18/11/2026
A zero-click search ends on the results page: the user gets their answer without visiting any site. For a marketing or SEO lead, the point is not to complain about it, but to know which pages are affected, what it really costs, and where to shift effort.
This page defines zero-click, describes what produces it, then gives a measurement method and four responses. It details a phenomenon that our page on the impact of AI on SEO places among the changes in search.
Zero-click: definition and origin
Zero-click (also written "zero click") refers to a search that leads to no site visit. The expression comes from the measurement of behaviour on search engines: across all searches, a large share closes without a click on a result.
What the term covers exactly
Three different situations are counted as searches without a click, and they do not mean the same thing for you:
- The answer is given on the page: a definition, opening hours, a calculation or a summary satisfies the request. This is zero-click in the strict sense.
- The user rewords: dissatisfied with the results, they run another search without having clicked.
- The user stays with the search engine: they open a business profile, a map, a video or a shopping result hosted by the engine itself.
Only the first takes away a visit you could have obtained. The second is ordinary search behaviour, and the third moves the interaction to a surface that you can, in some cases, feed. Before reasoning about an overall percentage, ask yourself which zero-click you are talking about.
An old phenomenon on a new scale
Zero-click was not born with artificial intelligence. Direct answers, entity panels and featured snippets have existed for more than ten years: the weather, a unit conversion or the date of an event never required a visit. What is changing is the range of questions the results page is able to answer.
The order of magnitude is now established: 60% of searches end without a click (Semrush, 2025). This measure covers all searches, across all subjects; it varies according to the datasets analysed and does not describe your sector. It says one useful thing: the results page has become a destination, and no longer just a place people pass through.
What produces a search without a click
Several elements of the results page answer in place of a site. Each affects different queries and leaves you a different amount of room:
Two lessons emerge. First, most of these elements cite or display a source: the visit disappears, not necessarily the presence. Second, a featured snippet itself only attracts a modest share of clicks: the click-through rate on featured snippets is 6% (SEO.com, 2026). Holding the placement is worth as much for visibility as for traffic.
Local searches illustrate this shift well. A user who checks opening hours, calls from the profile or starts a route has not visited your site, but they have interacted with your business. Counting that search as "lost" would be a misreading. The same goes for a brand query answered by an entity panel: the user wanted an address, a number or a confirmation, and they got it. The question to ask is therefore not "how many visits have I lost?", but "is the information displayed in my place accurate, and does it come from me?".
Zero-click and AI: what generated answers change
AI Overviews and conversational answers extend the same movement, with two differences in kind: they answer far more complex questions, and they write instead of extracting.
A change of scale
A direct answer dealt with a fact. A generated summary deals with an open question: comparing two approaches, explaining a method, summarising selection criteria. Queries that used to require several visits now find an answer on one screen. The effect can be measured: zero-click searches have multiplied by 2.5 since AI Overviews (State of AI Search, 2025).
The top-ranked position is no longer protected. When an AI Overview is present, the click-through rate of the first position is 2.6% (Squid Impact, 2025). A page can therefore keep its rank and lose most of its visits on a query, without any mistake on your part explaining it. How these overviews work and the criteria for being picked up are described in our article on AI Overviews and SEO.
From position to citation
These summaries display links to the pages that support them. Appearing there changes the situation: being cited as a source in an AI Overview brings +1.08% average click-through rate (Semrush, 2025). The gain is modest in visits, but it counts in another way: the brand is named in front of a reader who is forming an opinion.
The criterion for success therefore shifts. On the queries the summary covers, the objective is no longer only to be first, but to be the source the answer reuses. In conversational interfaces, where the list of links disappears from the screen, this logic becomes the only one: without a citation, there is no presence.
Which pages are exposed to zero-click
Not all pages run the same risk. Exposure depends on the intent the page serves, not on its quality.
The most exposed pages
Three types of page lose the most visits:
- Definitions: the answer fits in two sentences, and the results page provides them.
