ROI【🔏CLASSIFIED FILE】 No. X047 | What is the RICE Framework

Detective's Memo: The revolutionary prioritization framework "RICE" developed by Intercom. Many mistakenly perceive it as merely a "feature importance ranking system," but its true identity is "a quantification system that eliminates subjectivity and political influence, democratizing prioritization through data." Why is the opinion of the loudest executive not necessarily correct, and what is the real reason that "features everyone feels are important" may not actually deserve priority? Reach (number of people reached), Impact (magnitude of effect), Confidence (certainty level), Effort (work required)—the moment these four variables are distilled into the simple formula (R × I × C) ÷ E, organizational decision-making transforms from emotional reasoning to science. Eliminate the ambiguity of "because that person said so" or "it feels important somehow," and uncover the truth behind the transparent prioritization process practiced by Spotify and Airbnb.

What is the RICE Framework - Case Overview

The RICE Framework, formally known as "Quantitative Four-Factor Prioritization Evaluation Methodology," is a decision-making theory published in 2016 by Intercom's product management team. It is recognized among clients as a method that numerically evaluates four elements—Reach (number of people reached), Impact (magnitude of effect), Confidence (certainty level), and Effort (work required)—calculating a RICE score using the formula "(Reach × Impact × Confidence) ÷ Effort" and determining priority based on the highest scores. However, in actual practice, it is often superficially understood as "just a scoring system," with the majority of organizations failing to grasp its truly revolutionary value: the quantification of subjective judgments, transparent team consensus building, and the explicit incorporation of uncertainty through confidence levels.

Investigation Memo: RICE is not merely an "evaluation method" but a "democratization tool for organizational decision-making processes." Why is following HiPPO (Highest Paid Person's Opinion) dangerous, and how does quantification neutralize organizational power dynamics? It provides judgment criteria for "what to build" in MVP and scientifically supports sprint planning in Agile Development—we must decode this foundational prioritization system of modern product development.

Basic Structure of the RICE Framework - Evidence Analysis

Primary Evidence: Objective scoring through quantification of four elements

Overall RICE Formula

Basic Calculation:

RICE Score = (Reach × Impact × Confidence) ÷ Effort

Higher Score = Higher Priority
Lower Score = Lower Priority

Why This Formula Works:

Numerator (Reach × Impact × Confidence): - Represents "how much value will be generated" - Number of people reached × impact per person × probability of realization - Based on the concept of expected value

Denominator (Effort): - Represents "how much cost will be required" - ROI (Return on Investment) calculation structure - Higher cost = lower score

Result: High value + low cost initiatives = highest priority Low value + high cost initiatives = deprioritized

R: Reach (Number of People Reached)

Definition: Number of people affected within a defined time period

Measurement Unit:

People/Period

Examples:
- "1,000 people/month"
- "5,000 people/quarter"
- "20,000 people/year"

Measurement Methods:

Improving existing features:

Estimate based on current user numbers
Example: Monthly users of this feature = 2,500 people
→ Reach = 2,500 people/month

Adding new features:

Estimate from target segment
Example: 
- Total users: 10,000 people
- Target for this feature: Premium users
- Premium user count: 1,000 people
→ Reach = 1,000 people/quarter

Critical Insight:

Relative comparison matters more than absolute values:

Initiative A: 1,000 people/month
Initiative B: 500 people/month
→ Initiative A has 2x the Reach

Period standardization is essential:

❌ Initiative A: 1,000 people/month, Initiative B: 5,000 people/year
→ Not comparable

✅ Initiative A: 12,000 people/year, Initiative B: 5,000 people/year
→ Comparable

I: Impact (Magnitude of Effect)

Definition: Size of impact per person

Measurement Scale (Recommended):

3 = Massive Impact (revolutionary)
2 = High Impact (significant)
1 = Medium Impact (moderate)
0.5 = Low Impact (small)
0.25 = Minimal Impact (minimal)

Scale Selection Criteria:

Massive Impact (3): - User experience fundamentally transforms - Completely solves a major problem - Significantly improves competitive advantage Example: Drastically simplifying checkout process (10 steps → 2 steps)

High Impact (2): - Clear experience improvement - Solves important problems - Significantly increases usage frequency/satisfaction Example: Dramatic search speed improvement (5 seconds → 0.5 seconds)

Medium Impact (1): - Noticeable experience improvement - Partial problem resolution - A certain number of users benefit Example: UI improvements enhancing usability

Low Impact (0.5): - Slight improvement - Limited problem resolution - Few people notice Example: Improving error message wording

Minimal Impact (0.25): - Almost unnoticeable - Internal improvements - Only indirect effects Example: Standardizing log output format

Key Judgment Criteria:

Quantifying qualitative judgments:

"How much will users appreciate this?" → Predicted change in NPS → Customer interview reactions → Past performance of similar initiatives

C: Confidence (Certainty Level)

Definition: Confidence level in estimates (Reach, Impact, Effort)

Measurement Scale:

