GUIDES · DECISION

Make decisions with synthetic populations.

A decision cannot be reduced to an average.

A synthetic population reveals differences, hesitation, opposition, groups and tipping points that disappear in an average response.

The goal is not to mechanically predict what will happen. The goal is to see more clearly what can make a decision succeed, stall or reverse.

See before you decide.

THE SHIFT

Move from the average to the internal structure of the population.

An average can tell you whether a proposal is broadly liked or disliked. It tells you far less about why, by whom, under what conditions and with what degree of stability.

  • Average
  • Differences
  • Groups
  • Weak signals
  • Tipping points

STABLE

Those whose position changes little.

HESITANT

Those whose position depends on context.

OPPOSED

Those for whom the decision creates strong friction.

MOBILISABLE

Those who may shift the broader dynamic.

What matters is not only how many say yes or no. It is the structure of that yes and that no.

THE SCOPE

Test decisions that will encounter people.

A synthetic population becomes useful whenever a decision depends on how people will understand it, compare it, accept it, reject it or adapt to it.

  • Pricing
  • Launch
  • Offer
  • Brand
  • Transformation
  • Infrastructure
  • Reform
  • Communication
  • Crisis
  • Acceptability

The decision changes. The method stays the same.

THE METHOD

Five steps to turn a decision into an investigation.

  1. FRAME

    Turn the decision into a testable question.

    What choice must be made? Which alternatives should be compared? What would change the decision?

  2. TEST

    Build several scenarios and submit them to the population.

    Do not only ask “what do you think?”. Compare options, framings and consequences.

  3. READ

    Observe differences, groups, hesitation and friction points.

    The value lies in the structure of reactions, not only in their average.

  4. CHALLENGE

    Compare scenarios and look for what resists.

    Test a variant, investigate a group, reopen a question or explore an unexpected reasoning path.

  5. DECIDE

    Turn the investigation into an arbitration.

    Simulation informs the decision. It does not replace the decision-maker.

THE OUTPUT

An investigation dossier, not a single answer.

The result should make it possible to understand what happened in the simulation, not only to read its conclusion.

  • 01 · Summary
  • 02 · Compared scenarios
  • 03 · Groups and typologies
  • 04 · Friction points
  • 05 · Tipping points
  • 06 · Individual reasoning

SUMMARY

What changes the decision.

SCENARIOS

What works better or worse.

TYPOLOGIES

Which groups react differently.

FRICTIONS

What blocks, divides or weakens.

TIPPING POINTS

What can reverse a dynamic.

REASONING

Why some individuals react the way they do.

Every conclusion should be traceable back to the reactions that produced it.

AN EXAMPLE

One decision. Several scenarios. Several trajectories.

Launch a new offer.

  • DECISION
  • Several scenarios
  • Synthetic population
  • Reactions
  • Groups
  • Trajectories
  • Arbitration

SCENARIO A

  • Adoption
  • Hesitation
  • Rejection

SCENARIO B

  • Adoption
  • Hesitation
  • Rejection

SCENARIO C

  • Adoption
  • Hesitation
  • Rejection

The goal is not to automatically declare a winner. The goal is to understand what differentiates the trajectories.

STATED LIMITATIONS

What the method should not become.

A MECHANICAL PREDICTION

A synthetic population explores plausible reactions and their differences. It does not mechanically guarantee the future.

A REPLACEMENT FOR REALITY

When real data, observations or studies are available, they remain essential elements for calibration and challenge.

AN AUTOMATIC DECISION

The system informs arbitration. The responsibility for deciding remains human.

FAQ

Frequently asked questions.

How many scenarios should be tested?

There is no universal number. Enough variants should be tested to understand what genuinely changes reactions, without artificially multiplying scenarios.

Do we need to query the whole population?

No. The level of querying depends on the question, the population size and the level of precision required.

Can a specific group be investigated in greater depth?

Yes. A typology, segment or individual can be explored in greater depth when its behavior becomes decisive for the decision.

Can this approach be combined with real-world research?

Yes. Synthetic populations can be used before, during or after real-world research to prepare, deepen or challenge the results.

Does the system make the decision?

No. It produces evidence, scenarios and reasoning to inform arbitration. The final decision remains human.


YOUR NEXT DECISION

Test it before you make it.

Describe the decision you want to explore.

Talk about your project →

What if you tested
your next decision?

State your decision. See the future it produces.

Explore the product