hypothesis-tree

Start by framing a specific measurable question, then brainstorm first-level hypotheses organized to be mutually exclusive and collectively exhaustive. Decompose each branch until hypotheses are directly testable, usually within three levels.

A strong hypothesis is specific, testable, falsifiable, and actionable. It should state a measurable claim like "checkout abandonment is above 70% on mobile" with clear criteria for being proven wrong and a decision tied to each outcome.

MECE stands for Mutually Exclusive, Collectively Exhaustive. Branches must not overlap with each other, and together they must cover all possible explanations, such as splitting users into new versus returning rather than vague overlapping groups.

Score each hypothesis on available evidence, test effort, and impact if true. Prioritize hypotheses with high impact, low testing effort, and little existing evidence, then record them in a testing plan with method, timeline, and owner.

Use a hypothesis tree when you need broad MECE coverage of an ambiguous question with multiple possible causes. Use Five Whys to drill deep into a single confirmed hypothesis; the two methods complement each other.