A/B Test Prioritization Scorer
Turn a pile of test ideas into a ranked backlog. Score each idea on three factors and get an ordered list, so you run the highest-leverage experiments first instead of whatever shouts loudest.
Score your ideas
Impact (how much it could move the metric), Confidence (how sure you are), Ease (how cheap to build). Each factor is scored 1–10; the priority score is their average.
Test idea
Impa
Conf
Ease
Ranked backlog (ICE)
| # | Idea | Impa | Conf | Ease | Score |
|---|---|---|---|---|---|
| 1 | Add reviews to product page | 6 | 7 | 8 | 7.0 |
| 2 | Simplify checkout to one step | 8 | 6 | 4 | 6.0 |
| 3 | New pricing-page layout | 9 | 4 | 5 | 6.0 |
Score honestly, then sanity-check
- Define anchors before you score. Agree what a 3 vs an 8 means for each factor, and score from evidence (past tests, analytics) rather than gut — the frameworks are only as good as the inputs.
- Confidence and impact are guesses. ICE and PIE are fast but subjective; have two or three people score independently and compare, or use a more criteria-based method for high-stakes calls.
- A high score you can’t power is not runnable. Check that the top ideas have the traffic to reach significance in a reasonable window before committing.
For the frameworks in depth (ICE, PIE, PXL) and how to de-bias scoring, read how to prioritize A/B tests. Then size the top ideas with the traffic estimator and check they clear your minimum detectable effect.