Zero-Click Is the Best Thing to Happen to Paid Search in a Decade

The premise everyone gets backwards

The standard reaction to zero-click search is grief. In the first four months of 2026, 68.01% of Google searches ended without a click, and a decade ago that number was closer to 45%. The SEO industry read that as a eulogy.

For advertisers, it’s closer to a windfall.

The clicks that disappeared were never the valuable ones. They were the “what temperature to cook chicken” clicks, the “how many ounces in a cup” clicks, the top-of-funnel research clicks that landed on a blog post and bounced. Google didn’t take away demand. It took away the unqualified traffic layer sitting on top of demand — and in doing so, it concentrated commercial intent into a smaller, denser, far more purchasable pool.

The data backs this up in the least ambiguous way possible: an analysis of SERP click distribution across four major U.S. verticals from January 2025 to January 2026 found paid results were the single biggest winner in every category. In headphones, text ads and PLAs combined went from 16% to 36% of all clicks in one year. In jeans, paid went from 18% to 34%. Organic click share fell in parallel.

Same page. Same users. Paid doubled its share while organic collapsed. That’s not a coincidence — that’s what happens when the informational queries stop generating clicks and the transactional ones don’t.

Why intent beats keywords

The old model was crude and we all pretended otherwise. A user typed a string. We matched that string to a keyword we’d bid on. We inferred intent from a two-to-four word fragment and hoped.

The AI surfaces break that fragment open. Rather than responding to a single query, AI Mode evaluates the full conversational context, so ad placements are matched not just to keywords but to intent as it evolves over the session.

Think about what that actually means operationally. A user who starts at “best espresso machine under $500,” follows up with “does it need a separate grinder,” then asks “which ones are easiest to clean” has told the system more in three turns than a keyword list could capture in three hundred rows. The ad served at turn three is being placed against a fully articulated need, not a guess.

The formats follow the signal. Google introduced four Gemini-powered ad formats at Marketing Live 2026 designed to live inside AI Mode responses rather than alongside them. When AI Mode returns a list-style response, eligible ads can appear as a recommendation inside that list, with position depending on relevance rather than bid alone. For high-consideration purchases, Gemini pulls the product feed and writes a custom explainer on the fly, explaining why this product fits this shopper’s specific needs — creative generated per query, not per campaign. One format replaces the static lead form with a Gemini-powered chat agent trained on the advertiser’s website, letting users qualify themselves before ever leaving the results page.

That last one is the tell. Ad-to-qualification with no landing page in between is a shorter funnel than anything performance marketers have had access to before. The click was always friction. Removing it isn’t a loss — it’s the point.

And the ad relevance argument isn’t just spin: the ad isn’t interrupting the AI response; it’s selected for display because it’s relevant to the question the user asked. An ad that appears because the system understands the need is a better ad than one that appears because someone bid on a broad match term.

The honest downside list

Now the part most zero-click cheerleading skips.

Query visibility is degrading, and that’s a real loss. In May 2026, Google updated its help documentation to clarify that search terms shown in reporting for AI-powered surfaces may not reflect a user’s exact query — some reported terms represent Google’s interpretation of user intent instead. The change applies to AI Mode, AI Overviews, Lens, and autocomplete. That matters for negative keyword building, compliance review, budget decisions, and stakeholder reporting, and advertisers in regulated industries should be especially careful using search terms reports as evidence for brand safety or legal review.

If you can’t see the literal query, you can’t fully control what you’re paying for. Negative keyword lists — the primary defensive tool in search — get blunter.

You can’t target the placements you’re being served on. Existing Search, Shopping, and Performance Max campaigns can already become eligible for AI Overview placements, yet advertisers still cannot target those placements directly or pull clean segmented reporting. You’re in the auction whether you planned for it or not, and you can’t isolate the performance to judge it.

Eligibility is a compliance gate on automation. Eligible campaign types for the AI Mode ad test include Performance Max, AI Max with search term matching, Shopping, and broad match campaigns. Translation: the price of admission to the fastest-growing inventory is surrendering match-type control and manual bidding. If your account is built on exact match and manual CPC because that’s what your margins require, you’re structurally excluded from the surfaces Google is expanding into. That is a business decision Google is making on your behalf.

CPC pressure concentrates. As informational traffic compresses, advertisers reallocate budget toward bottom-funnel, revenue-dense queries — which intensifies competition and raises bid pressure in exactly those auctions. The efficiency gain from losing junk traffic gets partially eaten by everyone else chasing the same shrinking set of high-intent terms. Margin-sensitive verticals feel this first.

Attribution gets harder before it gets easier. Intent resolving on the SERP means the measurable click event is thinner. Offline conversion imports and CRM-to-ad-platform integrations are the tools that close the attribution gap AI Mode reporting is likely to widen — which is a polite way of saying the platform’s own numbers will be less complete and you’ll need your own data infrastructure to compensate.

The SEO side is genuinely damaged, and PPC inherits the problem. Open-web clicks fell from roughly 374 to 276 per 1,000 searches between 2024 and 2026 — and the shift isn’t only AI. Searches leading to another Google search rose 7.2 percentage points, meaning Google is getting better at persuading people to refine and re-query inside the results rather than click out. In AI Mode specifically, 92–94% of sessions currently end without a click to an external site.

That has second-order consequences for paid. Organic used to subsidize paid: content earned the top-funnel visibility, remarketing audiences got built off that traffic, and paid harvested the bottom. If organic can’t fill the top of the funnel, paid has to buy awareness it used to get free — which means budget that was allocated to conversion now has to cover demand generation too.

Creative control and disclosure risk. When Gemini writes the explainer per query, brand messaging is being generated rather than approved. Separately, Google’s AI ad disclosure labels launched July 9, 2026 across Search, YouTube, and Discover, with ads built using Google’s own generative tools labeled automatically and third-party AI creative relying on an advertiser self-attestation Google says it won’t verify — and Google’s own documentation states the label setting doesn’t guarantee compliance with specific regulations. EU AI Act Article 50 transparency obligations became enforceable August 2, 2026, with penalties up to €15 million or 3% of global turnover. The liability didn’t move to Google. It stayed with you.

One caveat on all the numbers above. Published zero-click and AI Overview coverage figures vary widely by methodology — 68% versus 80%, AI Overviews on 20% of queries versus 48% — depending on panel, device mix, and whether the Google app is included. Treat every headline stat in this space as directional, not precise.

The verdict

Zero-click is good for advertising and bad for the free half of search marketing. Those aren’t in tension; they’re the same mechanism viewed from two sides. Google removed the unpaid path to visibility and replaced it with a paid one that reads intent more accurately than any keyword-matching system ever did.

The trade is real: better targeting in exchange for less transparency, less control, and more dependence on automation you can’t audit. Whether that’s a good deal depends entirely on whether you were making money from the clicks that disappeared — and on how much you trust a system that both places the ad and grades the placement.