RICE score calculator: the formula, explained with 3 worked examples

The RICE formula is (Reach × Impact × Confidence) / Effort. Pasted into a spreadsheet, that's a calculator. But the formula is the easy part — the hard part is picking honest numbers for the inputs so the output isn't theatre.

This page is the calculator, the rules for each input, and three full worked examples with the arithmetic showing.

Want to run RICE on your phone? Framework for iPhone & iPad ships a RICE worksheet with the scoring rules built in — fill in your features, see the ranking. Or try the live RICE example on the framework page first.

The formula (and what each input really means)

RICE = (Reach × Impact × Confidence) / Effort

Higher score = higher priority. The four inputs are not interchangeable — each one has a specific definition and a specific scale. Mixing them up is the most common way RICE goes wrong.

Input	What it measures	Unit / Scale
Reach	How many people the work affects in a defined time period	Real count (users / customers / events) per quarter
Impact	How much it moves the metric per person when it lands	Fixed 5-point scale: 3 / 2 / 1 / 0.5 / 0.25
Confidence	How sure you are about Reach and Impact	100% / 80% / 50% only
Effort	Total person-months across everyone involved	Real number (sum of all roles' time)

The two "real number" inputs (Reach, Effort) want actual counts. The two scaled inputs (Impact, Confidence) use deliberately coarse buckets to prevent fake precision.

The Impact scale

Score	Meaning	Example
3	Massive	A retention fix that lifts engaged-user rate by 20%+
2	High	A new feature that moves a top metric by 5–15%
1	Medium	A solid improvement to an existing flow
0.5	Low	A small win, noticeable but not category-changing
0.25	Minimal	A polish that users may not notice

If you find yourself wanting 1.5 or 2.5, you are pretending the framework is more precise than it is. Pick a bucket.

The Confidence rule

Score	Meaning	Evidence
100%	Strong evidence	Quantitative data + tested hypothesis
80%	Some evidence	Qualitative signal + one supporting data point
50%	Gut feel	Opinion + analogy to other work

Anything below 50% means the score is noise. If your Confidence is < 50%, the idea isn't ready to prioritize — it's a research item, not a roadmap item.

How to calculate a RICE score (5 steps)

1. Define the one metric you're prioritizing against. Active users? Revenue? Tickets-deflected? RICE only compares apples to apples. One backlog = one metric.
2. Estimate Reach in real units for a fixed period. "8,000 users per quarter," not "lots of people." Pick the period (usually a quarter) and stick to it for every item.
3. Pick Impact from the 5-point scale. Don't invent values. If you can't decide between 2 and 3, pick 2 (the formula rewards being conservative).
4. Pick Confidence from one of three values: 100%, 80%, or 50%. Be honest about evidence. Confidence is where most teams lie to themselves.
5. Estimate Effort as total person-months. Sum across roles: 2 engineers × 2 months + 1 designer × 0.5 months = 4.5 person-months.

Then (R × I × C) / E, sort descending, look at the top quartile.

---

Worked example 1: A B2B SaaS Q3 backlog

A team running a $50K MRR B2B SaaS product is choosing between four candidates for Q3. Each is sized roughly the same — bigger than a bug fix, smaller than a strategic pivot.

Item	Reach (users/Q)	Impact	Confidence	Effort (PM)	RICE Score
Onboarding redesign	6,000	2	95% (round to 100%)	3	(6000 × 2 × 1.0) / 3 = 4,000
AI recommendation v2	8,000	3	80%	6	(8000 × 3 × 0.8) / 6 = 3,200
Mobile redesign	12,000	2	90% (round to 80%)	10	(12000 × 2 × 0.8) / 10 = 1,920
Export-to-CSV	3,000	1	100%	2	(3000 × 1 × 1.0) / 2 = 1,500

Ranking: Onboarding redesign > AI recommendation v2 > Mobile redesign > Export-to-CSV.

The interesting result: Mobile redesign has the highest Reach (12,000) but ranks third because Effort is 10 person-months. RICE is doing what it's supposed to do — punishing big swings unless the Reach × Impact justifies the cost.

The team's gut said "ship Mobile because it touches the most users." RICE says: not at that effort multiplier. Either cut Mobile's scope (lower Effort) or rank it lower.

For a fully written-up version of this exact scenario, see the RICE on a real SaaS Q3 backlog example.

---

Worked example 2: Growth experiments competing for one engineer-week

Growth teams often have one engineer-week to allocate among 5 experiments. Reach is the daily user count touching each surface, Effort in person-weeks.

