RICE Score

Use this when you need to prioritize a set of features, initiatives, or backlog items using the RICE framework (Reach, Impact, Confidence, Effort). Produces a ranked priority table with transparent scoring and rationale.

Related resources: rice-scoresheet.md and rice-scoresheet.csv -- downloadable RICE scoring spreadsheet with formula and scoring guide.

Framework attribution: RICE scoring was popularized by Intercom as a lightweight prioritization method for product teams.

Process

Step 1: Pre-prioritization gate

Before scoring anything, confirm the team can answer three foundational questions:

1. What problem are we solving? (If the team cannot articulate the problem, they are not ready to prioritize solutions.)
2. Who are we solving it for? (A ranked list without a named audience is just opinions.)
3. How are we measuring success? (Without a success metric, RICE scores are theater -- precise numbers disconnected from outcomes.)

If the team cannot answer all three, stop. Prioritization is premature. Help them articulate these first -- this is the context that makes scoring meaningful, not busy work that precedes it.

Also: if the team presents an unranked bullet list and says "these are all important," push back. Unordered lists are a symptom of avoiding hard prioritization decisions. Any first stab at ranking is better than no ranking -- treat prioritization like brainstorming, not legislation.

Step 2: Gather inputs

Ask the user to provide:

1. Feature list -- the features, initiatives, or backlog items to prioritize (names and brief descriptions)
2. Time horizon — what period are you scoring for? (e.g., this quarter, next 6 months)
3. User base context — total active users, user segments, or customer count (needed for Reach estimates)
4. Team capacity — team size and available engineering time (needed for Effort estimates)
5. Any existing data — usage metrics, research findings, or stakeholder input that should inform scoring

The user can paste items in any format — tracker exports, bullet points, spreadsheet data.

Step 3: Identify which personas are affected

Before scoring, prompt the user:

Persona check: Which user persona(s) does each feature serve? If you have defined personas (e.g., from /persona-create or /persona-draft), name them. If not, describe the key user types briefly — their role, goals, and what makes them distinct.

Which persona is currently most underserved? Features that serve underserved personas may deserve higher Impact scores.

This grounds your Reach and Impact estimates in real user segments rather than abstract numbers.

Step 4: Score each feature

For each feature, assign scores across the four RICE dimensions:

Dimension	What it measures	How to score
Reach	How many users will this affect per quarter?	Estimate a number (e.g., 500 users/quarter). Use data when available; estimate conservatively when not.
Impact	How much will this move the needle per user?	Use the standard scale: 3 = massive, 2 = high, 1 = medium, 0.5 = low, 0.25 = minimal
Confidence	How sure are you about these estimates?	Percentage: 100% = solid data, 80% = strong conviction, 50% = informed guess, 20% = speculation
Effort	How many person-months will this take?	Estimate in person-months. Include design, engineering, and QA.

RICE Score = (Reach x Impact x Confidence) / Effort

When estimating Reach, consider which persona segments are affected — a feature that reaches 500 power users may score differently than one reaching 500 casual visitors. When estimating Impact, weight features that serve underserved or strategically important personas higher.

Step 5: Generate the priority table

Output in this format:

---

RICE Prioritization: (feature set or project name)

Date: (today's date) Time horizon: (quarter or period) Team capacity: (team size and available effort)

Ranked priority table

Rank	Feature	Reach	Impact	Confidence	Effort	RICE Score	Rationale
1	(Feature name)	(number)	(0.25-3)	(%)	(person-months)	(calculated)	(1-sentence explanation of the score)
2	(Feature name)	(number)	(0.25-3)	(%)	(person-months)	(calculated)	(1-sentence explanation)
...

Scoring notes

(Feature name) — (explain key assumptions behind the Reach and Impact estimates. Note what data supports the score or what's missing.)

(Repeat for each feature that has notable assumptions or low Confidence.)

Recommendations

• Top priority: (Feature) — (why this should go first, beyond just the score)
• Quick wins: (Features with high RICE and low Effort — consider doing these first regardless of rank)
• Needs more data: (Features with Confidence below 50% — what would increase confidence?)
• Candidates for cutting: (Features with low RICE scores — discuss whether these belong on the roadmap at all)

---

Step 6: Review and adjust

Ask the user:

• Do the Reach estimates feel right? (This is the most commonly mis-estimated dimension.)
• Any features where Impact should be higher or lower based on strategic context?
• Any Confidence scores that should change based on data you have that I don't?
• Want me to run sensitivity analysis — what happens if we adjust the most uncertain scores?

Adjust scores based on feedback and regenerate the table.

Related skills

• /backlog-refine — use after RICE scoring to refine the top-priority items into well-shaped stories
• /story-write — write stories for the features that score highest
• /pre-mortem — run a pre-mortem on the plan that emerges from prioritization

Portfolio example: B Lab (2020-2021)

A product leader used RICE to prioritize 8 competing features for B Lab's roadmap. Key scoring decisions: "Measure actual company impact via SDGs" scored highest (Reach: 3500, Impact: 4, Confidence: 40%, Effort: 4, Score: 1400) because the SDG audience was comparable to the total B Corp population. "Fix Analytics data accuracy bugs" ranked second (Score: 666) with high confidence (60%) and low effort (1) because it was a bug fix affecting brand perception. The PM documented all assumptions with footnotes explaining proxies (e.g., LinkedIn headcount as a proxy for internal user reach, certification success rates to estimate monthly applicants).

Output location

Present the RICE analysis as formatted text in the conversation for the user to share with stakeholders or paste into their planning tool.