Context
Use this to convert unstructured idea lists into a shippable experimentation roadmap.
Procedure
- Compile experiment ideas from analytics and qualitative inputs.
- Write hypotheses in “If we [change], then [metric] improves because [reason].”
- Estimate impact, confidence, and effort scores.
- Rank backlog and identify dependencies.
- Define instrumentation and success/failure criteria.
Output Format
# Experiment Backlog
| Experiment | Hypothesis | Metric | Impact (1-5) | Confidence (1-5) | Effort (1-5) | ICE | Owner | Status |
| ---------- | ---------- | ------ | ------------ | ---------------- | ------------ | --- | ----- | ------ |
| | | | | | | | | |
## Top 3 Next Experiments
1.
2.
3.
## Instrumentation Notes
- Event names:
- Segment filters:
- Success threshold:
QA Rubric (scored)
- Hypothesis quality (0-5): testable and causal.
- Prioritization quality (0-5): scoring reflects constraints and upside.
- Measurement quality (0-5): success criteria are clear.
- Delivery readiness (0-5): owner and dependencies are defined.
Examples (good/bad)
- Good: “Changing CTA copy targets activation rate with defined lift threshold and event tracking.”
- Bad: “Try a redesign and see what happens.”
Variants
- Startup speed variant: weekly 1-week tests.
- Enterprise variant: risk-checked experiments with legal review.