Zero-click on Google: statistics, CTR decline and impact of AI (2019–2025)

I'll be direct: if your acquisition strategy depends on Google's organic traffic, you have a serious problem. And I'm not saying it for sensationalism. I'm saying it because I've spent months measuring what's happening in our clients' accounts, cross-checking that data with the most rigorous studies in the industry, and the conclusion is inescapable: Google is keeping your clicks.

The phenomenon is called zero-click (searches without a click) and it's not new. But in 2025 it's reached a magnitude that makes it a real threat to the P&L of any company that depends on organic visibility. According to the SparkToro and Datos (Semrush) July 2024 study, out of every 1,000 Google searches, only 360 clicks reach the open web in the US. In Europe, 374. Google keeps the rest.

And the worst part: with the massive rollout of AI Overviews (the AI-generated answers Google places above all organic results), the situation has accelerated dramatically. Data from Similarweb shows zero-click searches have gone from 56% to 69% between May 2024 and May 2025. Thirteen percentage points in twelve months. That's not a trend. It's a structural change.

What nobody is telling, and which to me is the most important strategic reading of all this, is that search volume keeps growing. Google processed between 9.1 and 13.6 billion daily searches in 2025, 21% more than in 2024. More people are searching than ever. But fewer people are clicking. That's not a demand crisis. It's a distribution crisis. And the difference matters a lot when deciding what to do.

Where does the 69% number come from and how did we get here?

The zero-click concept was popularized by Rand Fishkin, founder of SparkToro, in 2019. His first study using Jumpshot data revealed that 50.33% of Google searches ended without any click to an external website. It was the first time the 50% threshold was crossed. The industry was alarmed, but not enough.

In 2020, using broader SimilarWeb data (including mobile and iOS), the figure jumped to 64.82% globally. On mobile, the data was devastating: 77.2% of searches without a click. According to Search Engine Land, Google was turning its results into final destinations instead of intermediaries.

The most rigorous study to date came in July 2024, when SparkToro partnered with Datos (a Semrush company) to analyze tens of millions of users. The results: 58.5% of searches in the US and 59.7% in Europe end without a click. Almost 30% of clicks that do occur go to Google properties (YouTube, Maps, Images).

In the first quarter of 2025, the data worsened: organic CTR fell from 44.2% to 40.3% in the US, according to the State of Search report from SparkToro/Datos. Almost four points less in just twelve months. In Europe, the decline was from 47.1% to 43.5%.

And then there's the Bain & Company data (yes, Bain, not an SEO blog): 80% of consumers already trust zero-click results in at least 40% of their searches. Their estimate is that organic web traffic has been reduced between 15% and 25% as a direct consequence.

Between 2019 and 2025, the percentage of Google searches ending without a click has grown from 50% to 69%. Out of every 1,000 searches, only 360 clicks reach the open web. The 21% growth in search volume doesn't compensate for the drop in distribution: more people search, fewer people click.

A case that illustrates this perfectly: we had a client in the insurance sector, led by Nerea Forján, who depended 72% of their lead acquisition on informational organic traffic. Queries like "best life insurance for freelancers", "health insurance comparison". In six months, traffic to those pages fell 41%. Leads, 38%. But the interesting part is that their positioning hadn't dropped: they were still in position 1-3 for those keywords. The problem wasn't ranking. The problem was that Google was answering directly and the user didn't need to click.

For us, the strategic reading was clear: ranking is no longer enough. If you rank for a query Google answers directly, you're winning a battle that no longer matters. What matters is being in the answer. Or, better yet, being the answer.

Translation to GEO: Don't fight for clicks on queries Google already resolves without a click. Fight to be the source Google cites when it gives that answer. That's GEO.

How much traffic is AI Overviews really destroying?

Google launched AI Overviews (formerly SGE) in the United States on May 14, 2024. By May 2025, the functionality was active in more than 200 countries and 40 languages, with 1.5 billion monthly users. In Spain it was activated on March 26, 2025, under the name "AI-generated views". And Spain leads Europe in penetration: between 14% and 16% of searches already trigger an AI response, compared to 1% in Germany.

