Written by

Pragadeesh NatarajanContent Marketing Specialist
What happened

A revised working paper by Saharsh Agarwal of the Indian School of Business and Ananya Sen of Carnegie Mellon University, the first randomized field experiment on the question, found that AI Overviews reduce outbound organic clicks by 39.8% and increase zero-click searches by 34.5%. Sponsored clicks and overall search frequency did not change. The researchers used a browser extension to show more than 1,000 Google users search results with or without AI Overviews, so the effect could be separated from everything else changing in search.

The paper also tested Google's explanation for falling traffic, that AI Overviews mainly remove low-value visits. It found no difference in what happened after a click: about 4 in 10 same-tab clicks led back to the results in both groups, and time on site was statistically indistinguishable, as Search Engine Journal reported. The authors describe that as at odds with the view that AI Overviews mostly eliminate low-engagement visits. The paper has not completed peer review.

Why it matters for your business

This is the number to build a plan against, and the asymmetry inside it is more useful than the headline. Organic clicks fell by two fifths while paid clicks did not move at all, which means the surface is not shrinking, it is reallocating: the same attention is being served inside the answer or sold.

If you are defending an organic traffic line, that reframes the job. Recovering the lost 39.8% by ranking harder is not available, because the clicks are not going to a competitor's blue link, they are not happening. What is available is being the source named inside the overview, and making the part of your page a summary cannot reproduce obvious enough to be worth a visit.

Set the expectation internally now, before someone reads a flat traffic chart as a content failure. And if your board tracks sessions as the visibility metric, this study is the argument for changing the metric.

Sources