Most brands run A/B tests like they're throwing darts blindfolded. Someone has a hunch about a headline. Another stakeholder wants to try a red button. A few months later nobody can remember why that test was in the backlog or what it was supposed to prove. Sound familiar?

We've seen it dozens of times. The fix isn't more ideas — it's a system. One that forces you to compare every opportunity on the same criteria so you're always working on the thing that moves the needle most. That's where RICE comes in.

Why RICE (and why it fits CRO)

RICE stands for Reach, Impact, Confidence, Effort. It's been around in product for years. The twist we use: we apply it strictly to conversion and revenue. Every test idea gets a score. The highest scores get built first. No more "this feels important" — you've got a number.

Reach is how many users touch that part of the funnel. Impact is the lift you'd expect if you're right (we use a 1–3 scale: 0.25 for minor, 0.5 for moderate, 1 for massive). Confidence is how sure you are the test will show a clear result — usually driven by data you already have. Effort is dev/design time in person-weeks.

The formula: (Reach × Impact × Confidence) ÷ Effort. Higher = do it sooner.

How we actually score tests

Reach we pull from analytics. For a checkout test, that's everyone who hits the checkout page in a typical month. For a hero headline, it's everyone who lands on the homepage. No guessing.

Impact is where judgment comes in. A full checkout redesign might be a 1. Adding a trust badge might be a 0.25. We keep a simple internal rubric so the team stays consistent. If two people score the same idea and get wildly different numbers, we argue until we agree.

Confidence is the one that catches people out. If you've got heatmaps and session replays showing users bailing at a specific step, your confidence is high. If it's a gut-feel "maybe the button colour matters," confidence is low. We cap at 100% and rarely go above 80% for net-new ideas.

Effort we estimate in half-week chunks. One dev, one designer, two weeks = 4. Keeps the maths simple and forces the team to be honest about scope.

The roadmap in practice

Every sprint we take the top 3–5 RICE-scored ideas and slot them into the next two weeks. Nothing else gets in unless something from the list gets invalidated by new data. That discipline is what makes wins compound: you're not constantly context-switching or chasing the latest opinion.

One more thing. We re-score the backlog every quarter. Traffic shifts. New friction points show up. Old ideas that looked marginal might now be top of the list. Treat the roadmap as a living document, not a one-time plan.

If you're not scoring your tests, you're not prioritising — you're just picking. And picking is how you leave money on the table.

That's the framework. We've run it across 85+ engagements and it's the single biggest reason our CRO work keeps delivering. Start with your next ten test ideas. Score them. Build the top one. Then the next. You'll be surprised how fast the numbers add up.