The impact on CTR is documented by multiple independent studies, and they all point in the same direction. The most comprehensive is from Seer Interactive (3,119 informational queries, 42 organizations, 25.1 million organic impressions): organic CTR on queries with AI Overview fell 61%. From 1.76% to 0.61%. Paid CTR fell 68%. Read it again: 61% less organic clicks.

But there's a piece of data most articles omit and which to me is the most revealing: even queries without AI Overview experienced a 41% year-over-year drop in organic CTR. This suggests user behavior change goes beyond AI Overviews. People are learning to consume information directly in the SERP. They need to click less and less.

Ahrefs analyzed 590 million searches and found that AI Overviews reduce clicks to position 1 content by 58%. The GrowthSRC study with 200,000+ keywords puts position 1 CTR decline at 32% (from 28% to 19%) and position 2 at 39%.

There's an important methodological nuance Semrush contributes in their own AI Overviews study: when they track the same keywords before and after an AI Overview appears, zero-click actually drops slightly (from 38.1% to 36.2%). The interpretation: Google tends to activate AI Overviews on queries that were already prone to zero-click. Correlation isn't exactly causation. But aggregate impact on publishers remains devastating.

That said, and this is something I think is overlooked in almost every analysis I read: there's a brutal positive data point buried in the negative numbers. The Seer Interactive study found that brands cited within AI Overviews receive 35% more organic CTR and 91% more paid CTR than non-cited brands. Amsive documented that brand searches show an 18.68% increase in CTR even when AI Overviews appear. That is: if you get AI to cite you, you don't just not lose —you win.

AI Overviews reduce organic CTR between 30% and 61% according to the study. But brands cited within those AI responses receive 35% more organic clicks. The question is no longer "how do I rank" but "how do I get AI to cite me". That's exactly what GEO solves.

Another case of ours that illustrates both sides of this coin: a client (browniespain.com) had a blog with 340 informative articles that generated 28% of total traffic. After AI Overviews were activated in Spain, we did a keyword-by-keyword analysis. Of the 50 main informational keywords, 31 already showed AI Overview. Traffic to those 31 articles had dropped 47% on average. But transactional keywords ("buy long dress", "running shoes offer") were barely affected: less than 4% showed AI Overview. Google protects the queries that feed its advertising business.

Translation to GEO: Google doesn't apply AI Overviews equally. Pure informational queries are most affected. Transactional ones, the most protected. For GEO, this means prioritizing consideration content (comparisons, choice guides, analysis) over purely definitional content.

Who are the big losers (and winners) of zero-click?

This isn't theory. There are companies that have fallen. The most emblematic is Chegg, the American educational platform. Their non-subscriber traffic fell 49%. Revenue dropped 24% year-over-year. Subscribers fell 21%. The stock plummeted 99% —from $113 to approximately $1— and the company sued Google for using their database of 135 million questions without compensation. Afterwards they cut 45% of their workforce.

HubSpot saw its monthly organic traffic plummet from 13.5 million visits to less than 7 million between November and December 2024. Their content strategy was built on top-of-funnel informational queries ("famous sales phrases", "cover letter examples") that AI Overviews summarize effortlessly. Their CEO Yamini Rangan acknowledged in an earnings call that organic traffic "is falling globally".

According to data from Growtika and AdExchanger, losses in media are massive: Digital Trends lost 97% of their search traffic, ZDNet 90%, The Verge 85%, NerdWallet 73%, Healthline 50%. The Planet D, a travel blog that had been around for over a decade, had to close completely after losing 90% of its traffic. Aggregate publisher losses are estimated at around $2 billion in advertising revenue.

But there are also winners, and that's what I find most interesting. Dotdash Meredith (40+ brands, including Allrecipes and Investopedia) pivoted from a pageview-dependent model to contextual segmentation with its proprietary D/Cipher tool. Result: $522M revenue in Q4 2024, +10% year-over-year. HouseFresh, after losing 91-95% of its Google traffic, diversified toward YouTube and recovered its previous traffic in October 2025. Grow & Convert reports that clients focused on bottom-of-funnel keywords experience drops of only 10-20% in traffic, with stable or growing conversions.

