New simulation

Set up a run

Start from a written brief, a template, a blank sheet, or a campaign you already ran.

0 words

Pick the closest shape. You get a full set of segments, a market and three offer variants, all editable on the next screen.

Start with two placeholder segments and one offer, then define everything yourself — names, traits, channel preferences, peak hours, market conditions.

What are you selling?
How many units are at risk?

Paste the headline numbers from a campaign you already ran. We tune the population until the simulation reproduces what actually happened, then everything after that is calibrated to your business. Aggregates only — no customer records.

Past campaign

Open rate 18.0% · take rate 2.81%

Runs in isolationNo connection to production systems.

Aggregates onlySynthetic customers from segment behaviour, never records.

10,000 journeys per runTypically under a minute.

Reading your brief

Here's what we picked up.

Your words

Highlighted where something was extracted.

Extracted

Each fact traced back to the phrase it came from.

Approve the setup

We built a world. Change anything before it runs.

Nothing here is locked. Every segment, condition and offer is a guess from your brief — your corrections are what make the simulation yours.

Audience

Market

Conditions shift how deeply every segment needs to be discounted.
Live events

Offers

Three variants drafted from your brief. Edit, or leave them.
Uncalibrated — no past campaign loaded
Simulating

Every customer decides for themselves.

Running 0 / 10,000
Result

What the simulation says.

Variants

Ranked by revenue net of dilution, not by units moved.

Where the money went

Winning variant, gross through to net.
Revenue bridge
Best time to send

By segment

Winning variant.

Ten of them, in their own words

Individual simulated decisions.

Prototype. Behaviour models are illustrative. In production each population is calibrated against the business's own past campaign results before any forecast is issued, and the readout is relative — it ranks variants and finds the ones that destroy margin. It does not predict absolute revenue.