Experiment	Reach (users/day)	Impact	Confidence	Effort (PW)	RICE Score
Pricing page A/B test	2,500	1	80%	0.5	(2500 × 1 × 0.8) / 0.5 = 4,000
Email re-engagement campaign	8,000	0.5	50%	0.25	(8000 × 0.5 × 0.5) / 0.25 = 8,000
Onboarding tooltip tweaks	1,200	0.5	100%	0.5	(1200 × 0.5 × 1.0) / 0.5 = 1,200
Referral incentive bump	600	2	50%	1	(600 × 2 × 0.5) / 1 = 600
New homepage hero	4,000	1	50%	1	(4000 × 1 × 0.5) / 1 = 2,000

Ranking: Email re-engagement (8,000) > Pricing A/B (4,000) > Homepage hero (2,000) > Onboarding tweaks (1,200) > Referral bump (600).

The Email campaign wins despite a low Impact (0.5) and gut-feel Confidence (50%), because Reach is huge and Effort is a quarter-week. This is RICE telling you: with very low effort, even modest expected value beats higher-stakes bets. Ship the cheap experiment first; the data improves Confidence for everything downstream.

---

Worked example 3: Engineering platform investments

Platform teams hate RICE because Reach is awkward (your "users" are other engineers) and Impact lacks revenue language. Here's how to do it honestly.

The trick: define Reach as the count of internal users (engineers) affected, and define Impact in terms of saved time or prevented incidents.

Investment	Reach (engineers)	Impact	Confidence	Effort (PM)	RICE Score
Faster CI pipeline	40	2	80%	4	(40 × 2 × 0.8) / 4 = 16
Deploy automation	40	1	100%	2	(40 × 1 × 1.0) / 2 = 20
New observability stack	40	3	50%	12	(40 × 3 × 0.5) / 12 = 5
Dependency upgrade tooling	25	0.5	100%	1	(25 × 0.5 × 1.0) / 1 = 12.5

Ranking: Deploy automation > Faster CI > Dependency tooling > Observability stack.

The Observability stack has the highest Impact (3) but the lowest Confidence (50%) and biggest Effort (12 PM) — it's a high-stakes bet without evidence. RICE puts it last not because it's a bad idea, but because the team should reduce uncertainty before committing 12 person-months.

If the team can run a 1-week spike (Confidence → 80%) and confirm the value, the score becomes (40 × 3 × 0.8) / 12 = 8 — moving it ahead of Dependency tooling but still behind Deploy. Confidence is the most leveraged input in RICE.

---

When NOT to use RICE

RICE earns its keep when you have a backlog with 20–30 comparable items and a team that keeps relitigating priority. It is the wrong tool when:

• You have fewer than 5 candidates. With that few, the ranking is noisy. Use RICE vs ICE or pros/cons.
• The items aren't comparable in scope. Don't mix three-day fixes with six-month epics — the spread destroys the math.
• You're choosing between strategic directions. RICE is a tactical tool. For "should we enter market X?" use Ansoff Matrix or Five Forces.
• Confidence is below 50% for most items. That means you're prioritizing research, not work. Run discovery first.

Common mistakes

Mistake	Fix
Inventing Impact values like 1.5 or 2.5	Pick a bucket. The coarseness is the point.
Treating Confidence as "how much I like the idea"	Confidence is about evidence, not preference. Be honest.
Comparing Reach across different time periods	Pick one period (usually quarter) and use it for everything.
Estimating Effort in calendar time instead of person-time	A 2-month project with 3 people is 6 person-months.
Sorting and shipping the top item with no discussion	The score is a starting point, not a verdict. The team's job is to spot which scores feel wrong and find the input that's lying.

RICE vs other prioritization frameworks

RICE has a Reach term that simpler frameworks like ICE leave out. If you can quantify Reach with data, RICE produces more defensible rankings. If you can't, ICE is faster.

For a side-by-side comparison with worked examples, see RICE vs ICE: which prioritization framework to use. For larger enterprise contexts with SAFe, see WSJF.

The original source

RICE was published in 2017 by Sean McBride at Intercom as the framework Intercom's product team actually used internally. The format caught on because it converts vague debate into specific numbers without pretending to be more precise than its inputs.

---

Run RICE on your team's backlog: Framework's iOS app ships a RICE worksheet preloaded with the scoring rules above. Get it on the App Store. Or use the web version with the full live example on the framework page.