What to do when most searches end without a click

TL;DR SparkToro and Datos clickstream panel data covering September 2022 through May 2024 shows that 58.5 percent of US Google searches ended without any click. Only 360 of every 1,000 US searches reached a non-Google-owned property on the open web….

Flat clicks chart, AI summary panel with citation strip, and the brand impression rosette earned.
Flat clicks chart, AI summary panel with citation strip, and the brand impression rosette earned.

TL;DR

  • SparkToro and Datos clickstream panel data covering September 2022 through May 2024 shows that 58.5 percent of US Google searches ended without any click. Only 360 of every 1,000 US searches reached a non-Google-owned property on the open web.
  • Pew Research tracked 68,879 Google searches by 900 US adults in March 2025 and found that only 1 percent of AI summary-present searches generated a click on a source cited inside the summary.
  • A zero-click visit is not a zero-value visit when the AI summary cites your page by name. The brand impression fires inside the summary and surfaces later as a branded search.
  • Three query types still draw the click reliably — brand-name, transactional, and niche long-tail queries the engine cannot synthesise confidently. The broad informational long-tail is where the click rate fell furthest.
  • Two free changes this quarter: rewrite the first 80 words after every H2 to land an extractable answer with a number, a year, and a named source. Re-run target queries once a month from an incognito window and read the AI summary.

SparkToro and Datos found that 58.5 percent of US Google searches ended without a click in their 2024 clickstream study of tens of millions of US and EU users from September 2022 through May 2024.

Pew Research tracked the same behaviour from a different angle and found that only 1 percent of AI summary-present searches produced a click on a source cited inside the summary.

Both numbers describe the same gravity. A reader searches, reads the answer on the search page, closes the tab, and the visit never happens.

If you run a small site whose traffic depends on informational queries — the explainers, the definitions, the how-to articles — you have felt this gravity in your Search Console chart. Impressions hold steady or even rise. Clicks slide down.

The rank line does not explain the gap. What explains the gap is that the user got their answer before they reached you.

What does ‘zero-click search’ actually mean in 2026?

A zero-click search is a Google search session that ends without the user clicking through to any external website.

The user reads the AI summary, the Featured Snippet, the Knowledge Panel, or the People Also Ask box, learns what they needed to learn, and closes the tab. The session counted as a search and produced no visit. Across the SparkToro and Datos panel data covering September 2022 through May 2024, 58.5 percent of US Google searches ended this way.

The number is not a forecast or a hot take. It is behavioural panel data from tens of millions of devices over twenty months.

The same study also found that for every 1,000 US Google searches, only 360 clicks reached a non-Google-owned property on the open web. Roughly 30 percent of all clicks went to Google-owned platforms — YouTube, Maps, Images, Flights. The open-web slice that small sites compete for is a third of a third of all activity. Knowing the size of the pie changes how you spend your hour.

Why is the zero-click rate climbing on small-site queries?

Two engines drive the climb, and both run on small-site informational territory.

The first is the expansion of AI Overviews on informational queries — explainers, definitions, how-to questions, the heart of the small-site corpus. The second is the spread of Featured Snippets, People Also Ask boxes, Local Packs, and other in-SERP answer formats that already covered the same answer-shaped territory.

A small site that lives on informational long-tail queries feels the climb more than a brand-name site does.

The brand-name query still resolves the same way it did in 2018. A buyer types your name, a buyer reaches you, no summary stands between. The informational query is the one that lost half its clicks, because that is the territory the AI summary covers most aggressively. If your traffic mix is 80 percent informational and 20 percent brand, you absorbed more of the loss than a site whose mix is reversed.

What is the difference between zero-click and zero-value?

A zero-click visit can still carry value as a brand impression.

When the AI summary cites your page as one of its sources, the user reads your name and your one-sentence answer inside the summary even when they never click through. The brand impression fires.

The reader closes the tab. Three days later, the same reader types your brand name into the search box. That second visit lands in your Search Console as a branded search, not as a search-traffic visit from the original informational query.

This is the zero-click conversion pattern.

Bain Dynata’s December 2024 consumer survey of 1,117 US adults documented it under that name — a measurable share of buyers who learn a brand inside an AI summary and arrive at the site through a brand query weeks later. The first visit looks lost in your analytics. The second visit looks unattributable to anything specific. What actually happened is that the citation economy carried the buyer across the gap, and your Search Console shows only the second half of the story.

So the right question for a small site is not "how do I get the click back?" The right question is "is my page being cited inside the answers that have replaced the click?"

Which queries on a small site are still worth chasing for the click?

Three query types still draw the click reliably even in 2026.

