Title, meta description, canonical URL, OG tags, Twitter cards, breadcrumb schema, FAQPage schema, and WebApplication schema are configured for this page.
Calculate the original RICE score, show where each estimate came from, and test whether your ranking survives the alternative assumptions you actually believe are plausible.
The score structures the decision; it cannot manufacture evidence or make the final Product call for you. This calculator keeps the original Intercom RICE mechanics intact while exposing comparability problems and rank fragility.
No login. Inputs stay in this browser and are autosaved locally; CraftUp does not persist initiative text on its backend.
Calculator mode
Arithmetic period conversion only (week x13, month x3, quarter x1, year /4), not a forecast. For one-time or seasonal reach, enter a directly comparable period estimate. Import accepts v1 and v2 headers; export uses: name, reach_input, reach_timeframe, reach_unit, reach_source_type, reach_source_reference, impact, impact_goal, impact_rationale, confidence_percent, confidence_rationale, effort_person_months, effort_source_type, effort_source_reference, notes, alternative_reach, alternative_impact, alternative_confidence_percent, alternative_effort_person_months, alternative_note, rice_score, rank.
Impact (?), confidence (?), effort (?). Expand “Assumptions & alternative” on any row to record provenance or test fragility.
| Initiative name | Reach | Reach time frame | Impact | Confidence | Effort | Notes | RICE score | Rank | Actions |
|---|---|---|---|---|---|---|---|---|---|
Qtr: 1000 | 800 | #1 | |||||||
Assumptions & alternative case for “First initiative”Input provenance (optional) Alternative plausible case (optional) Only enter values you genuinely consider plausible. Blank means unknown, nothing is invented. | |||||||||
The decimal precision of the formula is not the precision of your knowledge. Ranking uses unrounded arithmetic; equal raw scores share a rank and are marked as tied.
Robustness not assessed, no alternative case entered.
| Initiative | Base rank | Alternative rank | Movement |
|---|---|---|---|
| First initiative | #1 | Not assessed | - |
Deterministic recomputation from your entered alternatives only. No invented ranges, no probability scores, no uniform-confidence scaling.
Relative to the current backlog, not a universal quick-win rule.
Lowest confidence values in this comparison, review, not an automatic discovery verdict.
RICE ranks mechanically. Record the actual Product decision separately, including when a non-RICE reason (dependency, reliability, commitment, strategy) overrides the raw order.
Loading saved data...
Use the calculator as an inspectable decision model, not a one-time scoring exercise. The value comes from keeping assumptions visible and rerunning ranking as evidence changes, so roadmap conversations stay grounded in current information.
Step 1
Define comparable candidates: one comparison horizon, one Reach population, one Product goal per comparison.
Step 2
Enter base RICE assumptions with the original formula, Reach x Impact x Confidence / Effort, and record where each estimate came from.
Step 3
Inspect the base ranking and the comparability review before trusting the order.
Step 4
Enter plausible alternative assumptions where uncertainty matters, then inspect whether the rank actually moves.
Step 5
Record non-RICE constraints and the final Product decision separately, with a revisit trigger for when evidence changes.
RICE is short for Reach, Impact, Confidence, and Effort, from Intercom's original RICE framework. The score is (Reach x Impact x Confidence) / Effort, with confidence entered as a decimal (80% enters the formula as 0.8). It is primarily useful for comparing candidate initiatives under consistent definitions, not an absolute measure of Product value, and never an objective roadmap answer.
How many people or events the initiative affects within a defined period, from real Product data where possible. Enter week, month, quarter, or year; the tool shows a normalized comparison estimate per quarter. That conversion is arithmetic convenience, not a forecast, for one-time or seasonal reach, enter a directly comparable period estimate instead.
Impact on one person reached, relative to one Product goal, using the original scale: Massive 3, High 2, Medium 1, Low 0.5, Minimal 0.25. These are relative categories, not universal KPI-change percentages. Name the outcome each row targets; rows with different goals may not be comparable.
A heuristic discount on uncertain estimates, starting from 100%, 80%, and 50%. These are not confidence intervals or statistical confidence, use a percentage your team can explain consistently, and write down what evidence supports the estimates and what remains uncertain.
Total relevant cross-functional effort in person-months (minimum 0.5, steps of 0.5), product and planning, design, engineering, and QA or testing where they apply. Record where the estimate came from and keep units consistent across the comparison.
What a RICE score can tell you: which candidate ranks first under one consistent set of assumptions. What it cannot tell you: whether those assumptions are true, whether rows are truly comparable, or what the Product decision should be once strategy, dependencies, and hard constraints enter the picture.
Keep scoring pragmatic and framework-faithful. The labels below are relative categories from the original framework, not universal numeric thresholds.
If your evidence is mixed, set a custom value and note what would increase certainty. A low value never automatically means discovery work, a cheap experiment can intentionally carry 50%.
A score like 426.67 can look precise while every input is approximate. The decimal precision of the formula is not the precision of your knowledge, and a gap between 412 and 405 is not automatically decision-significant. Enter the alternative Reach, Impact, Confidence, or Effort you genuinely consider plausible for each uncertain row, and the robustness lab recomputes the ranking deterministically: baseline rank, alternative rank, exact movement, and which single factor can flip the order. Rows without alternatives report robustness as unknown, never as stable.
RICE should not be treated as a hard-and-fast rule. A lower-ranked item can still be the right call for a visible non-RICE reason: a hard dependency, reliability or user harm, a contractual or security requirement, strategic fit, sequencing, capacity, or learning value. Record that reason in the decision record without touching the formula, adding strategy or risk weights would stop being RICE. When the remaining question is broader Product prioritization, continue in the prioritization workspace. Fast early comparisons with unknown Reach belong in ICE; release scope negotiation belongs in MoSCoW; satisfaction-response evidence belongs in Kano.
