Synthetic Markets
The market before the market.
A synthetic market is a model of your real market, built from official census data and personality science. You test prices, concepts and messages on it before you spend on the real one, and you see which customer types respond and why.
Official census data · personality science · zero personal data
How a synthetic market is built
Five layers. One Population Genome.
Every synthetic market is born from a country’s Population Genome: a recipe stored as mathematics, not a list of people. Markets are generated from it on demand, at the size each question needs.
- 01Census cellsOfficial statistics
The country’s population split into weighted cells — gender, age, socioeconomic level, education, region, urbanization and ethnicity — each with its real weight in the population.
- 02Identity ModelsPersonality science
The customer types of each market: personality (Big Five / OCEAN), motivations and consumer style, at four levels — national, regional, cultural and tribe, a small group tightly knit around a passion.
- 03Public data by territoryPublic sources
Income, development, connectivity, health and safety indicators for each area, so the market knows where it lives.
- 04Industry dataLicensed, where available
Licensed category data where it exists, for decisions that need the detail of a sector.
- 05Your own dataYours only
Your CRM, your studies and your results, isolated to your organization and never mixed into anyone else’s market.
Versions over time
Each enrichment creates a dated version of your market. Compare markets wave against wave and see what moved, and for whom.
Future scenarios
Describe a future — a new price level, a competitor, a change in the economy — and get a market calibrated to it, to test today what that future would do to your customers.
Your data and your scenarios never change the Genome. Every market can return to its census baseline, and no organization’s market touches another’s.
Four levels of resolution
Zoom from a country down to a tribe.
Identity Models — the customer types of a market — exist at four levels. Go only as deep as your decision needs: the same young urban explorer looks different across the country, on the West Coast, in a cultural community, and among the fans who organize their weekends around a passion.
Create your own market
Describe it in one sentence. Horizon builds it.
Say who and where. Horizon builds the market from the Genome, with its customer types and their weight, ready to test your alternatives.
- 01You describe the market: who and where.
- 02Horizon selects the census cells that match, with their real weights.
- 03The customer types in play appear, from national to tribe.
- 04You test your alternatives on that market.
Women 25–44 in São Paulo who buy natural cosmetics
B2C and B2B
Consumers, and the people who buy for companies.
Consumer markets
Consumer populations from the census genome of each country, with the customer types inside them.
- Built from each country’s official census data
- Customer types from national to tribe
- Public data by territory where it is available
Buying rooms
The people who decide inside companies, modelled as a buying room rather than a single buyer.
- The roles of the buying committee: who uses, who pays, who approves, who can block
- Seniority, sector and company size
- Country-specific practices, such as procurement and compliance norms
The science behind the decision engine
Mathematics for the market. Language only for its voice.
Fast or deliberate, as people are
Each synthetic identity decides fast and intuitively or slowly and deliberately (System 1 / System 2), depending on its personality and on what is at stake. Our Centaur cognitive router makes that call for every decision.
The market itself is mathematics
Sampling, weights and the assignment of customer types are mathematics, with zero language model, so the same market can be rebuilt exactly. Language models only give it a voice.
Better information beats more data
The engine learns which information actually improves fidelity in each kind of decision, and injects only that. A few well-chosen attributes beat dozens of raw ones.
Cover the market, don’t inflate it
Copies of one profile add no information. Effort goes where heterogeneity is: every customer type in the market, each with its weight.
Honest limits
A filter, not a forecast.
Synthetic results are read as order and comparison, never as exact percentages. They show what deserves investment and what should be dropped early.
Order between options
Which alternative wins, and which loses.
Direction of change
Where things move when a price, a message or a context changes.
Differences between customer types
Who responds, who resists, and why.
Coverage
Where synthetic markets exist today.
Census cells ready, Identity Models in progress. Canada, Poland, Ecuador, Portugal, Paraguay and more.
Census and personality studies gathered, genome build on the way.
Figures from the live genome. Coverage by country →
Klima · Market Thinker
Klima builds and tests these markets with you.
Klima, the Market Thinker, defines the market for your question, builds it from the Genome and runs the test. You read the result and decide.