Most users only read top third of Google AI Overviews

Four days ago, on May 12, 2025, the first comprehensive user experience study of Google's AI Overviews was published, revealing dramatic shifts in search behavior that will force marketers to rethink their digital strategies. The research conducted by Kevin Indig and Eric van Buskirk tracked 70 users across eight search tasks, documenting exactly how people interact with Google's AI-generated answers.

The study provides unprecedented insights into how users engage with AI Overviews and makes one thing abundantly clear: the era of optimizing purely for clicks is over, as visibility and trust now reign supreme in the search ecosystem.

Perhaps the most surprising finding is how little of the AI-generated content users actually consume. While 88% of users clicked "show more" to expand truncated AI Overviews, the median scroll depth was just 30% - meaning most users never even see content beyond the first third of these AI-generated summaries.

This pattern fundamentally changes the visibility game for brands. According to the study results, average scroll depth reached 75%, but this figure is heavily skewed by outliers who read more deeply. The median figure of 30% provides a more accurate picture of typical user behavior.

"If your brand isn't surfaced in the first third of an AI Overview, it's effectively invisible," Indig noted in his analysis of the findings.

Trust becomes the primary decision filter

The research discovered a profound shift in how users evaluate search results. Rather than selecting content based primarily on relevance to their query, users now apply a two-step filter where trust precedes relevance:

1. "Do I trust that this domain/brand/company can answer my question truthfully?"
2. "Can this result answer my question?"

One of the ways that the AIO usability study changed Indig's mental model of SEO is the role of trust. Before, he thought relevance was the most important filter that users apply when scanning search results, and results higher up have a higher chance of catching that attention. However, the study revealed this assumption was incorrect. Instead, the main filter is trust.

This represents a fundamental shift in search behavior, with significant implications for marketing strategies. The study found that when a recognized brand, authority site, or .gov/.edu domain appeared in search results, it was chosen first in 58% of cases where such links were present.

AI Overviews dramatically reduce click-through rates

The introduction of AI Overviews correlates with a measurable decline in organic visibility and clicks, particularly for top-ranking, non-branded keywords. The study revealed dramatic differences between mobile and desktop behavior:

• Desktop CTR drops by approximately half when an AI Overview appears
• Mobile clicks fall by about a third when AI Overviews are present

For desktop users, outbound clicks fell to just 7.4% when AI Overviews were present, while mobile users clicked external links 19% of the time. These figures represent a substantial reduction compared to searches without AI Overviews, where desktop users clicked through 28% of the time and mobile users 38%.

This pattern aligns with other research in the industry. According to a separate analysis by Seer Interactive examining approximately 10,000 informational intent keywords, organic CTR fell from 1.41% to 0.64% year over year for queries where AI Overviews appeared.

Demographics and device type influence AI Overview engagement

The study uncovered significant demographic differences in how users interact with AI Overviews. Younger users, particularly those aged 25-34, were considerably more likely to trust and rely on AI-generated answers compared to older demographics.

Device type also played a major role in user behavior:

• Mobile users scrolled an average of 54% through AI Overviews
• Desktop users scrolled only 29% on average

Search intent with higher stakes (financial or health-related queries) also drove deeper reading behavior, with users scrolling further to validate information for consequential decisions:

• Health YMYL (Your Money, Your Life) queries: 52% average scroll depth
• DIY how-to queries: 54% average scroll depth
• Financial YMYL queries: 46% average scroll depth
• Decision timing queries: 41% average scroll depth
• Promo code queries: 34% average scroll depth

Reddit and community content absorb residual traffic

The study revealed an interesting pattern in where users go when they do leave Google. Social proof platforms - particularly Reddit, YouTube, and other community forums - have become major beneficiaries of the traffic that doesn't stay within AI Overviews.

When users leave the SERP after looking at an AI Overview, community links receive a disproportionate share of those clicks (18% when AI Overviews are not present). The researchers observed that users, especially in younger demographics, use these platforms to validate the AI-generated information with human perspectives.

For particular query types, such as how-to searches, users sometimes intentionally skip AI Overviews in favor of richer media formats like videos, which received lengthy engagement times averaging 37 seconds - longer than the average AI Overview dwell time of 31 seconds.

Why this matters

This study represents a paradigm shift for digital marketing strategies. As van Buskirk noted, "Search has flipped from a click economy to a visibility economy."

For marketers, several critical implications emerge:

1. Shift KPIs from clicks to presence: Traffic metrics are no longer sufficient indicators of success. Brands must track how often, how prominently, and for which queries they appear in AI Overviews.
2. Prioritize authority building: The study shows that recognition from authoritative sources (.gov, .edu, expert reviewers) dramatically influences user trust, with 58% of users choosing a cited source first.
3. Optimize for skimmers: Content structure should prioritize key facts in the opening section, as most users will never scroll past the first third of an AI Overview.
4. Invest in community validation channels: Since residual clicks often go to Reddit, YouTube, and forums, maintaining presence on these platforms becomes increasingly important.
5. Differentiate mobile and desktop strategies: The significant behavioral differences between devices necessitate distinct approaches, with mobile remaining the more promising source of referral traffic.

The research leaves several critical questions for the industry to consider. These include how Google determines which sources to surface in AI Overviews, whether analytics tools will eventually expose AI Overview impressions, and how publishers can adapt their monetization models as outbound traffic continues to decline.

Gartner has predicted organic traffic will fall 25% due to AI Overviews, with some industries already seeing much greater impacts. According to Elisa Gabbert, Director of Content and SEO at WordStream by LocaliQ, this is going to make content marketing much harder for all businesses, but smaller businesses in particular.

The study authors suggest this may require fundamental changes in how we measure search success. "Marketing dashboards that track only visits are under-valuing visibility wins or hiding looming losses," Indig noted in the research publication.

As AI continues reshaping the search landscape, brands that adapt quickly to this new visibility paradigm - prioritizing trust, authority, and strategic positioning within AI Overviews - will likely maintain competitive advantage in an increasingly challenging environment.

Timeline

• May 12, 2025: Kevin Indig and Eric van Buskirk publish the first comprehensive UX study of Google's AI Overviews
• April 21, 2025: Google's AI Overview feature encounters another embarrassing error when it incorrectly stated that prominent SEO professional Lily Ray is 9 years old and not a dog, misinterpreting information from her website about her dog Marcy
• March 10, 2025: Google confirms to Adweek it will "explore bringing ads" to its new AI Mode search experience, using lessons from ads already running in AI Overviews
• February 5, 2025: Seer Interactive analysis shows organic CTR fell from 1.41% to 0.64% year over year for queries where AI Overviews appeared
• October 3, 2024: Google announces AI Overview Ads are going live for mobile U.S. users, with ads appearing directly underneath the organic AI-generated answer in the AI Overview
• May 14, 2024: Google announces the wide launch of AI Overviews (formerly known as Search Generative Experience) after running it as an optional experiment for a year