Creating Your First Experiment
Learn how to structure effective growth experiments that generate actionable insights and drive measurable business outcomes.
Why Experiment Structure Matters
A well-structured experiment eliminates ambiguity, ensures measurability, and increases your chances of generating valuable learnings - whether your hypothesis proves correct or not.
The Anatomy of a Growth Experiment
Every experiment in GrowthLab follows a proven framework:
1. Title
Write a clear, action-oriented experiment name that immediately communicates what you're testing.
- ✅ Good: "Test urgency messaging on pricing page"
- ❌ Poor: "Pricing page test"
2. Hypothesis (If/Then/Because)
Create a structured hypothesis using the If/Then/Because format:
"If we [action] for [audience segment], then [metric change], because [reasoning]."
Use the AI Hypothesis Refiner to automatically improve clarity, specificity, and measurability. The refiner scores your hypothesis across four dimensions: specificity, testability, measurability, and time-boundedness.
3. Target Metric
Identify the single primary metric you're trying to improve. Use the AI Metric Validator to check for common pitfalls like vanity metrics, lagging indicators, or confounding variables. The validator also suggests guardrail metrics and provides sample size guidance.
4. Funnel Stage
Specify where in the AARRR framework (Acquisition, Activation, Retention, Revenue, Referral) this experiment applies. This helps organize your experimentation roadmap and enables cross-funnel learning patterns.
5. ICE Score
Use the Smart ICE Editor which combines manual sliders with AI suggestions. The AI calibrates scores based on your business context, historical experiment data, and industry benchmarks - and even factors in learnings from past experiments that influenced the score.
6. Experiment Map (Optional)
Fill in the experiment map for richer AI context:
- Goal - What outcome are you targeting?
- Mechanics - Tracking metrics, success criteria, duration
- Audience - Who are you testing with?
- Resources - What tools and budget are available?
Using AI Features During Creation
| Feature | What It Does | When to Use |
|---|---|---|
| Hypothesis Refiner | Improves hypothesis clarity and structure | After writing your initial hypothesis |
| Metric Validator | Checks for metric pitfalls, suggests alternatives | After choosing your target metric |
| Smart ICE Editor | AI-calibrated prioritization scores | When scoring impact, confidence, ease |
| Evidence Panel | Shows market evidence and past learnings | When using Evidence-First ideation mode |
Pro Tips for Experiment Success
Build Momentum Early Start with high-confidence, easy-to-implement experiments. Early wins create team buy-in and establish your experimentation rhythm.
Define Success Criteria Upfront Before launching, specify exactly what constitutes a successful outcome. This prevents post-hoc rationalization and keeps experiments honest.
Use the Execution Wizard After creating your experiment, open the Execution Wizard to get a detailed, AI-generated implementation plan with step-by-step instructions, draft copy, and a pre-launch checklist.
Document Everything Even failed experiments generate valuable learnings. Thorough documentation builds institutional knowledge and prevents repeated mistakes.