- Factual questions: a figure, a date, a threshold, a formula.
- First-level guides: a list of generic steps that the summary condenses without loss.
These pages do not become useless. They remain the material answers are made from, and the gateway to your subject for a search engine. But judging them on sessions alone leads to abandoning them wrongly. Deleting a definition because it no longer receives visits also means removing the page that earned the brand its citation on the subject.
The pages that hold up
Conversely, a visit keeps its value as soon as the user needs more than a summary:
- Detailed comparisons: criteria, assumptions and use cases that people want to check for themselves.
- Complete methods: steps with their conditions, their errors and their checks.
- Proprietary data: a study, a survey, a methodology.
- Decision pages: offer, demonstration, conditions, evidence.
- Tools: a simulator, a template, a file to download.
The steering rule is simple. If the intent is satisfied in three lines, expect fewer clicks and aim for citation. If it involves a decision, evidence or a calculation, aim for the visit and the conversion. Between the two, mixed intents call for a page built in two parts: a short answer at the top, picked up by search engines, then the detail that justifies the visit.
Measuring zero-click on your site
No report displays your "searches without a click". You deduce them from three readings taken together.
The signals in Search Console
The Performance report gives, by page and by query, impressions, clicks, click-through rate and position. Zero-click is spotted through a precise pattern: a stable position, impressions that are stable or rising, and a falling click-through rate. The page is still seen, but it is no longer visited.
Compare two periods of the same length, at a comparable position, and isolate the queries where the gap in click-through rate is greatest. Then segment: brand or non-brand queries, mobile or desktop, by type of page. At the time of writing, Search Console counts impressions and clicks from AI Overviews in this same report, without distinguishing them: the reading therefore remains a deduction. The method for reading this report is detailed in our guide to the Google Search Console Performance report.
The reading of the results page
The Search Console signal shows where to look; it does not say what is answering in your place. For the most affected queries, record the results page: which elements appear above the links, is your page the source displayed, which competitors are. Repeat the reading on mobile and on desktop, where the layout differs, and at regular intervals, because the presence of an AI Overview varies. Log these readings in a dated table, query by query: that is what allows you to say, three months later, whether a page has become the source displayed or whether a competitor has taken its place.
The link with conversions
A fall in clicks is only a problem if it affects your business. In your analytics tool, check what the pages concerned used to carry: direct conversions, assisted conversions, sign-ups. Many informational pages generate high traffic and little value; their loss of visits weighs less than it seems. Conversely, an informational page that used to start sales journeys deserves to be defended first. So sort the affected pages into two groups before any action: those whose loss of visits has no measurable effect on the business, to be monitored without investing, and those that used to feed your sales enquiries, to be dealt with first.
Responding to zero-click: four strategies
There is no setting that brings the clicks back. There are four responses, to be combined according to the type of page.
Become the source of the answer
On the queries the results page handles itself, aim to be reused. Open each section with a complete answer sentence, structure in lists and tables, date your updates, cite your sources. An answer that cites you keeps the brand in the reader's field of vision, even without a visit. Check the prerequisites too: a page that is not indexed, or whose snippets are forbidden by a directive, cannot be reused. And keep your public information up to date, because entity panels and business profiles feed on what the web says about you.
Give a reason to click
A summary condenses; it does not replace what cannot be condensed. Add to your exposed pages what a summary cannot carry: a decision matrix, a worked example with figures, edge cases, a checklist, a template to reuse. Announce that value in the title and the description, so that the reader knows what more the visit will bring. On the other hand, avoid holding information back to force the click: a page that does not answer clearly is not reused as a source, and loses on both counts. The test is simple: read the summary shown for your query, then your page. If the page says nothing the summary has not already said, there is no reason to visit; if it shows how to apply the answer to a real case, with numbers and conditions, the reason exists and only needs to be made visible.
