ICE Prioritization (Impact, Confidence, Ease)
Score each idea on Impact, Confidence, and Ease to bubble the best ideas up without over-engineering the math
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 3
- Confidence
- 90%
ICE assigns three values to every idea: Impact (how much it moves the goals), Confidence (how sure you should be about your impact and ease guesses), and Ease (the opposite of effort — how easy it will be). Breaking the single question 'is this a good idea?' into these three produces a better discussion than 'my idea is better than yours.' Confidence is the crucial third element that stops teams from taking a gut-feel impact estimate and treating it as certain. ICE is meant as a simple heuristic to bubble the best ideas to the top — not a precise mathematical impact model.
Origin
ICE was invented by Sean Ellis (coiner of 'growth hacking' and the product/market-fit concept) — Gilad is explicit: 'I'm using the definitions of Sean Ellis. Sean invented ice.' Gilad adopts Ellis's original 'Ease' (rather than 'Effort') and prefers ICE over the RICE variant, folding Reach into Impact. He pairs ICE with his own Confidence Meter to make the C rigorous.
Core principles
- 01Splitting the judgment into Impact, Confidence, and Ease yields a better discussion than a battle of opinions
- 02Impact is estimated against clear goals — the North Star metric, top business KPI, or a local team metric — which is why goals must be clear and few
- 03Confidence assesses how sure you should be about your impact and ease guesses; without it, people give gut ideas false certainty
- 04Ease (Sean Ellis's original term) is the positive framing of effort
- 05Keep it simple — ICE is a heuristic to bubble the best ideas up, not a precise impact calculation; it always ends up as a spreadsheet
How to run it
- 1
Estimate Impact against clear goals
Assess how much the idea will move your defined goals (North Star metric, top business KPI, or a local team metric). Be explicit about which metric you're scoring against.
Pro tip Impact is the hardest to estimate — it may be gut feel, a back-of-envelope spreadsheet, or grounded in an actual test; all are legitimate at different confidence levels.
Watch out If you interviewed 12 customers and only one liked the thing, reduce your impact estimate accordingly rather than clinging to the guess.
- 2
Estimate Ease (not effort)
Assess how easy or hard the idea will be to build — the positive inverse of effort. Colleagues can help sharpen this estimate.
Pro tip Using 'Ease' keeps both dimensions oriented so that higher is always better, matching Sean Ellis's original definition.
- 3
Set Confidence honestly
Rate how sure you should be about your impact and ease guesses, using the Confidence Meter to ground the number in actual evidence class rather than personal conviction.
Pro tip Don't overcomplicate — a shared spreadsheet with three columns is enough to bubble the strongest ideas to the top.
Watch out People give a gut idea a high impact and then a high confidence too ('it's an eight and I'm pretty convinced'), which subverts the whole system.
In the wild
Gilad found teams estimate impact from instinct then assign themselves high confidence, defeating ICE's purpose. He built the Confidence Meter specifically as the rigorous C, so 'behind every terrible idea someone thought it was great' earns 0.01, while a live test earns a high score.
→ ICE stays simple but the confidence dimension becomes evidence-grounded, so weak ideas can't hide behind conviction.
Teams use ICE to objectively defer weak ideas: acknowledge the idea, then show how its impact, ease, and confidence stack up against better alternatives. Gilad calls it a 'classic PM move' — 'that was a great idea but what about this better idea.'
→ Rejection becomes an objective comparison rather than a personal veto, making no-decisions gentler and more defensible.
Common mistakes
Giving a gut idea both high impact and high confidence
Estimating impact from instinct and then self-assigning high confidence subverts ICE entirely; the confidence score must reflect the strength of actual evidence, which is what the Confidence Meter enforces.
Over-engineering the math
Piling on math formulas to estimate impact precisely misses the point — ICE is a simple heuristic to bubble the best ideas up; it doesn't need to be a perfect estimate and normally lives in a plain spreadsheet.
Is it for you?
Best for
Product teams flooded with ideas that waste time in opinion battles and want a lightweight, consistent way to rank ideas and defensibly say no
Not ideal for
Teams seeking a precise, defensible numeric forecast of impact — ICE is deliberately a rough heuristic, not a rigorous model
From the transcript
“in the book I suggest using ice impact confidence and ease”
“I'm using the the definitions the of Sean Alice Sean invented interesting ice”
“the impact tries to assess how much impact it will have on the goals”
“is is basically the opposite of effort how easy or hard it's going to to be”
“there's the third element which is confidence which tries to assess how sure are we or should we be about our first guesstimates”
“I prefer ice because I prefer to fold the reach into the eye”
From the episode
Becoming evidence-guided
Itamar Gilad (Gmail, YouTube, Microsoft)