Score every test idea across 8 variables. Add up the points. Run highest to lowest. No gut-feel, no guessing, just ranked priority.
Where on the site does this test run?
| Location | Score |
|---|---|
| Sitewide | 30 |
| Multiple locations | 30 |
| Product Detail Page | 20 |
| Cart | 20 |
| Bundle | 20 |
| Any Checkout step | 20 |
| Homepage | 10 |
| Product Listing Page | 10 |
| Edge Case | 0 |
Tests that touch more of the funnel score higher. Sitewide and multi-location tests reach every visitor, maximum exposure, maximum potential impact.
What percentage of your visitors will actually see this test?
| Exposure | Score |
|---|---|
| 100% | 30 |
| 75–99% | 20 |
| 50–75% | 10 |
| Under 50% | 0 |
A test only 20% of visitors see can't move the needle on overall RPV. Prioritize tests with broad reach.
What's backing this test idea?
| Research Basis | Score |
|---|---|
| All 4 (AI + Follow-up + Qual + Quant) | 50 |
| AI + Follow-up Test | 40 |
| Follow-up Test + Qualitative | 30 |
| Follow-up Test + Quantitative | 30 |
| Follow-up Test | 20 |
| AI | 20 |
| Qualitative + Quantitative | 20 |
| Qualitative only | 10 |
| Quantitative only | 10 |
The more evidence behind a test, the higher it scores. A follow-up to a test you already ran, where you know something works and you're dialing it in, is the highest-confidence bet outside of having all four signals aligned.
How hard is this to build?
| Difficulty | Score |
|---|---|
| Very Easy | 40 |
| Easy | 30 |
| Average | 20 |
| Hard | 0 |
Ease gets the highest individual weight in the formula. A high-impact, low-effort change always beats a high-impact, high-effort change when everything else is equal. Build the easy wins first.
Which devices does this test run on?
| Device | Score |
|---|---|
| All devices | 30 |
| Mobile only | 20 |
| Desktop only | 10 |
Mobile-first, all-device tests reach more visitors. Desktop-only tests have a smaller addressable audience in most niches.
How is this change being deployed?
| Type | Score |
|---|---|
| A/B Test | 20 |
| Direct Change | 10 |
A/B tests generate learning, not just change. A direct change tells you what you did. An A/B test tells you what worked and by how much, that data compounds over time.
What is this test trying to do?
| Intent | Score |
|---|---|
| Fix | 40 |
| Optimization | 30 |
| Redesign | 25 |
| Addition | 20 |
| Merchandising | 15 |
| Offer | 15 |
| Content | 10 |
| UX System | 10 |
Fixes score highest because they address something broken. A broken thing losing you money every day is a higher priority than adding something new that might make you more. Fix first, optimize second, build third.
Where does this test sit in the overall project?
| Workstream | Score |
|---|---|
| Baseline Fix | 40 |
| Experiment | 30 |
| Structural Addition | 20 |
| System Definition | 10 |
Baseline fixes, the foundational things that should have been right from the start, always run first. No point running experiments on a broken foundation.