> ## Content Index
> Fetch the complete content index at: https://www.promptthemarket.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Google's AI answers cost publishers 40% of their clicks, trial finds
- URL: https://www.promptthemarket.com/ai-overviews-cut-publisher-clicks-by-40-pay-rate-secret/
- Published: 2026-09-19T05:00:00.000Z
- Updated: 2026-09-19T12:43:24.000Z
- Description: Digital Content Next ran the first controlled test of AI Overviews and found a 40% fall in click-through. Google's reply is a payment scheme with a rate it will not disclose.
- Author: Editor News
- Tags: Marketing, Media, #author-james-dore, #by James Dore

![James Dore](https://storage.ghost.io/c/e4/01/e4017178-c1da-4131-80aa-35f02cb5b999/content/images/2026/09/james-dore.jpg)

By **James Dore** · 2026-09-19

TLDR

Two independent studies confirm Google's AI Overviews strip roughly 40 per cent of outbound clicks from publishers when they appear. Google launched a compensation pilot in June 2026 that pays publishers whose content feeds its AI answers, but has disclosed neither the rate nor how it calculates payments.

KEY TAKEAWAYS

01A randomised trial of 1,065 desktop users found 39.8 per cent fewer outbound clicks when an AI Overview appeared.

02Zero-click searches became 34.5 per cent more likely the moment Google generated an AI Overview.

03Only 1 per cent of visits to pages with an AI Overview resulted in a click to a cited source, per a 900-person panel study.

04Google's AI contribution pilot, launched June 2026, pays publishers for grounding AI responses but has not disclosed its rate.

05Across all searches, with and without AI Overviews, the estimated organic click reduction sits at 18.5 per cent.

## What the trial found

Digital Content Next (DCN) published a randomised controlled trial on 14 September 2026 that put hard numbers on what publishers had long suspected. Researchers recruited 1,065 U.S. desktop users and randomly assigned them to different versions of Google Search, with or without AI Overviews present.\[1\] Random assignment is the methodological difference between a study and a hunch.

When Google generates an AI Overview, users make 39.8 per cent fewer outbound organic clicks than when they do not see an AI Overview.\[1\] Put plainly: for every ten clicks a publisher would have received, they got six. The likelihood of a zero-click search, where the user reads the results page and leaves without visiting any site, rose by 34.5 per cent when an AI Overview was present.\[1\]

AI Overviews appeared in roughly 41 per cent of all searches during the experiment.\[1\] Apply that frequency to the 39.8 per cent click reduction and the blended effect across all searches is an estimated 18.5 per cent reduction in outbound organic clicks.\[1\] For a mid-sized news publisher drawing 500,000 monthly visits from Google, that translates to roughly 92,500 visits gone.

## A second study lands in the same place

An academic study posted to arXiv on 5 August 2026 tracked 900 U.S. adults through their real browsing sessions across March 2025.\[2\] The method was observational rather than experimental, capturing natural behaviour rather than an assigned treatment. The findings pointed in the same direction.

Users clicked on a search result in 15 per cent of visits to results pages without an AI Overview, but only 8 per cent when an AI Overview was present, a fall of nearly half.\[2\] Users ended their browsing session entirely on 26 per cent of pages with an AI Overview, compared with 16 per cent of pages without one.\[2\] One in four users who saw an AI Overview stopped searching, got what they needed, and left without reaching a publisher's site.

The citation-click figure is the hardest number in the study. Of all visits to pages featuring an AI Overview, only about 1 per cent resulted in a click to a source cited inside the AI-generated summary.\[2\] Google cites sources inside its AI Overviews as a feature; the data suggests users almost never follow them.

## Google pushes back, then launches a pay pilot

Google's Head of Search, Liz Reid, disputed the premise of studies like these. "This data is in contrast to third-party reports that inaccurately suggest dramatic declines in aggregate traffic, often based on flawed methodologies, isolated examples, or traffic changes that occurred prior to the roll out of AI features in Search," Reid said.\[4\] Google has not published a specific rebuttal to the DCN trial's randomised design.

Google launched what it calls an AI contribution pilot on 18 June 2026\. The program pays publishers when their content grounds generative AI responses, meaning it contributes to the factuality and freshness of what Google's AI answers.\[3\] Google VP Markham Erickson described it this way: "Beyond news, we are piloting a new way to partner with websites whose content meaningfully contributes to the freshness and factuality of generative AI responses through the process of grounding," Erickson said.\[3\]

Google has not disclosed the payment rate or calculation methodology for the AI contribution pilot, which operates through Search Console. Publishers in the program have no independent way to verify what their content is worth to Google's AI, or whether the payment offsets the referral traffic they are losing.

## What this means for Australian publishers and the brands behind them

Australian publishers face the same structural dependency as their U.S. counterparts. Most news and content sites built their audience models on Google referral traffic feeding advertising and subscription funnels. An 18.5 per cent reduction in organic clicks is a structural change to how revenue arrives.

The risk compounds for brands whose earned media value is calculated from publisher coverage. If an article placing a brand in a major publication draws half the readers it would have drawn two years ago, the value assigned to that placement needs revisiting. The DCN trial and the arXiv panel study give marketers and publishers a shared empirical basis for that conversation, even if Google disputes the methodology. The DCN study was published on 14 September 2026; the arXiv paper was posted on 5 August 2026\. Neither has been retracted.

SOURCES & CITATIONS

1. [AI Overviews Reduce the Search Traffic Reaching Publishers, Digital Content Next](https://digitalcontentnext.org/blog/2026/09/14/ai-overviews-reduce-the-search-traffic-reaching-publishers/?ref=promptthemarket.com)
2. [The Impact of AI Overviews on Web Search Behaviour, arXiv](https://arxiv.org/abs/2608.04831?ref=promptthemarket.com)
3. [Supporting the Information Ecosystem, Google Public Policy](https://publicpolicy.google/article/supporting-information-ecosystem/?ref=promptthemarket.com)
4. [AI Search Driving More Queries and Higher Quality Clicks, Google Blog](https://blog.google/products-and-platforms/products/search/ai-search-driving-more-queries-higher-quality-clicks/?ref=promptthemarket.com)

FREQUENTLY ASKED QUESTIONS

What is an AI Overview?

An AI Overview is an AI-generated summary box that Google places at the top of its search results page. It synthesises information from multiple sources to answer a query directly, before the user sees traditional organic links. Google launched them in 2024.

How was the 39.8 per cent figure calculated?

Digital Content Next recruited 1,065 U.S. desktop users and randomly assigned them to search experiences with or without AI Overviews. The 39.8 per cent figure is the difference in outbound organic clicks between the two groups. The randomised design controls for other variables that observational studies cannot.

What is Google's AI contribution pilot?

Google launched the program on 18 June 2026\. It pays publishers whose content contributes to the factuality and freshness of AI-generated search answers, a process Google calls grounding. The payment rate and calculation method have not been disclosed publicly.

Does the 18.5 per cent figure apply to every search?

Yes, it is the blended estimate across all searches, those where an AI Overview appeared and those where it did not. Because AI Overviews appeared in roughly 41 per cent of searches during the DCN experiment, the 39.8 per cent click reduction in AI Overview sessions averages out to 18.5 per cent across all sessions.

![James Dore](https://storage.ghost.io/c/e4/01/e4017178-c1da-4131-80aa-35f02cb5b999/content/images/2026/09/james-dore.jpg)

[James Dore](https://www.promptthemarket.com/james-dore/)

James Dore is Strategy Director at Bushnote, a leader in SEO and AEO in Australia. He writes on search, AI and the technology decisions shaping government and industry.