ChatGPT tops Google's paid-click share at 4.75%, and growth teams should reprice the auction

iPullRank's Q3 2026 zero-click and paid-click analysis found that ChatGPT had the highest paid-click share of any leading Google destination, with about 4.75% of Google traffic landing on ChatGPT arriving through paid results ^[1][2]. That is a small percentage attached to a large structural fact: according to iPullRank, Google is functioning as a paid acquisition and discovery channel for an AI destination, not only as the thing the AI destination is supposed to be replacing ^[9].

The comparison is what makes the number worth reading. ChatGPT's paid share sat well above the corresponding shares for destinations such as YouTube, Wikipedia, and Amazon ^[3], and ChatGPT was not the biggest destination in the study: it ranked around sixth by total Google clicks, behind YouTube, Google's own pages, Reddit, Facebook, and Wikipedia ^[5]. So the ranking and the paid share point in opposite directions. The established destinations pull enormous volume that is overwhelmingly not paid; the newer one shows up smaller and with roughly one in twenty-one of its observed Google arrivals coming through a paid placement ^[13]. Put the other way, about 95.25% of those arrivals were still unpaid ^[15], which is the honest scale of the effect.

The analysis covered roughly 200 million events ^[4], which is enough to make the relative comparison interesting and not enough to make it a map of AI discovery. iPullRank is explicit that the figure does not reveal OpenAI's advertising budget, bids, or total advertising activity ^[6], and that the metric counts paid Google clicks that land on ChatGPT rather than every path a user takes to the service, so it should be read as a traffic-distribution signal ^[7]. That boundary matters for anyone tempted to convert 4.75% into a spend estimate. It is a share of arrivals at a destination, measured at the landing end, and the dataset also does not speak to conversion quality or the long-term profitability of paid acquisition ^[12].

The panel is also predominantly desktop, at about 94.5% ^[8], which leaves roughly 5.5% of it everywhere else ^[14]. Mobile behavior, regional results, and campaign targeting can all move the relative contribution of paid clicks ^[8]. If your category is mobile-first, treat this as directional rather than as your baseline.

The operational consequence is a measurement one first. iPullRank's argument is that tracking only organic rankings or direct traffic can miss a material part of how people discover AI products ^[10], and that a visit arriving through a paid Google result carries different attribution, cost, and optimization implications from a direct visit or an organic referral ^[11]. For a growth team, that means the competitor set on a task-shaped query now plausibly includes an assistant, and the useful next step is to separate paid, organic, and direct discovery in your own reporting before deciding whether paid is complementing organic demand or substituting for it ^[11].

What to watch: whether the paid share holds or compresses in the next quarterly cut, and whether it survives a less desktop-heavy panel ^[8]. Watch also for the same pattern appearing at other assistants; a single destination with an elevated paid share is a company's choice, while several would be a change in how the category buys demand. And keep the boundary in mind when someone quotes the number at you: it describes arrivals at ChatGPT from paid Google clicks, not what OpenAI spends ^[6][7].