Google AI Mode Reduced Publisher Clicks in a Controlled Search Experiment

A recent experiment conducted by Google demonstrated that its AI Mode feature can significantly reduce the number of clicks to external publishers. In a controlled seven-day field experiment with 1,100 US participants, those assigned to AI Mode experienced an 18.8 percentage point decrease in clicks to external sites compared to the standard Google Search. This shift in user behavior suggests that AI-mediated search may not only alter rankings but also impact whether users leave the Google ecosystem entirely.

The study, titled "AI in Search Reduces Publisher Referrals Without Improving User Experience: Experimental Evidence," measured the impact of Google Search, both with and without AI Mode, using a browser extension. Participants were randomly assigned to one of three groups: a version without AI Search features, unmodified current Google Search, or a version that forced all queries into AI Mode.

This design allowed researchers to observe real user behavior rather than relying solely on surveys or aggregate traffic estimates.

Notably, the AI Mode group spent more time in the AI-mediated search experience, but this extra time did not lead to increased engagement with external websites or stronger user sentiment. In fact, participants who used AI Mode reported lower trust in the information found on Google and engaged in fewer daily Google search sessions. Click-through rates to news sites, Reddit, and Wikipedia were all 12.5%, 21.2%, and 9.9% lower, respectively, in the AI Mode condition compared to the standard search condition.

While the findings highlight the potential impact of AI Mode on publisher referrals, it's important to note that the results are specific to the experiment's design and implementation. The study does not provide a forecast for future versions of Google Search or every market. Nevertheless, the research offers valuable insights for website owners, marketers, and businesses that rely on search referrals.

It underscores the need for organizations to consider alternative strategies for measuring search demand, brand discovery, and content usefulness beyond just click volume.

For SEO and acquisition teams, the immediate takeaway is not to abandon Google Search. Instead, they should prepare for an environment where click volume alone may no longer accurately represent search demand, brand discovery, and content usefulness. Publishers, content-led businesses, and companies whose websites answer early-stage customer questions should monitor whether their clicks decline while visibility remains stable or grows.

In such cases, AI-generated answer experiences may be one possible explanation. It's crucial to differentiate between different types of search intent and focus on building content that earns deeper visits and engagement. Additionally, diversifying audience acquisition channels beyond search can help mitigate the potential impact of AI-mediated search on overall acquisition strategy.