Our core product · In development

Meet PROSE.

Product Optimization and Simulation Engine. A platform we’re building to help teams define experiments, measure outcomes, and investigate the effects of product decisions.

Discuss a pilot Explore the product

PROSE is not yet generally available. We’re inviting conversations about pilot use cases and integration needs.

01 · Experiment

Give each change a fair test

Manage the experiment and the assignment of customers to control and treatment groups.

A consistent experiment workflow

PROSE’s development version includes an experiment management console and a treatment-allocation API. Teams can define control and treatment proportions, manage experiment settings, and retrieve customer assignments.

Keeping an existing customer’s assignment consistent gives measurement a stable foundation. The next step is to connect those assignments to the outcomes the experiment is intended to change.

Development scope

  • Experiment creation and management
  • Control and multiple treatment groups
  • Persistent customer-to-group assignments
  • Analysis jobs linked to experiments and metrics

02 · Measure

Connect experiments to outcomes

Define which metrics matter and how they should be measured.

Metrics with context

The development workflow supports metric groups, customer- or session-level measurement, and analysis jobs. The aim is to keep the definition of an outcome connected to the experiment being evaluated.

Our broader direction is to connect these measurements to a company’s strategic goals. Metrics alignment workshops help teams establish that foundation while the product evolves.

Explore metrics alignment

Questions to settle first

  • What decision will this metric inform?
  • What counts as a conversion or an active customer?
  • Which population and time window are relevant?
  • Which other outcomes should we watch?

03 · Investigate

Make the assumptions explicit

Estimate the effect of a change using a stated model of how the relevant variables relate.

Causal modeling in development

PROSE includes a development workflow for specifying causal graphs, selecting a treatment and outcome, and estimating effects from experiment data or an uploaded dataset.

The estimation workflow includes diagnostic checks that challenge aspects of the analysis. Estimates depend on the data, the study design, and the assumptions in the model; a result is not an automatic proof of causation.

From a question to an estimate

  • Specify the proposed relationships between variables
  • Choose the intervention and outcome of interest
  • Estimate average or varying treatment effects
  • Review uncertainty and diagnostic checks

04 · Simulate

Explore before using live traffic

Use synthetic customer journeys to exercise an experiment’s allocation and measurement workflow.

A controlled development environment

PROSE’s simulator generates synthetic traffic through configurable customer journeys, including browsing and purchasing behavior. This helps test the mechanics of an experiment and its downstream analysis.

Synthetic results reflect the assumptions used to generate them. They help validate a workflow; they are not a forecast of customer behavior or evidence of commercial impact.

Example pilot question

Could a simpler onboarding journey help more customers complete registration?

  • Define completion and supporting metrics
  • Specify the control and proposed alternative
  • Exercise the workflow with synthetic events
  • Agree on the requirements for a real experiment

Built for underserved markets

We’re shaping PROSE around the budgets, data readiness, and operating constraints of businesses that have been underserved by traditional decision tools.

Start with a useful decision

Our initial focus is financial services, with related use cases in retail, e-commerce, telecoms, and media. We want to understand the decisions teams need to make before prescribing a tool or integration.

Develop with interested teams

Pilot conversations will help define data requirements, deployment needs, support, and commercial terms. Broader metrics alignment and scenario analysis remain part of the product direction as development continues.

Help shape a useful pilot

Bring a product question and a description of your current measurement setup. We’ll discuss whether PROSE is a fit.

Discuss PROSE