What Zero-Click Does and Does Not Mean

On Similarweb's measure, zero-click share for news queries rose from 56% to 69% within a year of AI summaries launching.

Pew Research Center's behavioural study (900 US adults, 68,879 searches): clicks on traditional results ran at 8% with an AI summary versus 15% without, and clicks on citations inside the summary at 1%.

⚠️ Category differences are large. The news figure does not transfer to B2B decision queries, where buyers still tend to visit the site to check details.

First, put branded search volume and direct traffic on the report. Someone who sees the brand in an answer and then searches its name or types the URL carries no AI referrer.

Second, accept that the impressions-to-clicks ratio trends down. That is not a content-quality signal and should be discounted before drawing conclusions.

Third, stop presenting AI referrals as a multiple. Cross-site studies average around 1% of traffic, and multiples off a tiny base persuade no one — use absolute numbers plus share of enquiries.

Being named has value on its own: discovery can complete in the answer, with the visit arriving later via branded search or direct traffic. So content goals should include getting the AI's description right, not only pulling people through.

That does not license neglecting the landing experience. Once someone does arrive, the page has to answer the specific question that brought them. Getting the first three layers right and dropping this one is a common, expensive waste.