Writing Good Hypotheses

Learn the framework for crafting testable, measurable hypotheses that drive successful experiments and actionable insights.

The If/Then/Because Template

All hypotheses in GrowthLab use the structured If/Then/Because format:

"If we [specific action] for [target audience], then [measurable metric change], because [reasoning/insight]."

The Four Essential Components

1. Specific Action: What Are You Testing?

Poor: "Improve the homepage" Strong: "Add a customer testimonial carousel with 5 reviews above the fold"

2. Measurable Outcome: What Metric Will Move?

Poor: "Users will like it more" Strong: "Increase demo request rate by 15%"

3. Target Audience: Who Is This For?

Poor: "Users" Strong: "First-time visitors arriving from organic search"

4. Rationale: Why Will This Work?

Poor: "Because it looks better" Strong: "Because social proof reduces purchase anxiety for high-consideration B2B purchases"

Hypothesis Quality Checklist

• Is the action described precisely enough for implementation?
• Can success or failure be objectively measured?
• Is the expected impact quantified?
• Can this be tested within 2–4 weeks?
• Is the target audience explicitly defined?
• Is there a clear rationale explaining why this should work?

Using AI to Improve Your Hypotheses

The AI Hypothesis Refiner automates hypothesis improvement:

1. Write your rough idea in the If/Then/Because textarea
2. Click "Refine with AI"
3. Review the clarity score (aim for 80%+)
4. Check the structured breakdown: Action, Segment, Metric, Reasoning
5. Address any warnings before launching

The Experiment Quality Score also evaluates hypothesis clarity as one of its four dimensions, giving you a second perspective on quality.

Common Mistakes to Avoid

All documentation