Simulating the world with 5.5 billion synthetic audience
Sixty-one markets, each built from its own census and kept calibrated against real survey waves. Describe any audience in a sentence and it is ready the same day, in the language it speaks.
For a century only the largest brands could ask what real people think. Now any team can ask 5.5 billion simulated people, calibrated against real human data, and trust the answer.
A niche or international segment means a panel vendor
- Week 0Find a vendor with reach in the market. Negotiate incidence and price.
- Week 3Recruitment stalls on the hard segment. Quotas get relaxed.
- Week 6A sample you cannot re-contact, in a language you translated twice.
Describe the segment. It exists by the afternoon.
A simulator is only as good as
its calibration loop.
A prompt gives you an average voice frozen at training time. The foundation is rebuilt and re-scored as the world moves.
- The census sets the shape
Age, income, education, region, household. Each market is rebuilt from its own official statistics as new releases land, down to states and prefectures.
- Real answers keep it honest
Tens of thousands of real survey measurements are on record across markets. Where a real answer exists it grounds the simulated one; nearby topics calibrate the rest.
- Real people close the loop
Recurring real-people interview waves and fresh behavioral and macro data re-calibrate attitudes between census years.
- Scored in public
Every calibration wave is measured on the open benchmark, blind and reproducible, before it ships.
A 6% group is 6% of the audience, and it answers like itself.
Because the foundation is a census, minorities are not rounded away. Cross-tabs by age, region or income come out of the same population the country reports.
The read below is the Japan slice of an Asia-Pacific snack study: over-55s are the group that says no, and they are exactly as large as they are in the country.
| 18–34 | 35–54 | 55+ | |
|---|---|---|---|
| Yes | 47% | 31% | 14% |
| Maybe | 33% | 39% | 28% |
| No | 20% | 30% | 58% |
Hit cell: the answer the question was really about. Where a real survey measurement exists it grounds the cell, and every cell opens into the people behind it.
Premium audiences, maintained by iMario.
Hard-to-reach groups, each built by iMario from real data about that group. Each one is a one-time purchase for your workspace, and studies on it use credits as usual.
- US Physicians, United States: About a million physicians in active practice, a little under two in five of them women.
- China Doctors, China: About five million licensed doctors, working everywhere from community clinics to top-tier hospitals.
- UK Doctors, United Kingdom: Doctors in hospitals and general practice, balancing patient care with training, workload and professional responsibility.
- US New Parents, United States: Parents with a baby or young child at home.
- US High Earners, United States: High-income Americans who think about money in terms of net worth, taxes and time.
- US Professionals on AI at Work, United States: College-educated working Americans across offices, creative fields and science, talking about AI at work.
- China Urban Graduates, China: Degree holders in Chinese cities who read, watch films and write reviews, and judge content and brands for themselves.
- Japan Consumers, Japan: Japanese adults who say what they really think in everyday conversation.
- Japan Parenting Families, Japan: Parents raising children in Japan.
New official audiences are added regularly.
61 markets, by region.
Each market is built from its own statistical office and national surveys. Every audience carries the sources it was built from.
- United States
- Canada
- Brazil
- Mexico
- Argentina
- and 1 more
- United Kingdom
- Germany
- France
- Spain
- Italy
- and 28 more
- United Arab Emirates
- Saudi Arabia
- South Africa
- Egypt
- and 4 more
- China
- Japan
- South Korea
- India
- Indonesia
- and 9 more
markets, each built from its own official census and survey statistics.
people in the simulated foundation.
agreement with real survey distributions across 11 benchmarked markets, against a 93% repeat-survey ceiling.
Sources · iMario Synthetic Audience Benchmark, public method and raw files · Pew Research Center, survey methodology · National statistical offices and national attitude surveys · Updated 2026-09-17, reviewed by iMario research
61 markets across the Americas, Europe, the Middle East and Africa, and Asia Pacific, including the US, Canada, the UK and EU, China, Japan, Korea, India, Southeast Asia, Brazil, Australia and the Gulf. Each market carries the sources it was built from, in the app.
