Zero-Click Search Strategy 2026: What Still Earns Value

The first version of this zero-click search strategy guide went up earlier in 2026, and a lot has moved since. The headline number changed, Google shipped a report that measures AI visibility directly, and Search Console has handed me unflattering data about this very page. All three are in this update, along with corrections to two figures the earlier version got wrong.

Verification note. Every statistic below was traced back to its original source on 13 October 2026, and the page's own Search Console data runs to 10 October.

What zero-click search means in 2026

A zero-click search is one that resolves on the results page. Someone types a question, Google answers it through an AI Overview, a featured snippet, a knowledge panel or a direct answer, and the session ends. There is no visit and no row in your analytics.

The current figure comes from SparkToro's analysis of Similarweb clickstream data. Between January and April 2026, 68.01% of US Google searches ended without a click to any destination [1]. In 2024 the figure was 60.45%, a rise of about 7.5 percentage points in two years.

Three caveats matter before you apply that number to your own site:

• The two years use different panels. The 2024 figure came from Datos, now part of Semrush, and the 2026 figure from Similarweb. SparkToro says plainly that the comparison is "a bit of apples and oranges", because the users and devices are not the same [1].
• The Google mobile app is excluded. The data covers searches in a browser. SparkToro expects zero-click behaviour inside the app to be even higher [1].
• 68% is a blended figure. It assumes searches are two-thirds mobile and one-third desktop [1]. If your readers are mostly on phones, the rate they live with may be higher than the average.

These are US figures. Search from Nepal, India or any market where AI Overviews cover fewer query types and you will see fewer AI answers for the same English query. Check your own results pages before assuming the US average describes them.

How the number reached 68%

AI Overviews are the steepest part of the curve, not the whole curve.

Featured snippets arrived in 2014, and knowledge panels earlier still. By 2016 roughly 45% of Google searches already ended without a click, and SparkToro's first study in 2019 put it at 49% [1]. The 2024 baseline of 60.45% came just before AI Overviews spread widely through 2024 and 2025. SparkToro calls the move to 68% the fastest acceleration it has seen in this metric [1].

Google AI Mode, the conversational search tab, has barely touched these numbers yet. Similarweb's panel put it at 0.34% of Google searches in January to April 2026 [4]. That is despite Google reporting more than a billion monthly AI Mode users at I/O 2026. When AI Mode sessions do happen they rarely leave Google, so that phase has not really started.

The tactics that worked against featured snippets in 2019 remain a reasonable starting point for a zero-click search strategy now. Both reward a page that states a clear answer a machine can lift. What is new is scale, and the fact that the answer is assembled from several sources rather than quoted from one.

Where clicks still happen

Not every search stopped clicking, and two sets of evidence agree on where the remaining clicks go.

Cyrus Shepard studied 400 websites whose traffic held or grew through the 2024 to 2026 decline, and Rand Fishkin discussed the findings on SparkToro [5]. Five features separated the winners: they offer a product or service, let the visitor complete a task, own proprietary assets, keep a tight topical focus, and carry a recognisable brand. Task completion was the sharpest divider. 83% of winning sites let visitors finish what they came to do, whether checking a lottery ticket, running a calculator or booking a slot. Of sites with proprietary assets, 92% went up. Tight topical focus predicted little on its own: 75% of winning sites had it, but so did 61% of losing ones.

The second line is query type. Branded searches make up roughly 44% of Google searches and still send clicks reliably, as do local and transactional queries [6]. Zero-click behaviour concentrates on informational queries, which is exactly what an AI Overview is built to answer.

For a small publisher this is uncomfortable. A page that explains a concept competes with the AI Overview on its own ground. A page that lets someone do something, compare real data or download a working template competes where the Overview cannot follow. This article is the first kind; the checklist near the end is my attempt to make it a little more of the second.

Why a zero-click search strategy is the second problem, not the first

Most content reporting still reduces success to one number, organic clicks, and that number now misleads in both directions. The fuller argument is in how to prove content ROI without click data.

Seer Interactive tracked 3,119 informational queries across 42 organisations from June 2024 to September 2025. On queries with an AI Overview, organic CTR fell 61%, from 1.76% to 0.61%. Brands cited inside the AI Overview earned 35% more organic clicks and 91% more paid clicks than uncited brands on the same queries [7]. Seer is careful to say it cannot prove that citation causes the higher CTR; stronger brands may simply be cited more often. The citation is still worth tracking, and a click count will never show it.

Pew's finding sharpens the point: only 1% of visits to a results page with an AI summary included a click on a link inside it [2]. Most of the value of being cited happens in the reader's head, before any click and often without one.

Before blaming AI Overviews, check your own ranking. I pulled Search Console for anobee.com in July 2026. The site-wide average position over the previous three months was 59.9, with a 0.8% click-through rate on 5.16K impressions. That is not a zero-click story. At page six, low clicks are explained by weak ranking, not by AI Overviews absorbing them.

