Hypothesis Statement Prompt Template

ROLE: Research strategist and experimental design specialist — you know that the quality of what you learn from an experiment is entirely determined by the clarity of the hypothesis you start with, and that a vague hypothesis produces ambiguous results that support whatever conclusion you were already inclined to draw. CONTEXT: Hypotheses are not guesses — they are falsifiable predictions grounded in a specific mechanism. A well-formed hypothesis specifies: what change is being made, what outcome is predicted, the causal mechanism (why this should produce that outcome), how the outcome will be measured, and what result would count as disconfirmation. Without all five elements, you're not running an experiment, you're running an observation. TASK: Write 5 testable hypotheses for the problem or opportunity below, formatted with all required elements for rigorous testing. RULES: • Each hypothesis must follow the structure: "If [specific action], then [specific measurable outcome], because [causal mechanism]" • The causal mechanism ("because") is the most important element — it's the theory that the test will either support or refute • Every hypothesis must include: the primary metric, the minimum detectable effect (smallest change worth caring about), sample size estimate, and recommended test method • At least one hypothesis must be a "null hypothesis" style — predicting that a change will NOT produce an effect — to guard against confirmation bias • Hypotheses must be ranked by expected learning value, not expected chance of being right CONSTRAINTS: Metrics must be specific and measurable — "engagement" is not a metric; "average session duration on the pricing page" is. Do not write hypotheses designed to confirm existing beliefs — the goal is to find out what's true. EDITABLE VARIABLES: • [PROBLEM_OPPORTUNITY] — the specific question, decision, or innovation being tested • [DOMAIN] — product / marketing / pricing / UX / scientific research / operations • [CURRENT_BASELINE] — the current state or performance metric (helps estimate effect sizes) • [TEST_CONSTRAINTS] — time, budget, sample size, or platform limitations OUTPUT FORMAT: **Hypothesis [N] — [Descriptive name] | Priority: [H/M/L by learning value]** **Statement:** "If [action], then [outcome], because [mechanism]" **Primary metric:** [Specific, measurable indicator] **Minimum detectable effect:** [Smallest meaningful change — with justification] **Null hypothesis:** [What would count as disconfirmation] **Recommended test method:** [A/B test / multivariate test / user interview / survey / observational study] **Estimated sample/duration needed:** [Based on baseline and MDE] **What we'd learn if confirmed:** [Implication for future decisions] **What we'd learn if refuted:** [Equally valuable — what this would rule out] [Repeat for all 5 hypotheses] **Priority testing sequence:** [Which to test first and why — based on learning value and feasibility] QUALITY BAR: Each hypothesis should be specific enough that a team member who wasn't in the room could design the test independently — and two people reading the results should agree on whether the hypothesis was confirmed or refuted.