 46 percent.
Does a lower search share mean publishers get less search traffic?
Not necessarily. Share is a ratio. It falls if search sends fewer visits, if other sources send more, or both, and the reports publish no absolute pageview counts to separate these.
Chartbeat's takeaway says overall traffic is steady, and the same sentence appeared in the fourth-quarter 2025 report. No figure backs it. If total traffic really were flat while search's share shrank, search would have lost visits in absolute terms. That is our inference, not Chartbeat's claim, and "steady" has no stated tolerance. Treat the chart as a pointer to where to look, not as a loss figure.
Is AI really the cause?
Chartbeat's earlier reports hedged that AI was behind the decline. The latest report gives no cause, and the data cannot isolate AI answers from any other shift in where readers come from.
The fourth-quarter 2025 report called the drop "likely a result of AI's growing impact on search". That is a plausible reading, but it is an opinion, not a measurement. Chartbeat also groups Google, Bing and every other search engine into one category, so the chart cannot show whether Google specifically lost ground. The reports do not say how many sites are in the sample or whether the same sites were measured each year. A changing panel could move the numbers without any change in reader behaviour. Chartbeat sells analytics to publishers, and the report closes with a demo pitch. That does not invalidate the data, but it does mean the framing is the vendor's.
How should a publisher use this benchmark?
Compare your own search share of pageviews with the network figure, then check your absolute search visits over the same period before drawing conclusions about visibility or AI.
If your share sits near the network's 12%, you are tracking the panel. If it is well above, you may be more exposed to search changes than your peers. If it is well below, the question is what replaced it. Either way, the share alone tells you nothing about the revenue effect.
The gap here is the measurement problem we raised in our earlier piece on AI search measurement challenges: a single percentage hides which sources moved. Ask your analytics vendor four things. What is in the sample? Is it the same set of sites each period? Can Google be separated from other search engines? Will they publish absolute counts alongside the shares? Without those, a share chart is hard to act on.
Frequently asked questions
How far has North American search share fallen?
From 21% in the 2025 third-quarter report to 12% in the 2026 one. The intermediate readings were 16% in the fourth-quarter 2025 report and 13% in the second-quarter 2026 report.
Which regions moved most?
Central Asia fell furthest, from 53 percent to 33 percent. Africa moved the other way, from 41 percent to 46 percent. The reports offer no explanation for either.
Can the report show Google lost traffic?
No. Chartbeat lumps all search engines into one category in these charts, so a Google-specific decline cannot be read from them.
Sources: Search Engine Journal, reporting on Chartbeat's Audience Reports.