RICE Scoring Explained (and When Its Math Lies) (2026)
RICE scoring is a four-variable formula that ranks initiatives by expected value per unit of effort. The math is simple: Reach times Impact times Confidence, divided by Effort. Sean McBride built it on the Intercom growth team in 2018 to settle internal arguments about which features to ship next. Most product, marketing, and operations teams now use some version of it.
The framework's strength is also its weak point. Multiplying by Confidence creates the illusion of math when three of the four inputs are still guesses. This guide explains how to score each variable honestly and walks through a worked example. It also stress-tests the model so you can see how a small confidence error reshuffles your backlog.
Quick answer: what RICE scoring is
RICE scoring is a prioritization formula: Reach times Impact times Confidence, divided by Effort. Each variable gets a numeric score, the formula produces a single number, and initiatives are ranked highest to lowest. The framework was invented at Intercom in 2018 to break ties between competing roadmap items.
Reach is the count of people a project will affect in a defined window. Impact is the size of effect on each person, scored on a 0.25 to 3 scale. Confidence is the percentage certainty in the other three numbers. Effort is the work required, usually counted in person-months. Higher reach, impact, and confidence raise the score; higher effort drops it.
Stress-test your scores in the widget
Type your own initiatives into the table below, or edit the three sample rows. The widget computes RICE in real time. Then drag the stress slider to drop confidence across every initiative by 5 to 40 points. Watch which ones lose the most score, and whether the top of the ranking flips.
Most product teams overestimate confidence by 20 to 30 points on first read. The widget makes that error visible.
Confidence Stress-Tester
Three sample initiatives below. Edit the values, then drag the stress slider to see what happens to the ranking when your confidence was higher than reality. Most teams overestimate confidence by 20 to 30 points.
Stress test: drop confidence by
0 pointsSlide right to simulate confidence calibration error. Watch which initiatives lose the most score, and whether the top of the ranking flips.
Live ranking
Score in the doc, ship in the board.
Rock pairs a task board with chat and notes, so RICE scores, the debate behind them, and the work itself live in one space. One flat price, unlimited users.
How to score the four variables
Each variable has a canonical scoring scale that Intercom published in the original post. Teams that drift off these scales lose the ability to compare scores across projects. That comparison is most of the point of using RICE.
| Variable | How to score it (Intercom canonical scale) |
|---|---|
| Reach | Count of people affected in a defined window. Pick a unit and stick with it: users per quarter, leads per month, signups per release. Pull from analytics, not from memory. |
| Impact | Massive = 3. High = 2. Medium = 1. Low = 0.5. Minimal = 0.25. Score per affected person, not in aggregate. Use the strongest available proxy: conversion lift, retention delta, support-ticket drop. |
| Confidence | High = 100%. Medium = 80%. Low = 50%. Under 50% means you should not score the item yet; run a discovery spike first. |
| Effort | Person-months of work across product, design, and engineering. Round to the nearest half-month for items under three months, full month for anything longer. |
The most common mistake is reusing a Reach number from a different window. A feature that touches 5,000 users per quarter is not equivalent to one that touches 5,000 users per year. Pick one window for the whole prioritization round and apply it to every initiative.
Confidence is the variable that quietly does the most damage. Itamar Gilad, who favors the related ICE framework, explains the tension well.
"Many ICE practitioners, me included, argue that Reach is simply a component of Impact, and not necessarily a component you always want to factor. An idea that only impacts a small, but very highly-engaged subset of users (power users) can be of high impact although it's low in reach." - Itamar Gilad, ICE Scores
A worked example, step by step
An agency considering three roadmap items for the next quarter scores each one using the canonical scale. The team uses a quarterly Reach window and counts effort in person-months across two engineers and one designer.
| Initiative | Reach | Impact | Confidence | Effort | RICE |
|---|---|---|---|---|---|
| Client comment threads on deliverables | 800 | 2 | 60% | 4 | 240 |
| Rebuilt reporting dashboard with custom metrics | 400 | 3 | 90% | 6 | 180 |
| Three-step onboarding wizard for new clients | 1,000 | 1 | 80% | 4 | 200 |
The math: 800 times 2 times 0.6 divided by 4 gives 240 for the comment threads. The dashboard, despite its higher impact and confidence, lands at 180 because of a longer build. The onboarding wizard hits a broad audience but at low impact, landing at 200. The team would ship in that order.
If the team had only looked at impact, the dashboard would have led. If only effort, the wizard. RICE is meant to surface this kind of cross-variable tradeoff. The risk is that the numbers feel more solid than they are, which is the next section.
