Two-Thirds of Searches Now End in Silence

About 68% of US Google searches now end without a click. A June 2026 study by SparkToro and Similarweb put the figure at 68.01%, up from roughly 60% in 2024 and about 45% a decade ago. For every 1,000 searches, only about 276 now reach the open web, down from 374 two years earlier: almost a quarter of that traffic gone in 24 months. The main driver is visible at the top of the page, Google AI Overviews, which now appear on more than a fifth of searches and, when present, cut click-through to traditional links by close to 60%.

Metric

2024

2026

US searches ending without a click

~60%

68.01%

Clicks to the open web per 1,000 searches

374

~276

Source: SparkToro and Similarweb, June 2026.

A non-click is still a problem because the answer is being given by someone other than you, and it may be wrong. When an AI summary describes your product, recommends a competitor or quotes an out-of-date price, it does so at the first-impression moment, and you had no say in it. Bain & Company found that about 80% of consumers now rely on AI answers for at least 40% of their searches, cutting organic web traffic by an estimated 15 to 25%. The flip side is the opportunity: Seer Interactive found in November 2025 that brands cited inside Google AI Overviews earn around 35% more organic clicks, and 91% more paid clicks, than uncited competitors on the same query. Being in the answer is the new being on page one.

Peer-reviewed research says four content-level changes get a brand cited, and none of them is a technical trick. The foundational study came from Princeton and was presented at KDD 2024. Researchers built a 10,000-query benchmark and tested nine ways of changing content; the headline result was that simple content changes lifted a page’s visibility in AI answers by up to 40%, and lower-ranked pages benefitted most, so a brand does not need to be the largest to be cited. Traditional keyword tactics had little to no effect in generative engines.

Change you make to your content

Why it works

Add named expert quotations

A direct quote from a credible person is easy for an engine to lift and trust.

Add real statistics

Specific numbers read as verifiable fact, which engines prefer to vague claims.

Cite your own sources

Linking to credible references signals thoroughness and lifts citation odds.

Write clearly and plainly

Simpler, fluent writing is easier for a model to parse and reuse.

Source: Aggarwal et al., "GEO: Generative Engine Optimization", KDD 2024 (arXiv:2311.09735).

Content structure does change citation rates, independently of the words. Researchers at the University of Tokyo kept a page’s wording identical and changed only its structure, the headings, the chunking and where the emphasis fell, then measured citations across six AI engines. Structure alone moved citation rates by about 17.3%: the same facts and claims, organised so a machine can find and lift them cleanly. This is recent preprint research, so treat the precise figure as directional, but the direction is consistent across independent studies.

Stop measuring success only in clicks and traffic. Most of your audience now gets answers without visiting you; track whether AI engines cite and describe you accurately instead.

Write to be quoted, not just to rank. Lead with clear answers, real numbers, named expert quotes and credible sources.

Make your content easy to read for machines too. Clean structure, clear headings and unambiguous facts measurably increase the odds of being cited.

Treat accuracy as risk management. If AI will describe your brand either way, the priority is making sure what it says is correct and current. Start with an AI visibility audit.