ICE Scoring Explained: How to Prioritise Ideas by Impact, Confidence and Ease

Teams rarely suffer from a shortage of ideas. They suffer from a shortage of agreement about which idea to do first. One person argues for the biggest opportunity, another for the safest bet, and a third for the quickest win, and without a shared framework the discussion goes in circles. ICE scoring is a deliberately simple framework for breaking that deadlock. Each idea is scored on three dimensions: Impact, how much difference it would make if it works; Confidence, how sure you are that it will work as expected; and Ease, how easy it is to implement. The three scores are averaged to give a single ICE score, and ideas are ranked by that score. The ICE Scoring Calculator does the averaging and the ranking, but the value of the exercise lies less in the arithmetic than in the conversation it forces about what each dimension means.

The formula is straightforward. If Impact, Confidence and Ease are each scored on the same scale, commonly 1 to 10, then the ICE score is the sum of the three divided by three. An idea that scores 8 for Impact, 7 for Confidence and 6 for Ease therefore has an ICE score of 21 divided by 3, which is 7.0. The averaging matters: a brilliant but very difficult idea can be outranked by a good and easy idea, and a high-impact idea with very low confidence can be outranked by a moderate idea that the team is confident about. That trade-off is the point. ICE does not claim that impact, confidence and ease are equally important in any absolute sense; it claims that making the trade-off explicit is better than making it implicitly, inconsistently, in each person's head.

Here is a worked comparison that shows the method in action. Idea A is a redesign of a checkout page: Impact 8, because checkout improvements affect every buyer; Confidence 7, because similar changes have helped before but the exact uplift is uncertain; Ease 6, because design and testing will take a few weeks. Its ICE score is 8 plus 7 plus 6, which is 21, divided by 3, which is 7.0. Idea B is a major new feature: Impact 9, Confidence 4, because it depends on unproven assumptions about user behaviour; Ease 5, because it requires significant engineering. Its ICE score is 9 plus 4 plus 5, which is 18, divided by 3, which is 6.0. Idea A therefore ranks above Idea B, even though Idea B has the higher impact, because Idea A's higher confidence and ease outweigh the one-point impact gap. A team that only debated impact would have chosen Idea B; a team using ICE at least has to justify why it would override the ranking.

Defining the scales is where most ICE exercises succeed or fail. Impact should be defined in terms of the goal the team actually cares about, such as revenue, activation, retention or cost saved, and the team should agree on what a 10 means versus a 5 before scoring begins. Confidence should reflect evidence, not enthusiasm: a 9 should require data, a prior test or a strong precedent, while a 3 should signal a guess. Ease should include the full cost of implementation, including design, engineering, testing, documentation and maintenance, not just the time to write the first version. Without those definitions, two people can score the same idea very differently while both believing they are being objective, and the resulting ranking will reflect who scored rather than what was scored.

ICE also has well-known weaknesses, and a good team names them before relying on the scores. First, averaging treats the three dimensions as equally important and interchangeable, which they are not: an idea with zero confidence should arguably be blocked, not merely penalised by a third. Second, scores are subjective, so a team can game the system by inflating the scores of a favoured idea. Third, ICE ignores dependencies, strategic fit and risk beyond the confidence score, so a high-ICE idea can still be the wrong idea if it pulls the product in the wrong direction. Fourth, small differences in ICE scores, such as 7.0 versus 6.8, are meaningless given the uncertainty in the inputs, and should not be treated as precise rankings. ICE is best used to sort ideas into broad groups, such as do first, consider next, and park, rather than to produce a strict one-to-twenty ordering.

A practical ICE session therefore has a structure. Start by listing the ideas to be compared and the goal they serve, so that impact has a common reference. Score each idea independently first, then discuss the largest disagreements, because those disagreements usually reveal different assumptions about users, effort or evidence. Record the reasons for each score alongside the number, so that a future reader can see why an idea scored as it did. Rank the ideas, then apply judgement: remove or defer any idea that fails a hard constraint, such as a legal requirement or a technical impossibility, regardless of its score, and consider whether the top-ranked ideas together form a coherent plan or merely a collection of easy wins. Finally, revisit the scores after the work is done: did the impact materialise, was the confidence justified, was the ease realistic? That feedback loop is what gradually calibrates a team's scoring.

ICE is often compared with other prioritisation methods, and the comparison is instructive. RICE adds Reach to the calculation and divides by Effort, which suits product teams that can estimate how many users an idea will affect. Simple impact-versus-effort matrices make the trade-off visual but lose the confidence dimension. Weighted scoring lets a team give impact more weight than ease, which may be appropriate when speed matters less than outcome. None of these methods is objectively correct; each is a lens. ICE's advantage is its simplicity and speed, which makes it suitable for workshops, backlog grooming and early-stage brainstorming, where a perfect method would be too slow to use. Its disadvantage is precisely that simplicity, which is why it should inform judgement rather than replace it.

Used in that spirit, ICE scoring earns its place. It forces a team to ask three useful questions about every idea: how much would this matter, how sure are we, and how hard is it really? Ideas that survive all three questions deserve attention. Ideas that score highly on only one dimension deserve a second look before they consume scarce time. And ideas that nobody can score with confidence are telling the team something important: that more evidence is needed before a decision, not after it. The calculator handles the arithmetic and the ranking; the team still has to supply the honesty, the definitions and the judgement that make the numbers mean anything.

Try it next: use the ice-scoring-calculator with your own figures, then compare with the ROI Calculator and the Pricing Simulator. For background reading, see Marketing Budget Allocator Guide.