"58%" gets quoted constantly in zero-click search content. It's actually at least three different numbers wearing the same figure. There's the share of all searches ending without a click, ~58.5%, SparkToro/Datos. There's the reduction in clicks specifically when an AI Overview appears for a query, a separate, conditional figure. And there's a modeled traffic-impact estimate from a consumer survey, 15–25%, Bain. Quoting any of them without its population is how the confusion started.
Headline numbers
of all US Google searches ended without any click, based on Datos clickstream panel data.
estimated reduction in organic web traffic attributed to zero-click behavior, from a consumer survey.
Three studies, one number
| What it actually measures | Source | Population |
|---|---|---|
| 58.5% — share of all US searches ending without any click | SparkToro / Datos, 2024 | All searches, clickstream panel |
| ~58% — reduction in clicks to a top-ranking page specifically when an AI Overview appears | Separate click-through analysis | Only queries that trigger an AI Overview |
| 15–25% — estimated organic traffic reduction attributed to zero-click behavior overall | Bain & Company, Dec 2024 | Survey of consumers, self-reported |
The SparkToro figure, in detail
Rand Fishkin's SparkToro, working with Datos clickstream panel data, found that 58.5% of US searches, 59.7% in the EU, resulted in zero clicks. No click to any result, organic or paid. This describes all Google searches, not specifically ones where an AI Overview appears. It's Traceable : a disclosed panel, a named methodology, a specific geography.
The Bain figure, in detail
Bain & Company's December 2024 consumer survey found that most consumers reported relying on "zero-click" results for a meaningful share of their searches. The firm modeled an estimated 15–25% reduction in organic web traffic as a consequence.
This is Traceable as a disclosed survey. But it's self-reported consumer perception, feeding a model. That's a different evidentiary category than a clickstream panel measuring actual behavior.
Three separate, legitimate studies produced numbers near '58%.' They measure completely different things. Quoting any of them without saying which is how a real statistic becomes a fake one.
Share on XThe AI-Overview-specific figure
A separate figure describes reduced click-through specifically when an AI Overview is present for a query. It's often quoted alongside the SparkToro number, as though they're the same measurement. They aren't. One is unconditional, across all searches. The other is conditional, on AI Overview presence.
The conditional figure is the one that gets misapplied to a site's whole traffic. That's the exact mistake documented in the provenance audit. See AI search CTR statistics for the fuller treatment of this specific figure.
Zero-click search predates AI by years
Worth stating plainly, since coverage sometimes implies otherwise. The zero-click phenomenon was already a majority pattern in Google search, well before AI Overviews existed. Featured snippets, knowledge panels, weather widgets, calculators, and unit conversions drove it. Any search interface piece that directly answers a query without a click counts.
AI Overviews are best understood as an accelerant on an already-large, already-growing trend, not the trend's origin. Treating "zero-click" as synonymous with "AI-caused" pins a multi-year pattern on a much more recent product.
What this means in practice
When quoting a zero-click statistic, always state the population. "Of all searches" and "of searches where an AI Overview appears" are not interchangeable. If you're estimating traffic risk for a specific site, a query-level tool checking which of your actual keywords trigger AI Overviews is far more useful than any of these aggregate figures. See the AI Overview Exposure Checker.
A worked example: quoting it correctly
Wrong: "58% of clicks are lost to AI." That sentence blends all three studies into one, and names no population at all. Right: "About 58.5% of all US Google searches end without a click, per SparkToro's 2024 clickstream panel. That figure predates AI Overviews and describes all search behavior, not AI-specific loss."
The second version takes one extra sentence to write. It also survives a fact-check, which the first version doesn't.
A quick checklist before you quote "58%"
- Which of the three studies above are you actually describing?
- Does your sentence state the population — all searches, or only AI-Overview-triggering ones?
- Are you applying a conditional figure to an unconditional claim (the exact error documented in §7)?
- Is the figure current, or has it since been updated or superseded by a newer measurement?
Verification status
Both headline figures trace to named, disclosed studies — Traceable . The confusion this page addresses isn't about either figure being wrong; it's about them being conflated with each other and with a third, distinct AI-Overview-specific click-reduction figure.
Namdev, R. (2026). Zero-click search statistics (v2). Retrieved from https://ritiknamdev.com/blog/zero-click-search-statistics Published under CC BY 4.0 — reuse freely with attribution.
This page is a direct application of the finding in where AI SEO statistics actually come from — read that piece for the full trace of how these three numbers collapsed into one in circulation.