Ask a synthetic population before you ask a real one

• Hogwarts: students and staff by house · 300 peopleA Ministry inspector at HogwartsShould the Ministry appoint an inspector to oversee teaching at Hogwarts?
  Support18% ±3

  Oppose27% ±3

  Undecided56% ±3
  Read the report →
• United States by state: Census 2020–2024 · 300 peopleEnding the twice-yearly clock changeWhat would you prefer the country do about the clock change?
  Keep daylight saving time all year31% ±3

  Keep standard time all year47% ±3

  Keep changing the clocks twice a year5% ±<1

  No preference16% ±2
  Read the report →
• Company: a regional logistics firm · 300 peopleMudança para um plano de saúde com franquia altaA partir de janeiro, a empresa passaria todos para um plano com franquia alta: franquia de US$ 2.000, mensalidades reduzidas para US$ 60 por mês só para você e US$ 200 com a família, e US$ 750 por ano depositados em uma conta poupança-saúde. Você é a favor ou contra a mudança?
  Support20% ±2

  Oppose35% ±2

  Undecided45% ±2
  Read the report →

How it works

1. 01Build

   Pick a template or a real place, then shape the people.

2. 02Ask

   Up to twelve questions. A coach checks the wording.

3. 03Read

   Who leans which way, by how much, and why.

• Hypotheses, tested

  Pro and up

  Say what you expect; Studio tests it.

  More →

• Honest error bars

  Every plan

  Every share with its sampling and model error.

  More →

• Drivers, held constant

  Every plan

  Which traits move answers, the rest held constant.

  More →

• Segments and crosstabs

  Every plan

  People who answer alike, and any two questions crossed.

• Compared with real results

  Pro and up

  Your result beside published surveys and rewordings.

  More →

• A second opinion

  Every plan

  Claude answers for 50 of the same people (beta).

  More →

• Trade-offs (beta)

  Team and up

  Conjoint and MaxDiff: which options win, and by how much.

  More →

• Real places

  Pro and up

  The US, Brazil, the UK and the EU, down to counties and regions.

  More →

• Data and reports

  Pro and up

  CSV, Excel or Parquet, and printable reports.

  More →

• Free

  $0

  • · 1 population of up to 1,000 people
  • · 30,000 responses a month
  • · 1 scenario: events, months and forks
  • · 3 coach checks a month
  • · 2 checks with Claude a month
  • · 1 member
  • · Every hand-built template, a blank start, and templates of your own
  • · Reports with error bars, key drivers, segments and crosstabs
  • · PDF reports
  • · Reports shared by link
  • · Answers kept 30 days

• Pro

  $15 a month

  • · 10 populations of up to 10,000 people
  • · 750,000 responses a month
  • · 10 scenarios: events, months and forks
  • · 50 coach checks a month
  • · 15 runs a month checked with Claude by themselves, with shares adjusted toward Claude's (beta)
  • · 1 member
  • · Every hand-built template, a blank start, and templates of your own
  • · Reports with error bars, key drivers, segments and crosstabs
  • · Wording robustness checks
  • · The United States and Brazil by state, the United Kingdom by region, the European Union by country and EU member states by region, from census records and Eurostat
  • · Benchmarks against real results
  • · 2 hypotheses per study, tested on every run
  • · Downloads as CSV and Excel
  • · PDF reports
  • · Reports shared by link or publicly
  • · Answers kept 1 year

• Team

  $59 a month or $472 a year

  • · 50 populations of up to 25,000 people
  • · 3,000,000 responses a month
  • · 50 scenarios: events, months and forks
  • · 300 coach checks a month
  • · 60 runs a month checked with Claude by themselves, with shares adjusted toward Claude's (beta)
  • · 3 members
  • · Every hand-built template, a blank start, and templates of your own
  • · Reports with error bars, key drivers, segments and crosstabs
  • · Wording robustness checks
  • · The United States and Brazil by state, the United Kingdom by region, the European Union by country and EU member states by region, from census records and Eurostat
  • · All fifty US states by county and the District of Columbia and twenty-six Brazilian states by município and the Distrito Federal, from census records and Eurostat
  • · Benchmarks against real results
  • · 5 hypotheses per study, tested on every run, with significance checks
  • · Downloads as CSV, Excel and Parquet
  • · PDF reports without the Studio mark
  • · Your own logo on reports
  • · Conjoint studies (beta)
  • · Reports shared by link or publicly
  • · Reports embedded in your own site
  • · REST API with keys, webhooks, and an MCP server for AI assistants
  • · Answers kept 2 years

• Enterprise

  From about $300

  • · Populations of up to 100,000 people, as many as agreed
  • · Responses a month agreed by contract
  • · Scenarios as agreed: events, months and forks
  • · Coach checks as agreed
  • · Every run checked with Claude by itself, with shares adjusted toward Claude's (beta)
  • · Members agreed by contract
  • · Every hand-built template, a blank start, and templates of your own
  • · Reports with error bars, key drivers, segments and crosstabs
  • · Wording robustness checks
  • · The United States and Brazil by state, the United Kingdom by region, the European Union by country and EU member states by region, from census records and Eurostat
  • · All fifty US states by county and the District of Columbia and twenty-six Brazilian states by município and the Distrito Federal, from census records and Eurostat
  • · Benchmarks against real results
  • · 5 hypotheses per study, tested on every run, with significance checks
  • · Downloads as CSV, Excel and Parquet
  • · PDF reports without the Studio mark
  • · Your own logo on reports
  • · Conjoint studies (beta)
  • · Reports shared by link or publicly
  • · Reports embedded in your own site
  • · REST API with keys, webhooks, and an MCP server for AI assistants
  • · Answers kept 10 years
  • · Custom geographies