The companies most affected by zero-click are those that built their acquisition on pure informational content (Chegg, HubSpot, media). Those that resist best are those that bet on transactional content, strong brand and channel diversification. Position on Google no longer guarantees traffic; citation in AI does guarantee relevance.

Translation to GEO: Losers optimized to rank. Winners optimize to be indispensable. In GEO terms, that means creating content with original data, proprietary frameworks and analysis no AI Overview can replace in two lines.

What do the experts who know most about this say?

Rand Fishkin (SparkToro) has evolved from documenting zero-click to embracing it as a marketing philosophy. His March 2025 research showed that Google receives 373 times more searches than ChatGPT. All AI tools combined represent less than 2% of the search market. His provocation: "That guy on your LinkedIn feed screaming AI will destroy Google has spent years without contributing data." But he also acknowledges that AI responses "kill click-through rates" even without reducing search volume. His proposal: measure brand influence, not clicks.

Lily Ray (VP of SEO Strategy & Research at Amsive) nuances the panic but doesn't dismiss it. Google still controls more than 90% of the global market. AI platforms represent less than 1% of referral traffic. A survey she published on LinkedIn showed that 70% say ChatGPT represents less than 2% of their web traffic. But she warned: if Google makes AI Mode the default mode, "it will have a devastating impact on the internet". And she pointed out something I 100% agree with: many "GEO" recommendations circulating are simply traditional SEO repackaged. That's why at Elevam we don't sell GEO as an isolated service: it's part of the CREF© growth system.

Kevin Indig (former SEO lead at Shopify, G2, Atlassian) conducted the first AI Overviews usability study with 70 participants. Result: clicks fell from 28% to 11% on desktop. His meta-analysis: traffic drops between 15% and 45%, but "the sales pipeline tends to be flat or growing". His conceptual framework seems powerful to me: we're moving from a click economy to an engagement model.

Gartner made the boldest prediction in February 2024: "By 2026, search volume in traditional engines will drop 25%." It seems premature —Google grew 21.64% in 2024— but the strategic direction they suggest (diversifying channels) has proven correct.

My reading of all this: Fishkin is right that Google isn't dying. Lily Ray is right that GEO as an isolated service is smoke. Indig is right that pipeline matters more than traffic. But what none of the three say explicitly is what for me is the key conclusion: zero-click doesn't destroy value, it redistributes value. And it redistributes it to brands AI systems consider authority sources. If you're that source, zero-click works for you. If you're not, it works against you.

Translation to GEO: Expert opinion is that SEO doesn't die but transforms. In practical GEO terms: building entity authority (mentions, proprietary data, citable frameworks) is more important than any on-page technical optimization.

Are ChatGPT and Perplexity really a threat or a hidden opportunity?

Beyond Google's AI Overviews, there's a second front: LLM-based search engines. ChatGPT reached 800 million weekly active users in March 2025 and processes approximately 1 billion daily queries. Perplexity AI already processes more than 780 million monthly queries with a valuation of $18-20 billion.

The zero-click implications of these systems are even more severe: between 92% and 94% of sessions in Google AI Mode end without any external click. The crawl vs. referral ratio is revealing: Google maintains a 10:1 ratio (crawls 10 pages per click returned), while OpenAI is estimated at 1,200:1 to 1,700:1. They extract enormously more value than they return.

But here comes the paradox every CEO should understand: traffic that does arrive from ChatGPT converts at a rate of 15.9%, according to Semrush. Perplexity at 10.5%. AI-referred visitors convert 4.4 times better than traditional organic traffic. Less volume, much more quality.

I verified this with a B2B SaaS client: we started tracking traffic from ChatGPT and Perplexity in September 2024, when they were still tiny quantities. By February 2025, that channel represented 3.2% of total traffic. Seems little. But that 3.2% generated 11.7% of requested demos. The conversion rate was 3.6 times higher than classic organic traffic. When we translated it to pipeline impact, acquisition cost was 68% lower. Volume is still small, but unit value is enormous.

