Attribution now sees only the 32% of Google searches that end in a click

Dynamic number insertion changes the phone number on a web page according to how the visitor reached it, so an inbound call can be traced back to the ad, campaign or search term that sent them ^[8]. The technique needs a site visit. Fisher writes that fewer people now click through to websites before calling, and that "they're arriving already sold on an AI summary they've read" ^[11]. Someone who read the answer on the results page and dialled from there never loads the page that would have swapped the number. So the method recovers calls from sessions analytics already logged, and the zero-click caller sits outside it.

Take the headline figure at face value. If 68% of Google searches in the first four months of 2026 ended without a click ^[1], then 32% produced one ^[13]. Click-based reporting prices that 32%. Fisher also puts the share of searchers who read the summaries at the top of results at 60%. He reports that about half of U.S. adults now use AI chatbots, up from a third in 2024, a rise of roughly 17 percentage points in two years ^[4][5][14]. None of what happens inside a chatbot thread reaches a company's website analytics, he writes ^[6].

Budgets follow credit, and credit follows traceability. Google Analytics and HubSpot make it straightforward to tie a site visit or a form submission back to the ad or email that prompted it ^[17]. The channels that still generate those traces keep accumulating attributed revenue, while demand created in a summary or a chatbot thread has nowhere to land. Spend then moves toward whatever still reports. Fisher writes that the implications "directly impact how businesses allocate spend and evaluate marketing's ROI" ^[12].

The article states the figure without naming the study behind it or splitting it by search intent ^[15]. Without that split there is no telling how much of the 68% was ever commercial, and a weather lookup that ends on the results page costs a business nothing. The number therefore bounds how much search behaviour is visible and cannot size the revenue that went missing.

Fisher is co-founder and CEO of CTM, formerly CallTrackingMetrics, and the piece ran under the Forbes Technology Council ^[2][16]. The diagnosis and the product line sit on the same P&L. The diagnosis can still be right. His claim about the past is the more durable one: "The click was never the whole journey; it was just a piece of the pie that your tools captured" ^[10]. Phone calls were an afterthought in most dashboards, when counted at all, long before AI overviews existed ^[7].

That leaves a narrow choice this quarter, which is whether to reallocate budget on reports that now describe under a third of search behaviour ^[13]. How much of the untracked path converts is not in this material. The sequencing consequence is clearer: cut a channel on click evidence and next quarter's click evidence will agree with the cut, because the traffic you stopped buying is the traffic the report could see. Fisher's recommendation is to pair dynamic number insertion with multi-touch attribution, which credits each step to its specific source instead of the last click before purchase ^[9].