Reach × Impact × Confidence ÷ Effort. A single score that lets you compare initiatives on the same scale — and stops the loudest voice from winning by default.
RICE is a quantitative prioritisation model from Intercom's product team. It produces a numeric priority score for each initiative from four factors — Reach, Impact, Confidence, and Effort — so competing initiatives can be compared on a common scale, surfacing non-obvious priorities that gut instinct misses.
Its real power is political as much as analytical. By forcing every initiative through the same four questions, RICE moves the argument from who advocated loudest to what the numbers say — and makes the reasoning visible, so a low score can be challenged on its inputs rather than its champion.
(Reach × Impact × Confidence) ÷ Effort = RICE score
Reach: users affected per period · Impact: per-user effect (e.g. 3 / 2 / 1 / 0.5 / 0.25) · Confidence: % evidence (100 / 80 / 50) · Effort: person-months
Without a shared scoring language, roadmaps get set by whoever presents last, argues hardest, or has the CEO's ear. RICE replaces advocacy with arithmetic.
| The shortcut | What it costs | What it gives you instead |
|---|---|---|
| Loudest voice wins | Capacity gets allocated by political weight, not user value. | A common score separates the initiative from the person advocating it. |
| Gut-feel priorities | “This feels important” can't be compared or challenged. | Four explicit factors make the reasoning visible and debatable. |
| Ignoring effort | High-impact ideas get chosen regardless of cost. | Dividing by effort surfaces cheap wins and expensive darlings alike. |
| Overconfidence | Estimates stated as facts inflate weak ideas. | The Confidence factor explicitly discounts guesses. |
How many users will this affect per time period? Use a real number from product data, not an aspiration. “All users” is almost never the honest answer.
How much does it improve the experience for each affected user? Use a fixed scale (e.g. massive 3, high 2, medium 1, low 0.5, minimal 0.25) so scores are comparable.
How much evidence backs your Reach and Impact estimates? 100% = solid data, 80% = some evidence, 50% = an educated guess. This is where wishful thinking gets taxed.
How many person-months across product, design, and engineering? Estimate honestly — the denominator is where pet projects get exposed.
Calculate (R×I×C)÷E, sort descending, and read the ranking. Don't worship it: if a score feels wrong, interrogate the inputs, which is exactly the productive debate RICE is designed to produce.
A team had eight proposals and four person-months. The loudest advocate wanted a flashy AI feature; the data-quiet PM wanted an unglamorous fix. RICE scored the AI feature's Reach honestly — only a fraction of users hit the relevant flow — and its Effort high.
The unglamorous fix, with broad reach and low effort, scored nearly double. The debate shifted from “whose idea is cooler” to “is that Reach number right?” — which was answerable from analytics in minutes.
The deliverable is a table of initiatives with all four factors and a computed score — sortable, challengeable, and transparent.
| Factor | Question | Watch out for |
|---|---|---|
| Reach | Users affected / period | “All users” when only a segment hits the flow |
| Impact | Per-user effect (fixed scale) | Inflating every item to “massive” |
| Confidence | % evidence behind R & I | Stating guesses as 100% |
| Effort | Person-months, all functions | Forgetting design/QA; underestimating |
RICE's value isn't the number — it's that the number is made of visible inputs. You can't argue with a gut feeling, but you can argue about whether Reach is really 10,000 or 2,000.
Treat the output as a conversation starter, not a verdict. When a score surprises you, the right response is to examine the factor that drove it, not to override the framework. That habit — challenging inputs, not conclusions — is what turns a roadmap debate from a popularity contest into a shared analysis.
Scoring “all users” when only a slice hits the flow. Use real data.
If every item is top-impact, the factor is doing no work. Be ruthless and comparative.
Unfounded certainty inflates weak ideas. Discount honestly.
RICE informs judgment; it doesn't replace it. Interrogate surprising scores via their inputs.
Reach and Impact are far more accurate scored against a real job than a feature assumption.
When speed matters more than precision, ICE drops Reach and Effort for a faster three-factor score.
RICE ranks; MoSCoW buckets. Use RICE to order, MoSCoW to draw the must/should line for a release.
A high RICE score still has to pass the strategic-fit test from the MVSR ladder before it earns a slot.
Pick three initiatives and score each: Reach (a real number), Impact (fixed scale), Confidence (%), Effort (person-months). Compute (R×I×C)÷E.
Check whether the ranking matches your gut. Where it doesn't, find which single factor is driving the gap.
The disagreements between the score and your instinct are the most valuable output — they're usually an input you were guessing or a bias you didn't notice.
PM Mapped is the complete product-management course — 156 tools, 6 modules, 25 AI deep-dives. Join the waitlist to get in first.
Join the waitlist →Browse all 156 →