Traffic from ChatGPT converts at 15.9% and Perplexity at 10.5%, compared to much lower rates from traditional organic. LLMs generate fewer visits but of radically superior quality. For B2B companies, pipeline impact can be disproportionate to volume.

Translation to GEO: ChatGPT and Perplexity aren't the enemy. They're an emerging acquisition channel with higher conversion rates. GEO work isn't defensive (avoid losing traffic) but offensive (gain presence in responses that generate business).

What does the Princeton GEO study propose to combat this?

The founding paper is "GEO: Generative Engine Optimization" by researchers from Princeton, IIT Delhi and Georgia Tech, presented at KDD 2024 (ACM SIGKDD). It's the first academic framework to optimize visibility in AI generative responses.

They tested nine optimization techniques on a 10,000-query benchmark and found three that consistently improve visibility by up to 40%:

1. Cite credible sources within content: improvement up to 40%. For websites ranking in position 5, the improvement was 115%. A small website that cites well can beat a large one that doesn't.

2. Add statistics and quantitative data: improvement up to 33.9%. LLMs prioritize content with specific data over generic content.

3. Include expert quotes: improvement up to 32%. The perceived authority of content improves when statements are attributed to people with name and title.

Critical data: keyword stuffing performed 10% worse than baseline. Classic SEO tactics not only don't work for GEO —they actively damage visibility in AI responses.

According to the Princeton/Georgia Tech GEO study (KDD 2024), the three most effective techniques to improve visibility in AI responses are: citing credible sources (+40%), including statistics (+33.9%) and incorporating expert quotes (+32%). Keyword stuffing makes it worse by 10%.

At Elevam we're approaching it like this: the CREF© model and HSA protocol applied to zero-click

We've integrated these findings into our CREF© framework and HSA protocol. It's not a rebranding of SEO. It's a growth system that starts from the P&L and decides which channels, content and optimizations make sense for each business. Within that system, the response to zero-click is structured in five layers:

Layer 1 — Structured extractability. Write one idea per paragraph, front-load key points, create 40-60 word blocks ready to be extracted as a snippet, use semantic headings that directly answer questions. This is basic, but almost nobody does it well.

Layer 2 — Proprietary data density. AI systems prioritize content with original research. We publish our own studies (like our ranking of GEO agencies in Spain with real tests on ChatGPT, Claude, Gemini and Perplexity) because we knew AIs would cite unique data before generic content.

Layer 3 — Entity authority building. AI tracks brand mentions even without links. Building an entity map —presence on Reddit, review platforms, sector publications, consistent NAP citations— directly feeds the probability of being cited. Brand searches show AI Overviews only 4.79% of the time. Brand is the most powerful moat.

Layer 4 — Multimodal content. Pages with multimodal content receive 67% more AI references. Video, infographics, tables, calculators: it all adds up.

Layer 5 — Triple AI visibility. Pre-training (being present in sources LLMs absorb: Reddit, Wikipedia, sector publications), search layer (classic SEO still matters because ChatGPT searches in Bing and Perplexity queries Google), and selection filter (GEO optimizations determine what content is cited).

A case that validates this approach: with a client in the online training sector (AulaVall), we applied these five layers to their cluster of 45 main pages for four months. Traditional organic traffic dropped 19%. But mentions in Google AI Overviews rose from 0 to 12 queries. Visits from ChatGPT and Perplexity grew 340%. And most importantly: total qualified leads rose 8%, with a 23% lower cost per lead. Less traffic, better business.

Translation to GEO: Elevam's CREF© applies GEO as part of a system, not as an isolated tactic. The five layers (extractability, proprietary data, entity authority, multimodality, triple AI visibility) are designed so content is citable by language models, not just indexable by crawlers.

Why does this affect us in Spain more than anyone in Europe?

Spain has a peculiarity that makes zero-click have a proportionally greater impact: Google controls approximately 96% of the search market. There's no Yandex, no Baidu, no relevant local alternative. If Google decides not to send you clicks, you have nowhere to go.

