Rebundle
/rice-prioritize
Skill

RICE Prioritize

Prioriza features con RICE (Reach × Impact × Confidence ÷ Effort). Convierte 30 ideas en una lista ordenada por valor.

SKILL.md
---
name: rice-prioritize
description: Score and rank a backlog of product ideas using RICE — Reach (how many users), Impact (how much per user), Confidence (how sure are we), Effort (eng-weeks). Output a sorted CSV with rationale per row. Used for quarterly planning, JIRA grooming, roadmap defense to leadership.
---

RICE Prioritize

Created at Intercom. The most defensible prioritization framework because it forces you to declare your assumptions on each axis instead of waving your hands.

When to use this skill

  • Quarterly planning with 20+ candidate features.
  • Defending the roadmap to leadership.
  • Saying no to a stakeholder request without it being personal.
  • Resolving "but my idea is more important" debates.

The formula

RICE = (Reach × Impact × Confidence) / Effort

Reach

How many people the feature will affect per quarter. Use absolute numbers. Examples:

  • "Every user logging in" → MAU
  • "Users on Pro plan only" → count of Pro accounts
  • "Users hitting this bug" → support ticket volume

Impact

How much it moves the metric per user. 5-point scale:

  • 3 — massive
  • 2 — high
  • 1 — medium
  • 0.5 — low
  • 0.25 — minimal

Confidence

How sure are we about the inputs. Percentage:

  • 100% — high confidence (data, evidence, well-scoped)
  • 80% — medium (reasoned estimate, some signal)
  • 50% — low (gut feel, no data)

If confidence drops below 50%, run discovery before scoring.

Effort

Eng-weeks (or person-weeks) total. Estimate generously.

Process

  1. List every candidate feature in a CSV.
  2. For each, fill in Reach, Impact, Confidence, Effort with one-line justifications.
  3. Compute RICE score.
  4. Sort descending.
  5. Cut at the line where you run out of capacity for the quarter.
  6. Show the bottom of the list to stakeholders so they see what got cut and why.

Output format

Feature,Reach,Impact,Confidence,Effort,RICE,Rationale
Auto-save drafts,12000,2,80%,2,9600,"Hit by every active user; clear support tickets; 2 eng-weeks based on prior similar work"
Dark mode,8000,0.5,100%,3,1333,"Universal but low impact; many users ask for it but no churn signal"
SAML SSO,150,3,80%,8,45,"Specific to enterprise tier; gates 3 deals worth combined $80K ARR"
...

Common failure modes

  • Inflating Confidence to win. Fix: peer-review the percentages.
  • Comparing apples and oranges (one feature scoped, another speculative). Fix: same scoping depth.
  • Forgetting opportunity cost (Effort 8 means 8 weeks not spent on the next thing).
  • Treating RICE as truth instead of a forcing function. The score helps you argue, not decide.

Constraints

  • Never accept "high impact" or "easy" without numbers. Push back.
  • Always include the bottom 5 cuts visible to stakeholders.
  • Re-score quarterly — Reach and Confidence change as data accumulates.