This page is the sharper example. Between 21 June and 14 September 2026, the earlier version recorded 77 impressions, no clicks and an average position of 29.6. In the three months to 10 October it recorded 276 impressions across its two URL versions and still no clicks, but the canonical URL's average position improved to 20.6. The queries that surface it are still mostly the wrong ones; "zero click search strategy" itself brought one impression at position 19. The top query is the quoted search "datos" "clickstream", where this page sits at position 1.7. It is a search about Datos, the clickstream company behind SparkToro's 2024 data. My earlier version of this article wrongly described it as a Spanish-language query.

A page that ranks for the wrong questions is not losing to AI Overviews. It has not yet earned the right to be summarised. That is why this refresh starts with intent and ranking, and why the zero-click search strategy below treats citation as a second-stage goal.

The ANCHOR framework for a zero-click search strategy

Optimising harder for clicks that are structurally not coming does not work. What does work is a set of habits that earn citation and recall while ranking work continues. I call it ANCHOR: Answer-first structure, Named authorship, Corroborated claims, Homepage-level entity consistency, Original data and Repeat brand exposure.

A framing note first. Google's guidance for its generative AI features says SEO best practices remain relevant because AI Overviews and AI Mode are rooted in its core ranking systems. It also lists things you do not need: llms.txt files, special markup, content chunking, or rewriting for AI systems [8]. ANCHOR is not a workaround for that guidance. It is ordinary good publishing, done in an order that suits how answers are now assembled. The difference between SEO, AEO and GEO is covered separately.

A: Answer-first structure

Open every major section with a direct, self-contained answer to the question in its heading, in the first sentence or two, then earn the detail. Retrieval systems lift these passages. Bury the answer three paragraphs down and a competitor's cleaner passage gets used instead, even when your page is deeper. This is not chunking: a full article with clear openings reads better for people too.

N: Named authorship

Google's people-first content guidance asks whether it is "self-evident to your visitors who authored your content" [9]. It also asks whether pages carry a byline where one is expected, and whether that byline leads to background about the author. A bylined page from a verifiable person is easier to place and trust than an anonymous one. Every Anobee article carries a byline linked to an author page. The work is keeping that consistent and pairing it with the tactics that make a page citable by ChatGPT.

C: Corroborated claims

When the same fact about your brand or data appears independently across your own site, third-party mentions and community discussion, it is more likely to be surfaced. Consistency across sources does what confidence on one page cannot. Google draws a firm line at inauthentic mentions [8]. The version that works is the slow one: publish something specific enough that other people repeat it with your name attached. The same principle drives entity optimisation for AI search.

H: Homepage-level entity consistency

Your brand name, author name and core facts should match exactly across your site, About page, social profiles and any third-party listings. Anobee has a small but real problem here. The bare word "Anobee" collides with unrelated personal accounts on several platforms, and with Amobee, an ad-tech company one letter away. So the site anchors every entity signal to the domain and the author name together. The earlier version of this article flagged a broken url field in the site's Organization schema. It has since been fixed, and the structured data guide for bloggers shows how to check yours.

O: Original data

Primary data, whether a benchmark, a survey or an observed result, gets cited far more than commentary that repeats someone else's numbers. Google's guide says creating content people find "unique, compelling, and useful" is likely to matter more for generative AI search than anything else it suggests [8]. The Search Console figures two sections up are original data. They are small and unflattering, and they are the most citable thing on this page because nobody else has them.

R: Repeat brand exposure

Zero-click value builds up over time rather than landing in one shot. A reader who sees your name inside three different AI answers over a month is more likely to search for you directly later. That branded-search lift is one of the few measurable proxies a small site has for zero-click influence. It sits at the centre of the measurement section below.

Should you opt out of AI features?

You now can. For most publishers, you should not.

Alongside the reports, Google added a property-level Search generative AI control under Settings in Search Console. It removes a site's links and content from AI Overviews, AI Mode and generative AI features in Discover. Google says the setting is not used as a ranking signal elsewhere in Search. It has no effect on AI training either, which the separate Google-Extended control governs [10]. Britain's Competition and Markets Authority has given Google until March 2027 to add page-level controls, and expects click-through data too [10], so finer options are coming.

The trade-off is simple. Either way, the query still gets answered. Opting out means someone else answers it, and you keep neither the click you were not getting nor the mention you were. The narrow case for opting out is a site whose whole value is in the visit, such as a tool or a paywalled archive. There, being summarised genuinely replaces being used. For a publisher who wants to be known for a subject, the mention is the asset.

