How to Prioritize Features with RICE Framework (Step-by-Step)
RICE (Reach, Impact, Confidence, Effort) quantifies prioritization and reduces HiPPO influence in backlog management.
This guide covers correct application and common mistakes that make RICE scores meaningless.
Step-by-step guide
Step 1: Score Reach
How many users affected per quarter. Use analytics data, not guesses. If guessing, mark Confidence as low.
Step 2: Score Impact
Scale: 3 (massive), 2 (high), 1 (medium), 0.5 (low), 0.25 (minimal). Most features are 1.
Step 3: Score Confidence
100% (data-proven), 80% (strong evidence), 50% (some evidence), 20% (speculation). Low confidence correctly penalizes uncertain features.
Step 4: Score Effort
Total team time in person-months. Include design, engineering, QA, and PM time.
Step 5: Calculate and rank
Score = (Reach x Impact x Confidence) / Effort. Rank highest to lowest. Review: does it feel right?
Common mistakes
Inflating Impact
If every feature is a 3, you are not using the scale. Most are 1. A few are 2. Very rarely a 3.
Guessing Reach
If guessing, set Confidence to 20-50%. This correctly deprioritizes unvalidated estimates.
RICE as final decision
RICE is input to discussion, not the final answer. Strategic bets and dependencies can override scores.
Tips
- Create a spreadsheet with RICE formulas for all backlog items
- Re-score quarterly as data improves
- Use Vantage-generated requirements as scoring units
How Vantage helps
Vantage generates requirements with priority levels based on PRD content. These priorities serve as a starting point that teams refine using RICE.