Modeled Communities: The Synthetic Respondents That Don’t Replace Humans with Dan Braker of Breakthrough Research

Welcome back to Intellicast! On today’s episode, Brian is joined by Dan Braker, President and Founder of Breakthrough Research, for a conversation about one of the most talked-about topics in market research right now: synthetic respondents. Dan returns to the podcast to discuss Breakthrough Research’s new Modeled Communities capability and explain how his team is approaching synthetic respondents differently.

Kicking off the episode, Dan gives some background on Breakthrough Research’s innovation efforts and how the company’s work has expanded from behavioral science methodologies into artificial intelligence. That work ultimately led to Modeled Communities, custom communities of synthetic respondents that clients can use for quantitative surveys, qualitative exploration, concept testing, and other research applications.

The conversation then turns to how these communities are actually built. Rather than creating generic synthetic personas, Dan and his team start with real people who represent a client’s customers, prospects, or target audience. Those individuals participate in detailed intake interviews covering their personal histories, attitudes, behaviors, lifestyles, experiences with the brand, and how they make decisions within a category. Each modeled respondent is then built from one of those real human participants.

Brian and Dan also discuss an important limitation of modeled respondents: unlike real people, they don’t continue evolving after they are created. Real customers change their behaviors, discover new brands, experience major life events, and react to changes in the world. Because of that, Dan explains why Modeled Communities need to be replenished periodically with fresh human respondents, especially when major changes occur within a brand, category, or broader environment.

The discussion then shifts to one of the biggest questions surrounding synthetic data: validation. Dan walks through Breakthrough’s two-step validation process, including holding back information from the initial modeling process and comparing the modeled respondent’s answers with those of the real individual. The team also compares aggregate results from Modeled Communities with traditional human research to determine whether the synthetic respondents produce similar themes and, most importantly, lead researchers toward similar decisions.

From there, Brian and Dan explore where Modeled Communities fit within an existing research program. Dan makes one point particularly clear: the goal is not to replace human research. Instead, modeled respondents can help researchers explore more ideas, test more concepts, iterate more quickly, and identify the strongest options before investing in a traditional human study. A team that might normally narrow 50 ideas down to five internally can instead use a Modeled Community to help determine which ideas actually deserve to reach real customers.

Dan also discusses several potential use cases, including hard-to-reach audiences, concept and message testing, and highly confidential research. Because synthetic respondents don’t suffer from survey fatigue or share confidential concepts with friends or on social media, they can provide researchers with additional flexibility before going to the market with higher-risk research.

The episode wraps up with a broader discussion of how synthetic and human research can work together. As Dan explains, the voice of the customer remains valuable precisely because it takes time and money to access. Modeled Communities offer another way to make sure that when researchers do go back to real people, they are bringing their best questions and strongest ideas with them.

You can connect with Dan via email at DBraker@bthrough.com or on LinkedIn: https://www.linkedin.com/in/danbraker/

You can also visit Breakthrough Research’s website here: https://bthrough.com/

Thanks for tuning in!

If you missed the live version of our recent webinar, Most Wanted: The New Fraudsters Affecting Market Research, you can still check it out on-demand here: https://us02web.zoom.us/webinar/register/WN_jEsQAF2sSleFPIGOcTqf4w

The 2026 Edition of The Sample Landscape is now available! You can download your copy here: https://content.emi-rs.com/the-sample-landscape