Our method

From assumptions to executable definitions

Most analytics and AI projects fail because critical assumptions stay unverified until it’s too late. Our method makes them explicit, structured, and validated — before execution begins.

Diagonal cut illustration: assumptions become structured, validated, and executed

Why analytics and AI projects go off track

  • Assumptions about data and KPIs stay implicit
  • Definitions are incomplete or inconsistent
  • Issues surface during implementation, not before
  • Teams rely on meetings instead of structured input
  • Responsibility for answers is unclear
Principles

Principles behind the method

  • Deliver value from the first project

    Most approaches require building a complete data foundation first. Our method starts from a real project and delivers value immediately, while improving every project that follows.

  • Work with real inputs, not ideal ones

    Projects rarely start with clean or complete inputs. Our method uses what already exists — documents, catalogs, spreadsheets, and people’s knowledge — and makes it actionable.

  • Make assumptions visible early

    Unverified assumptions are the main source of failure. Our method surfaces and validates them early, before they impact execution.

  • Define at the level where projects fail

    Not everything needs detail. Focus on what can delay or break delivery — data availability, grain, logic, and dependencies.

  • Use AI to accelerate, not replace judgment

    AI helps explore, structure, and coordinate information. Final decisions and validation stay with people.

  • Own it in-house, reuse it across projects

    When your team owns the definitions, they don’t get lost. Each project builds on the previous one instead of starting from zero.

A structured way to define before you build

The method starts with what already exists, surfaces hidden assumptions, and validates definitions before execution.

Context Assumptions Structure Validation Execution
  1. Start from what exists

    Use existing documents, data catalogs, and prior work instead of starting from scratch.

  2. Surface assumptions

    Identify what’s known, what’s assumed, and what’s missing across data, KPIs, and expected outcomes.

  3. Structure the definition

    Define decisions, KPIs, business questions, and outputs in a clear format that supports execution.

  4. Validate continuously

    Catch missing data, inconsistencies, and scope gaps as you define — not during implementation.

  5. Execute with confidence

    Use validated definitions to guide teams, vendors, or AI tools without costly surprises.

What changes when you use our method

Typical approach
  • Start with assumptions
  • Define inconsistently
  • Discover issues mid-delivery
  • Rely on meetings
  • Rework and delays
With our method
  • Start from real context
  • Structured, consistent definitions
  • Issues identified early
  • Targeted, asynchronous input
  • Predictable execution
Compounds over time

More valuable with every project

Validated definitions don’t disappear after delivery. They become reusable building blocks, reducing effort and improving consistency across future projects.

Project 1
Project 2
Project 3
Reuse · efficiency

Our method is built into Decision Formula

Decision Formula operationalizes the method so teams can apply it consistently across projects — without relying on ad-hoc processes or individual experience.

See how it works