What CTAs work in a zero-click search strategy

A call to action on a page most people never visit has a different job: to be remembered, not only clicked. Three types hold up better than "read more":

• Brand-recall CTAs. "Search anobee.com for the ANCHOR checklist" gives a reader something to search for later. "Click here to learn more" does not.
• Named-resource CTAs. A specific, distinctly named tool, template or framework gives the reader a handle, even if they never click.
• Community and email CTAs placed early. A specific invitation to a newsletter or discussion in the first few hundred words gets seen. Nobody reaches a closing pitch on a page they did not open. SparkToro's advice to own your audience through email and community rather than rented reach is the same idea [6]. The email list building guide covers the mechanics.

One pushback on my own advice. Do not swing so far toward brand recall that you strip out every functional link. Readers who do click still need a clear next step, or the page rewards recall and punishes action. The shift is not "remove CTAs" but "stop designing every CTA as if the click is certain".

How to measure a zero-click search strategy without clicks

Clicks alone will keep understating performance as this trend continues. Four measurements close most of the gap, and all four live in tools you probably already run.

1. Impressions in Search Console, including the Generative AI report. Google announced Generative AI performance reports on 3 June 2026 with a small group of UK sites. It rolled them out to all sites worldwide on 31 August 2026 [10]. The Search report shows how often links to your site appeared in AI Overviews and AI Mode, by page, country, device and date [11]. It shows no clicks, CTR or queries, and may not appear at all for sites with too few AI impressions [11]. Treat it as a visibility gauge. In the main Performance report, track impressions on informational queries separately from clicks; a widening gap between the two is the zero-click signature. If your property is not set up cleanly, start with the Search Console setup guide.

2. Branded search volume, also in Search Console. Filter the Performance report by your brand name, with a qualifier if your name is ambiguous. Rising branded search after publishing zero-click-prone content is a reasonable proxy for recall built inside AI answers. Pew's 1% figure means most of that influence never shows up as a click [2].

3. GA4 direct and branded organic sessions. Watch for a rise in direct visits or branded organic sessions without a matching rise in ordinary organic referrals. That pattern often reflects a reader who saw your name in an answer and came back another way. A behaviour tool such as the one in the Microsoft Clarity guide helps confirm those visitors engage.

4. Manual AI checks against a fixed prompt set. Once a month, run the same ten to twenty questions through ChatGPT, Perplexity, Gemini and Claude, and record whether your site is cited or your brand is named. Keep the prompts stable so the trend means something. The check AI citations guide has a repeatable method. Anobee's September 2026 AI citation repeat test shows what a repeated run looks like in practice. Perplexity's freshness-weighted behaviour is worth checking on its own.

Here is how the four fit together. Suppose a post shows 4,000 impressions and 90 clicks in a month, a 2.25% CTR that looks weak on its own. Now suppose branded searches for your name rose after the publish date, and direct GA4 sessions ticked up too. Then the post is probably doing real work in AI answers. It is a failure by clicks, and possibly your best article by the full picture. The reverse holds as well, and my own numbers above are the reverse. Low clicks at position 20 to 30 mean the page needs ranking work, not a new dashboard.

A 30-day zero-click search strategy checklist for a small site

This is the sequence I am running on Anobee. Order matters, because each step depends on the one before it.

1. Capture a baseline first. Export three months of Search Console data for the page and note the Generative AI report figures, or the before-and-after cannot be measured.
2. Check ranking and intent. If the page sits past position 20 or surfaces for the wrong queries, fix the title, opening and headings to match the query you actually want. No zero-click search strategy can rescue a page that is not in the retrieval set.
3. Rewrite section openings answer-first. One clear sentence under each heading that could stand alone if quoted.
4. Audit the byline and entity facts. Same author name, brand string and About-page facts across the site and every profile.
5. Add one piece of original data. A Search Console before-and-after, a small test, a real number nobody else has. Label it honestly, including when it is small.
6. Rewrite the CTAs. One memorable brand or resource prompt near the top, one functional link further down.
7. Fix a monthly prompt set of ten to twenty questions and record AI citations and mentions.
8. Review at day 30 against the baseline, in the same market and device, allowing for normal volatility.

Search "anobee.com ANCHOR checklist" if you want to find this list again without bookmarking it. That sentence is a brand-recall CTA doing its job.

Bottom line

About 68% of US Google browser searches now end without a click [1]. AI Mode, where 92 to 94% of sessions never leave Google, has barely begun to shift that average [3]. Optimising harder for clicks that are not coming is a losing strategy. But a zero-click search strategy is the second problem, not the first: ranking and intent come first, and everything in ANCHOR builds on them.

Fix entity gaps early, because they undercut everything else. Measure impressions, branded search, direct sessions and AI citations alongside clicks, not instead of them. If you have not looked at older articles through this lens, the content audit guide is a sensible next stop. None of this means clicks stop mattering. It means they stop being the only number that tells you whether an article did its job.