How to prioritize CRO tests without opinion soup

Every growth team hits the same wall: you have 15 possible tests, three opinionated stakeholders, and a backlog that grows faster than your sprint cycles.

The problem isn't ideas. It's prioritization. Without a framework, you end up with "opinion soup" — whoever talks loudest wins the next test slot.

Step 1: Score every idea on two axes

Impact and effort. That's it. Not seven criteria, not a weighted matrix that takes a PhD to interpret. Two numbers, each from 1 to 5.

• Impact: If this test wins, how much revenue lift will it realistically drive? A hero headline rewrite on a high-traffic page is a 5. Changing a button color on a low-traffic page is a 1.
• Effort: How many hours, people, and dependencies does this test require? A copy change is a 1. A full checkout redesign is a 5.

Plot them on a 2x2. Top-left quadrant (high impact, low effort) is your "do first" pile.

Step 2: Kill the zombie ideas

Set a rule: if an idea has been in the backlog for 60 days without moving, it gets archived. No discussion, no meeting, just archived. You can always resurface it later.

Step 3: Limit work in progress

No more than 2-3 tests running simultaneously. More than that and you can't read results cleanly, you can't allocate enough traffic, and your team is context-switching constantly.

The best teams run one high-impact test and one quick win per sprint.

Step 4: Pre-commit to a decision

Before the test starts, write down: "If variant B beats A by X%, we ship it. If not, we revert." This prevents post-hoc rationalization that kills experiments.

Step 5: Review results as a team

Set a 30-minute weekly review where you look at last week's test results together. Screen-share the data. Let the numbers speak. This builds a culture of evidence over opinion.

What we see at Fixly

Teams that follow this framework run 3-4x more tests per quarter — not because they work harder, but because they spend less time arguing about what to test.

When you can see exactly which issues have the highest estimated impact, prioritization almost does itself. The data replaces the debate.