What is the RICE scoring model?
RICE is a prioritization formula: Reach × Impact × Confidence ÷ Effort. Sean McBride described it on Intercom's blog as a way to compare product ideas without a tangle of gut feeling. The calculator uses his scales:
- Reach is how many people or events the idea affects in a fixed period, such as customers per quarter.
- Impact is how much it moves each of them: 3 massive, 2 high, 1 medium, 0.5 low, 0.25 minimal.
- Confidence is how much you trust the other estimates: 100% high, 80% medium, 50% low.
- Effort is the total work in person-months, across everyone involved. It divides the score.
What is ICE scoring?
ICE scoring multiplies three 1-10 ratings: Impact × Confidence × Ease. It's usually credited to Sean Ellis, who used it to rank long lists of growth experiments fast. Ease is the inverse of effort, so a quick win scores high. With no reach and no units, ICE takes seconds per idea, which is its strength and its weakness.
How to use the RICE and ICE calculator
- Choose RICE or ICE at the top. Each idea keeps its numbers for both, so you can switch and compare.
- Add your ideas, one per row, and score them. Load an example to see a filled-in table first.
- Read the ranked list under the table. It re-sorts as you type; equal scores share a rank.
- Export CSV for a spreadsheet, or copy the share link. The whole table lives in the link, so nothing is stored on our servers.
RICE vs ICE: which should you use?
Use RICE when you can count reach from real data, and ICE when you need a fast first cut of many small experiments.
| RICE | ICE | |
|---|---|---|
| Formula | Reach × Impact × Confidence ÷ Effort | Impact × Confidence × Ease |
| Inputs | A count, two fixed scales and person-months | Three 1-10 ratings |
| Best for | Product and growth work with usage data | Triage of many small experiments |
| Time per idea | Minutes, for the reach estimate | Seconds |
| Main risk | Reach dominates the score | Ratings drift toward favourites |
When RICE and ICE scores break down
Both models turn guesses into a number, and the number looks more certain than the guesses. The usual failure points:
- Reach swamps everything. An idea touching every visitor beats a sharper idea for your best customers, even when the second one matters more.
- Scores don't travel. Two people scoring the same idea get different numbers, so lists merged from several people rank the scorers as much as the ideas.
- Confidence gets counted twice. People already shrink impact when they're unsure, then lower confidence as well.
- Timing is invisible. Neither formula knows that a decaying page loses traffic every week you wait, or that one fix unblocks three others.
- Small differences mean nothing. A 410 and a 395 are a tie; treating the order as exact wastes the debate the score was meant to save.
The fix is to use scores to sort a long list roughly, then decide the top few by talking through them in pairs.
How LogNorm ranks moves head to head instead
LogNorm doesn't score growth work with a formula. Every finding becomes a move (a page to write, a fix to ship, a refresh), and LogNorm ranks open moves by asking head-to-head questions: if the team can only do one of these two this week, which should come first? The question weighs upside, the odds of working, effort and whether waiting makes things worse.
Each pair is asked both ways round, because the order you read two options in can bias the answer, and the answers are fitted into one strength per move. Moves still carry impact, confidence and effort signals from 1 to 5; they feed the comparisons, but priority is never calculated from them directly.
You keep the last word. Rank higher or lower on a move, or drag it in the weekly plan, and LogNorm stores the call and applies it on every later pass. The moves and ranking docs explain how new moves get placed, and the Growth Plan shows how the ranked list becomes your week.