Where RICE's math lies
The formula multiplies four numbers, three of which are estimates. Multiplication amplifies error. A 20-point miss on Confidence does not feel large in conversation, but in the math it changes the score by 20 percent. In a competitive backlog where the top three items sit within 10 percent of each other, that error reshuffles your priorities.
Jens-Fabian Goetzmann put the problem plainly after years of using the framework.
"Tweaking that assumption slightly will dramatically change the overall score of an idea." - Jens-Fabian Goetzmann, The Problem with Prioritization Frameworks
The five honest weaknesses are worth naming, because every team using RICE will hit them eventually.
- False precision from confidence Multiplying by a percentage gives the formula the look of math, but Confidence is a gut score. The widget above shows how a 20-point shift in confidence can flip the top of a ranking. Treat scores within 15 percent of each other as a tie, not a decision.
- Reach is the easiest number to fudge Teams pull Reach from whatever analytics view supports the case they want to make. Pick the window once for the entire round, source the number from a defined report, and write the source next to the score so the next person can audit it.
- Impact is subjective and not benchmarked "High" versus "Massive" is judgment, not data. Teams that score Impact in isolation usually compress everything to 1 or 2. Calibrate by scoring three known features after launch first, then anchor new estimates against them.
- Effort underestimation is the rule, not the exception Person-months is the unit Fred Brooks discredited in 1975 because the work does not scale linearly with headcount. Use effort as a relative comparison between items in the same round, not a hard project estimate.
- No room for dependencies, tech debt, or strategic bets RICE assumes initiatives are independent and benefits are tactical. Foundational work, debt cleanup, and "we need to be in this market in 18 months" bets always score low. Keep a separate lane for those, or they will be cut every quarter.
The original framework's author, Sean McBride, acknowledged this directly in the post that introduced RICE.
"RICE scores shouldn't be used as a hard and fast rule. There are many reasons why you might work on a project with a lower score first." - Sean McBride, Intercom Blog
RICE vs ICE vs MoSCoW vs WSJF
Four prioritization frameworks come up repeatedly. They are not interchangeable. RICE assumes you can quantify reach and effort. ICE drops the Reach factor and runs faster. MoSCoW sorts items into categories rather than ranking them. WSJF, from SAFe, swaps Effort for Job Size and adds a Cost of Delay factor for time-sensitive work.
| Framework | Formula | Best for |
|---|---|---|
| RICE | Reach times Impact times Confidence, divided by Effort. | Product roadmaps where you can attach numbers to reach. Quarterly planning across 10-plus items. |
| ICE | Impact times Confidence times Ease (or 1/Effort). | Faster scoring rounds, growth experiments, marketing campaigns where reach is uncertain or always large. |
| MoSCoW | Categories: Must, Should, Could, Won't. | Stakeholder agreement on release scope. Faster and more political than RICE; no ranking inside categories. |
| WSJF | Cost of Delay divided by Job Size. | SAFe environments with regulatory deadlines or time-bound opportunities. Surfaces work that loses value if delayed. |
| Eisenhower | 2x2 grid: Urgency vs Importance. | Daily task triage by a single person. Not designed for team backlog scoring. |
The honest move is to use two frameworks in parallel for different decisions. MoSCoW lets stakeholders agree on what ships in a release. RICE ranks items inside the "Must" and "Should" categories. Eisenhower handles the daily triage. WSJF flags the work that has a clock attached to it. None of them is the right answer for every kind of decision.
When not to use RICE
RICE is a quantitative tool. It works when the inputs are real numbers and the team trusts those numbers. Several common situations break those assumptions, and forcing RICE on them gives a false sense of rigor.
Skip RICE when reach is genuinely uncertain or the user base is small. A team of five customers cannot be scored on Reach without absurd math. Use ICE or qualitative judgment instead. Skip it when the work is foundational. Refactors, migrations, and tech-debt cleanup always lose to user-facing features in RICE math, even when the foundational work is urgent. Run those in a separate lane.
Skip it when the decision is political. If three executives need to agree on what ships next quarter, the conversation is not about scores. It is about who gets credit and what message the roadmap sends. MoSCoW or a stack-rank workshop reaches agreement faster than any formula. Skip it when the team is brand new. Scoring requires calibration across past work; teams in their first six months have nothing to anchor against.
RICE for agency teams
Most RICE writing assumes one team, one product, one goal. Agencies do not work that way. A designer ships work across five clients. An account manager juggles eight retainers. The four variables behave differently when the work spans clients, and the framework does not adjust for it on its own.
