The trend, with numbers
Rand Fishkin's analysis of Similarweb clickstream data put US zero-click searches at 68.01% for January to April 2026, up from 60.45% in 2024. A decade earlier the figure was around 45%.
AI Overviews accelerated this rather than causing it. Knowledge panels, featured snippets, and weather and calculator widgets had been absorbing clicks for years before generative answers arrived.
| Period | Share ending without a click | Source |
|---|---|---|
| 2016 | around 45% | SparkToro |
| 2019 | 49% | SparkToro |
| 2024 | 60.45% | SparkToro and Similarweb |
| January to April 2026 | 68.01% | SparkToro and Similarweb |
Why this breaks the usual reporting
Every traditional search metric assumes a click. When two thirds of searches never produce one, session counts stop describing search performance, and a flat traffic line can hide either real decline or real growth in visibility.
The practical fix is to report exposure and outcome separately. Impressions and citation share measure whether you are present. Self-reported attribution and pipeline measure whether presence is worth anything. Conflating them is how teams end up reporting growth that never reaches revenue.
What still earns the click
- Queries where the answer is genuinely too long or too conditional to summarise
- Anything needing a tool, a calculator, a template, or a download
- Comparison and pricing research, where buyers want to verify the summary themselves
- Original data, which a summary can cite but cannot replace
- Branded queries, where the person already decided where they are going
Why it matters for B2B marketing teams
Two thirds of searches sending no click breaks the only metric most B2B marketing teams report on. The fix is not despair, it is a second number. Track citation share and self-reported attribution alongside sessions, so falling traffic and falling visibility can be told apart.