RICE Scoring: A Practical Guide to Roadmap Prioritization

TL;DR. RICE scoring compares product ideas using reach, impact, confidence and effort: multiply the first three, then divide by effort. The framework was described by Intercom, whose version estimates effort in person-months; this guide uses person-weeks and local scales, which you should adapt to your team. Use it to make estimates and trade-offs visible before choosing roadmap work. The score is a starting point for a decision; dependencies, commitments and weak evidence still need judgement.

What does RICE scoring measure?

Start with a shared objective, such as helping new accounts complete setup. Compare ideas that serve that objective, rather than scoring every request in one list. Product Strategy Explained provides context for deciding which problems deserve attention.

Then define each input before assigning numbers.

Input	Question to answer	Working rule
Reach	How many people or accounts will encounter this change?	Use one unit and one time window
Impact	How much could it improve the chosen outcome for each person or account reached?	Use a written scale with clear anchors
Confidence	How strongly does the evidence support the estimates?	Record the evidence and use a consistent multiplier
Effort	How much total work will delivery require?	Include all contributing roles in one unit

The resulting score has meaning only within those definitions. Do not compare scores from teams using different reach windows or effort units.

How should you estimate reach?

Choose the population that could actually encounter the change. For an onboarding improvement, that might be new accounts starting setup during the next planning window.

Do not use every registered account simply because that number is easy to find. Exclude accounts that cannot access the feature, do not use the relevant workflow, or will not receive the change during the window.

Count each account once within an initiative. If an account repeats the workflow, those repetitions are not additional accounts reached.

Also distinguish exposure from success. Reach counts eligible accounts encountering the change; impact estimates the improvement it could produce. Defining reach as accounts already expected to succeed mixes those two inputs.

How should you estimate impact and confidence?

Write impact anchors around the outcome you want. The following is an illustrative local scale, not a universal RICE standard:

• 1 — Small: removes a minor obstacle while leaving the main workflow unchanged.
• 2 — Moderate: removes a meaningful obstacle to completing the target workflow.
• 3 — Large: addresses a central obstacle that currently prevents completion.

Use the same anchors across candidates. A polished description should not earn a higher impact rating than a plainly written idea.

For confidence, record what supports the estimate: observed workflow failures, customer conversations, usage records or an experiment. Separate evidence that a problem exists from evidence that a proposed solution will solve it.

An illustrative confidence convention could use 0.5 for substantial uncertainty and 0.8 for stronger supporting evidence. Choose your convention in advance. Do not treat a confidence multiplier as a measured probability unless you have evidence for that interpretation.

What belongs in effort?

Estimate total person-weeks across discovery, design, engineering, testing and rollout. Include migration or documentation work when delivery requires it.

Person-weeks measure work, not elapsed time. Two people contributing one week each represent two person-weeks, even when they work simultaneously.

Use the same delivery boundary for every candidate. Comparing a small experiment with a fully supported feature creates a misleading ranking unless that difference is explicit.

For ongoing maintenance, add a separate note. A delivery-effort denominator alone does not describe the continuing cost of supporting a feature.

What does a worked RICE example look like?

All numbers below are illustrative. Imagine a B2B product team choosing work to help new accounts finish setup.

Reach means distinct accounts encountering the change during one shared planning window. Impact uses the local scale above. Effort includes total delivery work in person-weeks.

Candidate	Reach	Impact	Confidence	Effort	RICE score
Guided setup checklist	240	2	0.8	4	96
Import error explanations	120	3	0.8	2	144
Reusable setup templates	180	2	0.5	3	60

For import error explanations:

RICE score = (120 × 3 × 0.8) ÷ 2 = 144

That candidate ranks first under these assumptions. It reaches fewer accounts than the checklist, but its estimated impact and smaller effort produce a higher score.

Now test the estimate that could change the decision. If import error explanations require four person-weeks, their score falls to 72. The checklist then ranks first.

The useful next step is to resolve the effort uncertainty. Inspect the error-handling work before committing to the ranking. A score can identify the question worth investigating as well as the work worth building.

How does RICE compare with ICE and value/effort?

ICE, created by Sean Ellis, rates impact, confidence and ease and multiplies them. Value versus effort scores each idea on expected benefit and required resources, then compares the two.

Their usefulness depends on what you need to distinguish.

Approach	Useful when	What to make explicit
RICE	Candidates reach meaningfully different populations	Reach unit, time window, impact anchors and effort boundary
ICE	You want a lighter comparison of ideas or experiments	What ease means and how ratings combine
Value/effort	A rough comparison is enough to support discussion	Which outcome defines value and what effort includes

Choose RICE when reach deserves a separate estimate. If every candidate serves the same population, the extra reach input may add little to the decision.

Choose a lighter comparison when detailed estimates would be guesswork. Writing four precise-looking numbers does not create four pieces of evidence.

For any method, document the calculation locally. Do not assume a score copied from another team's worksheet uses the same scales or aggregation rule.

Where does RICE scoring mislead?

When the candidate list ignores strategy. A high score does not establish that a problem belongs on your roadmap. Agree the objective and target customer first. How to Build an ICP can help define whose needs matter in that decision.

When reach dominates customer importance. A broadly used convenience can outrank a serious obstacle affecting a smaller target segment. Review that trade-off explicitly.

When work is mandatory. Handle security obligations, accessibility requirements and binding commitments as constraints before ranking discretionary work.

When dependencies disappear. An enabling change may have little direct reach while making another initiative possible. Assess the dependent package together, or reserve capacity for the enabling work.

When overlapping ideas claim separate benefits. Two onboarding changes may address the same obstacle for the same accounts. Do not add their estimated benefits without examining the overlap.

When uncertainty becomes false precision. Close scores with weak inputs do not justify a firm ordering. Recalculate using plausible alternative estimates and see whether the ranking survives.

How do you turn scores into a roadmap?

Use this checklist before accepting the order:

• Agree one outcome, reach unit and planning window.
• Write impact anchors before scoring candidates.
• Record evidence and assumptions beside each estimate.
• Include all delivery roles in effort.
• Identify mandatory work and dependencies separately.
• Recalculate the leading candidates using plausible alternative estimates.
• Record why the final choice follows or overrides the score.

Then turn the chosen work into a sequence. A Now, Next, Later Roadmap can communicate direction without turning uncertain estimates into fixed delivery promises.

Revisit scores when evidence changes, scope changes or the objective changes. After delivery, compare the observed result and actual effort with the estimates. Use those differences to improve the next decision.

Related reading

• Now, Next, Later Roadmap — communicate priorities while leaving room for learning.
• Product Strategy Explained — define the direction behind prioritization decisions.
• How to Build an ICP — identify the customers your roadmap should serve.