Reach is the variable that breaks first. An 800 for Client A, a direct-to-consumer brand with 200,000 monthly visitors, is not the same 800 for Client B. Client B is a B2B firm with 30 named accounts. The percentages mean different things to each business. RICE without normalization rewards the loudest client. The fix is to score Reach within each client's backlog separately, and never compare RICE scores across clients.
The cleaner pattern for cross-client decisions is to swap Reach for Revenue or Strategic Value. A retainer worth $20,000 per month is a different priority than one worth $5,000, even if both have 100 affected end-users. This is closer to WSJF territory: Cost of Delay divided by Effort. Agencies that have tried strict RICE across clients tend to migrate toward this version within two quarters.
One operational rule helps inside a single client's backlog. The client should see the scores. Most teams hide them, treating RICE as an internal scoring exercise. Sharing the rank with the client surfaces disagreements early, before the team ships the wrong thing. Why does the dashboard rank below the comment threads? That is the conversation the team needs to have before sprint planning, not after the work is half-built.
What we recommend at Rock
Rock is not a dedicated product management tool, but most of our customers run prioritization rounds inside the same workspace they use for chat. From that vantage, the pattern that works is small: keep the scores, the debate, and the work in one place. Spreadsheets and standalone tools split the conversation from the artifact, which makes the scores easier to ignore once the sprint starts.
Practical setup: a scoring doc lives in the Notes section of the relevant space. The four columns are Reach, Impact, Confidence, Effort. A fifth column holds the source for each number, so the next person can audit. Once the round is complete, the top-ranked items move into the task board as cards. The card description carries the score and the reasoning. When a stakeholder asks why item three ranks above item seven, the answer is one click away.
For agencies running multiple client backlogs, each client gets its own space with its own scoring doc and task board. The format is identical across spaces, so a project lead moving between three clients reads three docs that look the same. The strict template matters more than the choice of variables, because consistency is what lets the team scale beyond two or three clients.
Free resource: the Agile Sprint Planning template ships with backlog, sprint, and review columns ready to receive your scored items.
One last move that compounds: re-score the top five items at the end of each quarter. Compare the post-launch reality to the original scores. The team that does this for four quarters running gets meaningfully better at the Impact and Confidence numbers. The calibration is built from their own past work. The team that never re-scores keeps making the same calibration errors.
Frequently asked questions
What does RICE stand for in scoring?
RICE stands for Reach, Impact, Confidence, and Effort. The first three are multiplied, the result is divided by Effort, and the output is a single score used to rank initiatives against each other. The framework was created by Sean McBride at Intercom in 2018.
What is the RICE scoring method used for?
RICE is used to rank competing product, marketing, or operations initiatives by expected value per unit of effort. It is most common in quarterly roadmap planning when a team has more ideas than capacity and needs a defensible way to decide what ships first.
What is a good RICE score?
There is no absolute number. A RICE score is only meaningful when compared to other RICE scores in the same round, scored against the same Reach window and effort unit. Pay attention to the relative ranking, not the headline figure. Treat any two scores within 15 percent of each other as a tie.
What is the difference between RICE and ICE?
RICE multiplies Reach, Impact, and Confidence, then divides by Effort. ICE drops the Reach factor and multiplies Impact, Confidence, and Ease (the inverse of Effort). ICE is faster and works for cases where reach is uncertain or always large. RICE is better when you can attach a real number to reach.
Who invented RICE scoring?
Sean McBride, then a product manager on the Intercom growth team, published the original RICE post on the Intercom blog in 2018. The framework grew out of internal arguments about which experiments to ship next and was meant to give the team a shared structure for making the call.
What are the disadvantages of RICE scoring?
RICE multiplies four numbers, three of which are estimates, so small errors compound. Confidence is gut-judged. Effort is hard to estimate honestly. Foundational work and strategic bets always score low. The math also assumes initiatives are independent, which is rarely true on a real roadmap.
How is RICE different from MoSCoW or Eisenhower?
RICE produces a numeric ranking. MoSCoW sorts items into Must, Should, Could, and Won't categories without ranking inside them. Eisenhower is a 2x2 grid for daily task triage by an individual. Most teams use two or three of these for different decisions rather than picking one.
Can agencies use RICE scoring across clients?
Strict RICE breaks down when scoring across clients because Reach numbers from different client bases are not comparable. The cleaner pattern is to score RICE inside each client's backlog separately and use a different model, often closer to WSJF, for cross-client decisions about where to put agency capacity.
RICE scoring is simple math built on top of four estimates. Used honestly, with the stress-test in mind, it is a faster way to defend a roadmap decision than a long argument. Rock combines chat, tasks, and notes in one workspace so the scoring, the debate, and the shipped work stay together. One flat price, unlimited users. Get started for free.
