Plan and prioritise · Add-on
Build and run an experiment backlog
Growth comes from many small, well-chosen tests. Keep a ranked backlog of experiments and run them steadily.
About 1-2 daysEvergreenLast reviewed 5 November 2026
Who this applies to
- Skills it builds
- Data and measurementPrioritisation and roadmapping
- PM types
- Growth
- Product types
- All
- Starting points
- All
- Ways of working
- All
- Experience
- All
- Size of organisation
- All
Steps, by experience
Beginner (under 1 year)
- Collect ideas for the priority funnel stage.
- Write each as a hypothesis with an expected result.
- Rank them by expected impact, confidence and effort.
- Run the top ones, one change at a time.
- Record every result and what you learned.
Intermediate (1 to 4 years)
- Maintain a ranked experiment backlog.
- Run experiments at a steady pace.
- Record and share learning.
Experienced (5+ years)
- Track experiment velocity and win rate.
- Combine learning into bigger product bets.
- Protect long-term value with guardrails.
Who can do this
Owner: Product manager
Can be delegated to
- Data Scientist / ML Engineer or Data Analyst: Design and analyse experiments
- Frontend Engineer or Full-stack Engineer: Build variants
Reviewed by: Data Scientist / ML Engineer
Tool examples
- Notion
- Keep an experiment backlog with hypothesis, score, status and result.
- Amplitude
- Run experiments with guardrail metrics.
Definition of done
- Ranked backlog in place
- Experiments running regularly
- Results recorded
Common pitfalls
- Testing tiny changes with no real impact
- Stopping tests early
- Not sharing failures
If you lead the team
What good looks like: Growth teams keep a ranked experiment backlog and run tests steadily.
Ask your PM
- How many experiments did we run last month, and what did we learn?
- How do we rank the backlog?
Warning signs
- Experiments are run one at a time, slowly.
- Results aren't written up or shared.