Answer extracted from the Pod Save America podcast — listen to the full episode below.
How Priorities USA Analyzed 6 Million Searches Across AI and Social Platforms
Priorities USA examined approximately six million searches from opted-in users across multiple platforms—Google, Gemini, ChatGPT, YouTube, Instagram, TikTok, and Facebook—collected during late 2024 and a portion of 2025. The organization processed this large dataset using AI extensively to identify behavioral patterns and understand how less engaged voters discover political information.
A Comprehensive Cross-Platform Behavioral Dataset
The study's scope was deliberately broad, capturing search and content-consumption behavior across the digital landscape that actually influences undecided and less politically engaged voters. Rather than relying on traditional polling or media consumption estimates, Priorities USA worked with a third-party provider to gather real behavioral data from users who explicitly opted in and agreed to share their information.
The two-period timeframe—late 2024 and part of 2025—allowed the organization to analyze how search behavior and content discovery patterns evolved across the critical pre-election window. As Danielle Butterfield explains in the episode, this methodology prioritized authenticity over convenience, measuring actual behavior rather than reported habits.
Why Scale and AI Processing Mattered
The magnitude of the dataset—six million searches—required extensive AI analysis to extract meaningful patterns. Without computational processing at that scale, identifying which platforms, content types, and search queries most influenced voter behavior would have been nearly impossible.
This approach revealed insights that traditional media research could not capture. The platforms included in the study—particularly TikTok, YouTube, and AI chatbots like ChatGPT—represent the actual information ecosystem for millions of voters, especially younger and less politically engaged audiences. The full episode dives into how Priorities USA discovered that YouTube consumption patterns were more predictive of vote shifts than traditional demographic factors.
"Nine out of the ten top YouTube ads in 2024 were just 30-second ads that ran on TV first. We are quite literally practically taking 30-second ads that we run on TV and then putting them on YouTube and expecting people to watch them — and that is quite literally lighting money on fire."
Danielle Butterfield — Executive Director of Priorities USA. Since joining the organization in 2017 following the Hillary Clinton campaign, where she served as Digital Director, Butterfield has led the transformation of Priorities USA from a traditional super PAC into a digital-research-first organization. She negotiated to build an in-house innovation team focused purely on cutting-edge strategy rather than outsourcing to consultants driven by profit margins, positioning the organization as a leader in understanding how modern voters actually consume political information.
The research methodology also illuminated a critical gap in how political campaigns deploy resources. The podcast discusses how this six-million-search analysis revealed inefficiencies in consultant-driven ad strategies, where the same 30-second TV formats are repurposed for platforms with entirely different user expectations and attention spans.
See also
How does Priorities USA study the media habits of less politically engaged voters?
Priorities USA runs an always-on survey operation where participants are paid to send screenshots of their screen-time apps, which are then validated to ensure accurate data collection.
What did Priorities USA's post-2024 research reveal about YouTube consumption and vote behavior?
Priorities USA found that YouTube consumption was more predictive of vote choice and vote swing from 2020 to 2024 than age, demographic, or geographic factors.
What role do political consultants play in perpetuating outdated digital strategies?
Butterfield points to the consultant-driven nature of campaign decision-making as a core obstacle: producing 20 tailored digital ads instead of one TV spot limits reach and wastes resources on fragmentation.