These three prefilled scenarios are explicitly hypothetical product initiatives. Load one to see how the arithmetic behaves, and which assumption could change the rank.
Hypothetical: Reach 1200 per quarter, impact High (2) on activation, confidence 80%, effort 1.5 person-months.
Ranks high on base inputs, but the rank is fragile: if only half of signups are eligible (Reach 600), the order can change. The arithmetic is correct; the Reach assumption is what needs validation.
Hypothetical: Reach 350 per month (normalized comparison estimate), impact Medium (1) on trial starts, confidence 50%, effort 0.5 person-months.
Cheap to run and honestly uncertain. A low confidence value does not automatically mean discovery work, here it marks a deliberate cheap experiment, not a blocked decision.
Hypothetical: Reach 120 enterprise accounts per quarter, impact Massive (3) on expansion, confidence 80%, effort 4 person-months.
Ranks lower because effort is substantial. If engineering discovery revises effort upward, the rank moves further, enter that as an alternative case instead of debating the score.
Symptom: Scores look inflated across every initiative.
Cause: Impact is set to Massive by default without a shared outcome.
Fix: Name one target outcome per row and challenge every Massive score against that outcome with concrete reasoning.
Symptom: A weekly initiative outranks long-term bets unexpectedly.
Cause: Reach timeframe was mixed and normalization was treated as a forecast.
Fix: Check the normalized comparison estimate and, for one-time or seasonal reach, enter a directly comparable period estimate instead.
Symptom: Top-ranked rows change every meeting with no clear reason.
Cause: Inputs are edited without updating assumptions in notes.
Fix: Require one note or provenance update whenever reach, impact, confidence, or effort changes.
Symptom: Import CSV fails even though values look correct.
Cause: Headers or order differ from the required template.
Fix: Use the downloadable sample CSV. Version-1 files still import; new fields arrive blank, never invented.
Symptom: Initiatives with low confidence still dominate ranking.
Cause: Confidence was set to 100% despite limited evidence.
Fix: Lower the estimate to a value you can explain and record what evidence would raise it.
Symptom: Effort inputs are rejected with validation errors.
Cause: Effort values are below 0.5 or not in 0.5 increments.
Fix: Use person-month estimates with minimum 0.5 and step size 0.5.
Symptom: Stakeholders reject the ranked table as too technical.
Cause: Raw math is shared without practical rationale and assumptions.
Fix: Switch to the calculation-explanation view and present base inputs, provenance, and alternative-case movement together.
Symptom: The ranking feels certain but every input is a guess.
Cause: No alternative assumptions were entered, so fragility stays invisible.
Fix: Add one plausible alternative per uncertain factor and inspect whether the rank survives, that is the real robustness check.
Yes. The RICE score calculator is fully free, requires no login, and supports batch comparison and single-initiative checks. You can compare a backlog, inspect assumptions, and export results to CSV, Markdown, or JSON without any paywall steps.
No server storage is used for your initiative inputs. Data stays in this browser and is autosaved locally so you can resume work later. You can remove everything at any time with Clear data. CraftUp does not persist your initiative names, notes, or scores on its backend, but share URLs do encode your workspace, so review them before posting externally.
A chat tool can brainstorm, but it does not enforce consistent RICE structure by default. This calculator standardizes inputs, normalizes reach per quarter, validates effort steps, ranks initiatives, records where each estimate came from, and recomputes rank under your own alternative assumptions. That consistency is what makes recurring prioritization reviews credible across cycles.
Yes. The output format is built for practical product planning, including stakeholder reviews and backlog triage docs. Teams commonly export the Markdown decision record into planning notes and share compressed URLs internally. Just review notes and provenance references before sharing externally if they contain sensitive assumptions.
Treat 100%, 80%, and 50% as heuristic RICE levels, not statistical confidence. Use a percentage your team can explain consistently and write one sentence saying why: what evidence supports the estimates and what remains uncertain. The tool asks for that rationale but never calculates the percentage for you.
Teams often mix weekly, monthly, and quarterly reach estimates. Without normalization, ranking is misleading because time windows differ. The calculator converts entries to a normalized comparison estimate per quarter (week x13, month x3, year /4) while preserving your original input. These factors are arithmetic conveniences, not forecasts: for one-time or seasonal reach, enter a directly comparable period estimate instead of scaling a rate.
Estimate the total relevant cross-functional effort for the scoped initiative, product and planning, design, engineering, and QA or testing where they apply, in consistent person-month units (minimum 0.5, steps of 0.5). Record which estimate the number came from. If effort is uncertain, enter your plausible alternative in the robustness lab and watch whether the rank survives.
Yes. Share URL compresses your batch table, provenance, alternatives, and decision record client-side and recreates the session in a fresh browser. Everyone sees the same rows and assumptions. Share links are not encrypted or private, anyone with the link can read the encoded workspace, so trim customer names or internal references before sharing outside your team.
You optionally enter one plausible alternative per uncertain factor, for example a lower Reach or a higher Effort you genuinely consider possible. The lab recomputes the ranking with only your entered alternatives substituted and reports exact rank movement (such as #1 to #3), one-factor-at-a-time findings, and which areas have no alternative entered. It never invents uncertainty ranges or probability scores.
RICE structures comparison; it does not make the Product decision. When an item is mandatory (reliability, security, legal, committed scope), when candidates target incomparable goals, when sequence or dependencies dominate, or when strategy is unresolved, record the deviation reason in the decision record and use the broader prioritization workspace. Never add strategy or risk weights into the RICE formula, that would stop being RICE.
If inputs are not comparable, the framework choice is unclear, or trade-offs remain after scoring, review the Product decision in the broader workspace or compare frameworks first.
Last updated: 2026-09-22