Additionally, Spain leads Europe in AI Overviews penetration: between 14% and 16% of searches already show them. Germany is at 1%. The language barrier that traditionally "protected" Spanish websites from Anglo-Saxon content no longer exists: LLMs are native multilingual.

But there's an opportunity practically no one is taking advantage of: in the Spanish market there is no systematically GEO-optimized content. Consultancies and agencies in Spain keep talking about classic SEO as if it were 2019. Nobody measures Share of Answer. Nobody does citation engineering. Nobody publishes studies with verifiable methodology on brand performance in AI searches.

That's exactly the gap we're attacking from Elevam. And that's why I write this article with this level of detail and data: because I believe the content LLMs cite is content that brings new data, proprietary conceptual frameworks and analysis that can't be found anywhere else.

Spain combines the largest Google market share in Europe (~96%), the highest AI Overviews penetration on the continent (14-16% of searches) and the practical non-existence of GEO strategies among local consultancies and agencies. For companies that move first, the window of opportunity is enormous.

Translation to GEO: In Spain, whoever builds GEO entity authority now will dominate a space their competitors aren't even seeing. It's the equivalent of doing SEO in 2008: few did it, those who did it well are still reaping the fruits.

What should you do tomorrow if you're a CEO or CMO?

I'll be specific. If your company bills between 1 and 20 million and your acquisition partly depends on organic traffic, these are the five actions I would prioritize:

1. Audit how much of your traffic is vulnerable to zero-click. Review your 50 main keywords. How many show AI Overview? How many have featured snippets that fully answer the query? If more than 40% are pure informational queries, you have a concentration problem.

2. Rebalance your content mix toward consideration and transactional queries. AI Overviews appear in less than 4% of transactional queries. "What is X" articles are dead as traffic generators. "Best X for [specific use case]", "how to choose X when your situation is Y" articles still work. Grow & Convert reports drops of only 10-20% in traffic for clients focused on bottom-of-funnel, with stable or growing conversions.

3. Measure your AI visibility, not just on Google. Start tracking what ChatGPT, Perplexity, Claude and Gemini answer when someone searches your category. At Elevam we use what we call the Prompts x Models Matrix: 15 strategic prompts tested on 4 LLMs, with periodic Share of Answer measurement. If your brand doesn't appear in those responses, you're invisible to a growing part of the market.

4. Invest in brand, not just content. Brand searches show an 18.68% CTR increase even when AI Overviews appear. Users searching for your brand convert 4 times more. In a zero-click world, brand is the asset Google can't take from you.

5. Optimize to be cited, not to rank. The most important finding from the Princeton study: citing sources, including specific data and incorporating expert quotes can improve your visibility in AI responses by up to 40%. Keyword stuffing makes it worse by 10%. We're moving from optimizing for clicks to optimizing for citations.

My conclusion: less traffic, better business

Zero-click isn't going to disappear. It's going to get worse. Gartner predicts drops of 50% or more in organic traffic by 2028. They may be wrong on timing, but the direction is unequivocal.

But here's what almost no one is saying and which for me is the central thesis of this article: zero-click doesn't destroy market value. It redistributes it. It redistributes it from websites that accumulated informational traffic without conversion toward brands AI systems consider authority sources. If you're that source, zero-click works for you. If you're not, it works against you.

Companies that understand this and adapt their acquisition system —measuring influence in addition to clicks, building brand in addition to content, optimizing for citations in addition to positions— will gain share in a market where most keep looking at metrics that no longer reflect reality.

That, by the way, is exactly the reason Elevam exists as a growth consultancy and not as an SEO agency. Because SEO is a tactic within a system. And the system needs to be rewritten.

Related reading

• How to expose data to AI with schema, feeds and entity
• Why AI doesn't recommend the same to everyone
• Public GEO baselines from Elevam Labs

Shall we work together?

If you want to apply this in your company with a team that combines technical SEO, GEO and paid acquisition measured against the income statement, request a no-commitment audit. You can also check real case studies or read the public GEO baselines that Elevam Labs publishes every quarter.