RICE prioritization compares initiatives using reach, impact, confidence, and effort. Calculate a score by multiplying reach, impact, and confidence, then dividing by effort. Use the same reach period, impact scale, and effort unit for every option. The score supports a prioritization conversation; it is not a forecast of revenue or a substitute for judgment.
RICE is most useful when several discretionary initiatives compete for the same team’s capacity. It is less useful when choices have different objectives, mandatory obligations, or dependencies that make them impossible to evaluate independently. Decide what belongs in the comparison before filling in a spreadsheet.
Intercom’s original explanation of RICE identifies the four factors and their combined formula. The workshop examples and decision rules below are our own practical application of that method.
Define the decision before choosing numbers
Write one sentence: “We are choosing the next improvement to help newly activated teams reach their first useful meeting.” This gives impact a common meaning. Without that sentence, one person may score revenue, another satisfaction, and another technical elegance.
Define a reach population and period. For example, use eligible teams per quarter. Do not compare monthly users for one option with annual organizations for another. If an option affects existing accounts once, make that distinction explicit rather than treating it as a recurring benefit.
Put obligations outside the discretionary ranking. A required accessibility correction or contractual commitment needs its own decision process. A low estimated reach does not mean the team can ignore it. Similarly, identify prerequisite work before pretending that the highest-scoring feature can begin immediately.
Use a shared workshop board to collect assumptions, but keep the calculation in a table or spreadsheet that the team can inspect. Nextretro can support the discussion; it should not be represented as automatically calculating RICE scores.
The RICE formula with clear units
RICE score = reach × impact × confidence ÷ effort.
Reach is a count within the chosen period. Impact is a relative estimate tied to the agreed outcome. Confidence is entered as a fraction: 80% becomes 0.8. Effort is the estimated total contribution across relevant disciplines, expressed consistently, such as person-weeks.
Intercom’s published method uses a particular impact scale and person-months for effort. Teams adapting the units should document the adaptation and use it consistently. Switching units does not create an objectively universal score; it creates a comparable index within this decision.
| Input | Question to answer | Evidence to attach |
|---|---|---|
| Reach | How many eligible people, teams, or events? | Defined population and observation window |
| Impact | What change do we expect for each reached unit? | Research, experiment, or explicit assumption |
| Confidence | How well supported are these estimates? | Evidence quality and gaps |
| Effort | What contribution is required across roles? | Scope, dependencies, and estimate range |
Avoid decimal precision that your evidence cannot support. An impact estimate of 1.37 usually communicates less clearly than an agreed category and a note explaining why it applies.
Worked example: three onboarding improvements
Consider a hypothetical team comparing three improvements. All reach figures refer to eligible teams in one quarter, impact uses the same relative scale, and effort is measured in person-weeks. The figures below are illustrative, not Nextretro product data.
| Option | Reach | Impact | Confidence | Effort | Score |
|---|---|---|---|---|---|
| Clearer invitation guidance | 400 | 1 | 0.8 | 2 | 160 |
| Guided first-session checklist | 300 | 2 | 0.5 | 3 | 100 |
| Expanded export options | 100 | 1 | 0.8 | 4 | 20 |
The first score is 400 × 1 × 0.8 ÷ 2 = 160. That calculation does not mean the invitation change will create 160 successful teams. It means the option leads this particular index under the stated assumptions.
The checklist has lower confidence because the team has not observed where first-session setup actually fails. Instead of dismissing it, the team can fund a small research task. An uncertain high-value option may deserve learning before implementation.
Test whether the ranking survives uncertainty
Change the assumptions that matter. If invitation guidance reaches only 200 eligible teams, its score becomes 80. If the checklist’s confidence rises to 0.8 after research, its score becomes 160. The order reverses.
That result is useful. It tells the team that reach and evidence quality are decision-sensitive. Collect those facts before arguing over tiny effort differences. If a broad range of plausible values leaves the order unchanged, more estimation may not be worth the time.
Confidence also needs scrutiny. A stakeholder’s enthusiasm is not the same as strong evidence. Record whether uncertainty comes from reach, effect, delivery complexity, or all three. A single multiplier cannot explain every unknown, so keep a short narrative beside it.
Do not repeatedly adjust inputs until a favored project wins. If strategic judgment overrides the ranking, say so directly and record the reason. Honest exceptions are easier to revisit than engineered certainty.
Turn the result into a decision record
A useful record contains the selected option, assumptions, rejected alternatives, constraints, and a review condition. Name what evidence would cause reconsideration. For example: “Revisit the checklist when five observed first-session walkthroughs establish where participants get stuck.” The sample size here is a proposed research scope, not a statistical guarantee.
After implementation, compare the actual outcome with the mechanism you expected. A team retrospective can examine whether the decision process exposed uncertainty early enough. Connect the learning to the next planning round rather than using the original score to defend a sunk cost.
Browse retrospective ideas for a reflection format when the team needs to inspect its prioritization habits. The important artifact is the reasoning that future colleagues can understand.
Frequently asked questions
Is a higher RICE score always better?
It leads the ranking only within the options, objective, and assumptions used. Dependencies, obligations, risk, and strategic choices may justify a different order. Record the exception instead of disguising it in the inputs.
Can we use story points as effort?
Only if the options are estimated consistently by the same team and the measure captures the relevant work. Story points from different teams are not a shared capacity unit. A cross-functional effort estimate is often easier to explain.
What should we do when confidence is low?
Identify the assumption most likely to change the decision, then consider a small learning task. Low confidence can justify investigation; it does not automatically mean an idea is bad or should be implemented.