From Synthetic Data to Synthetic Societies
Written by: Ray Kong, CMO of Arima
If you were at SocialWest 2025 conference in Calgary, you may have heard about synthetic data for the first time. And judging by the buzz and the registrations for our July 16th webinar, you’re ready to learn more.
Let’s step back. Synthetic data isn’t new. It first appeared in the 1970s and 80s to test statistical models and software systems. But today, it's at the forefront of innovation. Why? Because the way we access and use data has changed dramatically. At the same time, the risk and limitations of traditional data are becoming more apparent.
Marketers are under pressure. Privacy regulations are tightening. Audiences are fragmented. Traditional research is suffering from oversampling, fraud, and fatigue. Even AI bots can fake human responses, undermining the reliability of survey data. The result? A growing data integrity crisis.
Meanwhile, industries like healthcare, finance, and cybersecurity have been quietly advancing synthetic data techniques for years. Why? Because they need to innovate without compromising real people’s privacy or safety.
Healthcare uses synthetic patient records to train AI models without breaching HIPAA or GDPR.
Finance simulates customers to test fraud systems and financial stress models.
Autonomous vehicles train on rare or dangerous driving scenarios.
Cybersecurity conducts simulations without touching real systems.
These sectors show us what’s possible. And while marketing has been late to the party, now is the time to catch up.
At Arima, we’ve developed the Synthetic Society: a privacy-first, high-fidelity dataset designed to reflect the entire populations of Canada and the U.S. but without relying on personally identifiable information (PII). Each synthetic individual is modeled using over 40,000 data points, drawn from census records, commercial research, purchase behavior, mobility data, and more.
This synthetic population lets marketers safely explore real-world behaviors, attitudes, and preferences. It’s a dataset built for insight, not surveillance.
More importantly, we’ve wrapped it in a platform made for agencies. Arima’s analytics tools were designed with small and independent teams in mind, and the tools offer support from pitch through to post-campaign analysis.
With tools like Persona Builder, Cross Media Planner, and Location Intelligence, agencies can define and analyze custom segments, evaluate local markets, and tailor strategies by geography and audience type. One agency, for example, used Arima to help a payday lending client identify underserved communities and develop a location-based entry strategy.
Our Marketing Mix Modeler (MMM) lets teams isolate the impact of different campaign elements. A Canadian university used MMM to shift budget from traditional media to digital and on-campus channels. The results boosted applications by 18%, and improved enrollment opportunities by 30%.
Even better, our InstaPoll feature combines synthetic data with language models to simulate survey responses. This approach offers insights where traditional polling falls short.
Synthetic data isn't just a workaround, it’s a leap forward. And with Arima, it’s finally within reach for marketers ready to lead.
Join us July 16th as we dive deeper into how Arima’s Synthetic Society is reshaping modern marketing. RSVP
About the Writer: Ray Kong has always believed that marketers should have better data: faster, clearer, and closer to the communities they serve. As a Partner at Arima, he’s turning that belief into real tools that help organizations across Canada plan smarter campaigns and make confident decisions.
Today, Ray is helping demystify synthetic populations and showing marketers how to move from assumptions to action. He shares his perspective by answering two of the questions we posed as part of this broader conversation around synthetic data and marketing strategy.