Shift effort towards decision queries
Rebalance your output. Comparison, choice and implementation queries have lower volumes, but they keep their visits and weigh more on the business. Link each informational page to a decision page, so that the remaining visit has somewhere to go. In practice, for each subject you cover, check that you have a reference page, a comparison and an offer page, then direct new publications towards the missing link. An editorial calendar made up solely of definitions is the most exposed. This does not mean giving up informational content: it means giving each informational page a job in a journey, and checking that the next step exists.
Count visibility without a visit
An impression in an answer, a brand named in a summary, a call from a business profile: these are results. Add them to your measurement: presence in answers over a fixed set of questions, impressions, interactions on the search engine's surfaces. The indicators that complement clicks are defined in our article on GEO KPIs. This wider measurement is not a consolation: it describes the part of your visibility that influences a buyer before they visit a site, and that has so far been missing from your dashboards.
What zero-click changes in reporting
Reporting built on organic sessions alone describes reality less and less well. It signals a fall where visibility is holding, and pushes towards the wrong decisions: abandoning pages that feed answers, or rewriting pages that have lost nothing in ranking.
Three adjustments make reporting readable for senior management:
- Separate pages by role: reference pages, judged on their presence and their citations; decision pages, judged on visits and conversions.
- Present three levels: visibility (impressions, presence in answers), traffic (clicks, sessions), value (conversions, contribution to sales).
- Compare each page with itself, before and after, rather than with a site average.
Say clearly what will not come back: part of the visits on simple informational queries. And what is being aimed for instead: a brand present in answers, and traffic more concentrated on the pages that serve a decision. A gap observed after an action remains a correlation over time; present it as such.
Change the targets along with the indicators. A team still judged on organic sessions alone has every reason to keep producing the simple informational pages that generate volume and that summaries replace. Set one target for presence in answers on a fixed set of questions, another for visits to decision pages, and a third for the conversions those visits produce. Review the three together, at the same meeting, so that a fall in one is read in the light of the other two rather than in isolation.
FAQ on zero-click
What is zero-click?
Zero-click refers to a search that ends without a site visit: the user finds their answer on the results page, rewords their request or stays on one of the search engine's own surfaces. People also speak of zero-click search or no-click search.
What share of searches is zero-click?
According to Semrush (2025), 60% of searches end without a click. This figure covers all searches and varies according to the datasets analysed. It does not describe your sector: measure your own exposure, page by page, in Search Console.
Does zero-click mean the end of SEO?
No. The answers displayed rely on pages that are indexed and well ranked: search optimisation remains the condition for being their source. What changes is the objective: being reused and cited on informational queries, winning the visit on decision queries.
How do I measure zero-click on my site?
In Search Console, find the pages whose position is stable and whose click-through rate is falling, then record the results page for the queries concerned to see what is answering in your place. Finally, link those pages to their conversions, to measure what is really at stake.
Which queries are most affected by zero-click?
Simple informational queries: definitions, facts, generic steps, along with local and brand searches, which are handled by profiles and panels. Comparison, choice and implementation queries are less affected, because a summary is not enough to decide.
How should you react to a fall in clicks caused by zero-click?
Combine four responses: make your pages citable so as to be the source displayed, add what a summary does not condense, shift effort towards decision queries, and include visibility without a visit in your measurement. None of them brings back all the lost clicks.
Continue reading
Depending on your situation:
- Your traffic is falling and you do not know why: if you have not yet distinguished a loss of ranking from a loss of click-through rate, start with the full diagnosis → SEO traffic drop.
- Your users are moving to conversational search: if the list of results gives way to a written answer, only a citation makes the brand visible → Google AI Mode.
- You have physical premises: if part of your searches ends on a profile, a map or a call, visibility is worked on away from the site → local search optimisation.
- You need to back up your analysis: if you are looking for market benchmarks on click-through rates and search behaviour → SEO statistics.
- You want to track visibility, traffic and value in one place: the Incremys performance reporting module brings together positions, traffic from AI engines and citations per question, and compares each page with itself before and after an update.
.png)
.jpeg)

.jpeg)
%2520-%2520blue.jpeg)
.avif)