| name | niopd-pd-experiment |
| description | Designs product experiments (A/B tests, pilots). Use for feature validation, optimization, or data-driven decisions. |
Experiment Design Skill
This skill designs rigorous product experiments for validation.
Theoretical Foundation
Experiment Components
- Hypothesis: What we believe will happen
- Variables: What we change (independent) and measure (dependent)
- Control: Baseline for comparison
- Sample Size: Statistical significance requirements
- Duration: How long to run
Instructions
Step 1: Define Hypothesis
"We believe [change] will cause [effect] for [users]."
Step 2: Design Test
| Element | Control | Treatment |
|---|
| [Variable] | [Baseline] | [Change] |
Step 3: Determine Metrics
- Primary: [What success looks like]
- Secondary: [Supporting metrics]
- Guardrails: [What not to break]
Step 4: Calculate Sample Size
Based on:
- Minimum detectable effect
- Baseline conversion rate
- Statistical significance (typically 95%)
Step 5: Plan Duration
Consider:
- Sample accumulation rate
- Weekly patterns
- External factors
Step 6: Generate Document
File path: 04-plans/[YYYYMMDD]-experiment-v0.md
Output Specifications
- File Naming:
[YYYYMMDD]-experiment-v0.md
- Location:
04-plans/
- Template:
references/experiment-template.md