Brand-name queries, because the user came to find you and no summary can interrupt that intent. Transactional queries — buy, hire, book, contact, near-me — because the user came to choose a destination, not to read an answer.

Niche long-tail queries the engine cannot synthesise confidently from a generic source still default to the traditional ranked list, where the click goes to the best-matching page.

Those three keep the click economy alive for a small site even as the broad informational click rate falls. They are also where most of your business actually lives — people who already know your name, people ready to act, and people with the specific question only you have written about deeply. A small site that prioritises those three query types in its content map preserves the click base while the rank-as-floor reframe handles the informational ground.

How should you rewrite informational content for the no-click era?

Write the answer block first, the supporting story second.

The first 50 to 90 words after every heading carry the extractable answer the AI summary will quote. A specific number. A year. A named source.

A clear single-sentence claim a chunker can lift out of the page and quote inside its synthesised answer. The rest of the section earns the click from the small share of readers who scroll past the summary and want the depth behind the headline answer.

The structure is not a trick. It is honest in the same way an executive summary is honest.

You name the answer, you cite the source, and you give the reader who needs depth the depth. The chunker that builds the AI summary lifts the first paragraph cleanly because it reads as an answer.

The reader who clicks through reads the same answer with the story underneath it. Both audiences get served by the same paragraph. There is no separate version of the page for the AI and a different version for the reader.

If you have ever wondered how to structure content for featured snippets, the answer-first paragraph is the same structure — it was the right shape for Featured Snippets before AI Overviews existed, and it stayed the right shape afterwards.

How do you measure progress when the click line has gone flat?

Three measurements together replace the old click-only dashboard.

Impressions on the target query, read directly from Search Console. AI citation presence on the target query, read directly from the AI summary by running the query yourself once a month. Branded-search volume in Search Console over a 90-day window, which is where the zero-click conversions surface six to twelve weeks after the brand impression fires.

You write these three numbers in a small text file at the start of each quarter.

You re-read the file at the end of the quarter. The numbers either trend the right way together or they do not. If impressions and citation presence rose and branded search lagged, the gravity is correct and the lag is real — branded search is the slowest of the three signals because it requires a buyer to remember you and act later. If impressions rose but citation presence did not, the answer block is not extractable enough yet, and the next pass is rewriting the first 80 words under each H2.

The dashboard is a single sheet of paper. The work it tracks is six months long.

Other questions worth answering

How does mobile behaviour on Google compare to desktop?

77.2 percent of mobile Google sessions end without any click. Desktop sits at 46.5 percent. SparkToro and Datos clickstream panel data from September 2022 through May 2024 produced both numbers.

Phone screens fit one answer block above the fold. Readers rarely scroll past that first answer.

How often does an AI Overview show up when the prompt is longer versus shorter?

AI Overviews showed up on 8 percent of one or two word prompts. They showed up on 53 percent of prompts ten words or longer. Question-shaped prompts triggered an AI Overview 60 percent of the time.

Pew Research Center tracked this in their 2025 measurement of 68,879 Google queries by 900 US adults. The longer and more question-shaped the input, the more often the engine answers in place.

Where do readers end up after an AI summary fully resolves their intent?

Often nowhere. Pew Research Center tracked 68,879 Google queries by 900 US adults during 2025. They found about 26 percent of AI-summary-present sessions ended with the reader closing the browser.

Sessions without an AI summary ended at 16 percent. That is a 63 percent jump in pure abandonment.

Why do the few visitors who do arrive tend to act on what they read?

Because the AI summary filters them before they arrive. It surfaces your one-sentence answer only when their problem matches your page. The reader who clicks through already understands what you do.

Pre-qualification sharpens the audience. SparkToro and Datos’s 2024 panel data found only 360 of every 1,000 US Google queries reach an open-web page. So each arrival represents stronger intent than the older click-rich era did.

What should you change on your site this quarter?

Two changes, both free, both finishable in a quarter on a five-page site.

Rewrite the first 80 words after every H2 to land an extractable answer with a number, a year, and a named source. Re-run the target queries once a month from an incognito window and read the AI summary to see whether the citation set has shifted — whether you are in it, whether a competitor entered, whether the synthesis itself has changed shape. Those two habits compound into the brand impression Search Console cannot yet show you.

If you read this and still feel that flat click line in your stomach, that is normal. Most small site owners are looking at the same flat line right now and reading it through a 2018 lens. You can contact me here and we can open Search Console together for one or two pages, run the citation-presence check live, and read what the numbers actually say.

There is no charge for the call, and there is no pitch at the end. I will tell you which pages already earn citations you cannot see in the dashboard, which ones need an answer-block rewrite, and the one or two queries where a brand impression is forming even though the clicks have stayed flat.

Similar Posts