Help & support
Open gardenHome

The Lean Loop

Validated learning, one build–measure–learn turn at a time.

Last updated 2026-08-13

The lean loop optimizes for one thing: how fast you can learn whether the bet is right. Each season is one third of a turn — build the smallest thing that tests the hypothesis, measure what actually happens, and let the learning decide what the next turn builds.

Unlike a staged plan, this lineup is a cycle. Advancing the trunk out of Learn does not end the project; it starts the next turn with everything you now know carried forward.

  1. 1

    Build the smallest test

    Code session

    Name the riskiest hypothesis, then build only what tests it — a Code session or the leanest artifact that could possibly disprove you. Scope is the discipline here.

    You’re here when: The Build season is open and the current hypothesis has no test in the world yet.

  2. 2

    Measure — in the world

    Outside Leefy

    Ship the test to real people and collect what happens: numbers, replies, behavior. This happens outside Leefy; paste the raw results back onto the tree before they fade.

    You’re here when: The Measure season is open and the test is live but its results are not on the tree.

  3. 3

    Learn, then decide the turn

    Harvest

    Synthesize the results into one banked learning: held, or refuted, and what that changes. Then advance the trunk — persevere into the next build, or pivot the hypothesis.

    You’re here when: The Learn season is open and raw results are waiting to become a verdict.