100% = High Confidence (certain)
80% = Medium Confidence (moderately certain)
50% = Low Confidence (low certainty)

Scale Selection Criteria:

High Confidence (100%): - Data-driven estimates - Past similar cases exist - Clear measurement methods available - Technical feasibility certain Example: Minor improvements to existing features, data-validated initiatives

Medium Confidence (80%): - Partial data available - Similar cases exist but circumstances differ - Some uncertainty present Example: New feature with established technology, predicted market response

Low Confidence (50%): - Minimal data available - Experimental/innovative initiatives - Technical feasibility questionable - Market response difficult to predict Example: Completely new concept, no precedent

Strategic Meaning of Confidence:

Making uncertainty explicit:

Traditional: "This feature is important (confidence unknown)"
RICE: "Reach=1000, Impact=2, Confidence=50%"
→ Uncertainty incorporated into numbers

Risk management:

Initiative A: Reach=1000, Impact=3, Confidence=100%, Effort=5
→ RICE Score = (1000×3×1.0)÷5 = 600

Initiative B: Reach=2000, Impact=3, Confidence=50%, Effort=5
→ RICE Score = (2000×3×0.5)÷5 = 600

Same score but Initiative A has lower risk

E: Effort (Work Required)

Definition: Total work required for implementation

Measurement Unit:

Person-Months

Examples:
- 0.5 person-months = 1 person for half a month (approx. 10 business days)
- 2 person-months = 1 person for 2 months or 2 people for 1 month
- 5 person-months = 5 people for 1 month or 1 person for 5 months

Estimation Scope:

Include all phases:

- Design and specification
- Development and implementation
- Testing and QA
- Deployment and release
- Documentation creation
- Stakeholder coordination

Entire team's effort:

❌ Developer effort only
✅ Designer + Developer + QA + PM total

Improving Estimation Accuracy:

Utilizing historical data:

Performance of similar features:
"Previous search feature improvement = 3 person-months"
→ If current is similar, estimate 3 person-months

Establishing Baseline of Measurement (BOM):

"Minimal feature addition = 0.5 person-months" as baseline
Current feature is "3x more complex" → 1.5 person-months

Evidence Analysis: The revolutionary nature of the RICE Framework lies in decomposing subjective "importance" into four objective variables, constructing a transparent decision-making system through formulas where everyone arrives at the same conclusion.

RICE Implementation Procedure - Investigation Methods

Investigation Finding 1: Intercom's Practical Process

Case Evidence (Real example from RICE Framework developers):

Phase 1: Initiative Listing (Feature and improvement idea identification)

Situation:

50+ initiative candidates in product backlog
- Requests from each team
- Customer feedback
- Executive directives
- Technical debt resolution

Problem: Unclear which to tackle first

Traditional prioritization (pre-RICE):

Method 1: HiPPO (Highest Paid Person's Opinion)
→ Executive's word is final
→ Frontline voices unheard

Method 2: Loudest person criteria
→ Sales director strongly requests
→ Actual value unknown

Method 3: Intuition/feeling
→ "Feels important somehow"
→ Retrospectively turns out to be failure

Result: Team dissatisfaction, inefficient development

Phase 2: RICE Introduction Decision

Background of decision:

CEO Brian Halligan:
"Make data-driven decision-making organizational culture"

Product team challenges:
- Multi-hour meetings for each prioritization
- Low team conviction after decisions
- Swayed by emotions and power dynamics

Solution: 
"Quantify four elements, decide by formula"

Phase 3: Concrete Initiative Evaluation (Actual Examples)

Initiative A: Add message read receipt feature

Reach: 5,000 people/quarter
(Predicted 50% of monthly active users will use)

Impact: 1 (Medium)
(Convenient but not revolutionary, similar features exist elsewhere)

Confidence: 80%
(Technical feasibility certain, usage rate estimated)

Effort: 2 person-months
(Frontend + Backend + Testing)

RICE Score = (5000 × 1 × 0.8) ÷ 2 = 2,000

Initiative B: Onboarding process improvement

Reach: 1,000 people/quarter
(All new registrants)

Impact: 3 (Massive)
(Initial experience dramatically improves, directly affects retention)

Confidence: 100%
(Effect proven through A/B testing)

Effort: 3 person-months
(Redesign + implementation of multiple screens)

RICE Score = (1000 × 3 × 1.0) ÷ 3 = 1,000

Initiative C: Admin dashboard design refresh

Reach: 200 people/quarter
(Administrators only, 2% of total)

Impact: 2 (High)
(Work efficiency significantly improves)

Confidence: 80%
(Design plan exists but implementation complexity uncertain)

Effort: 8 person-months
(Redesign, implementation, testing of all screens)

RICE Score = (200 × 2 × 0.8) ÷ 8 = 40

Phase 4: Priority Decision

Score ranking:

1st: Initiative A (Message read receipts) = 2,000
2nd: Initiative B (Onboarding) = 1,000
3rd: Initiative C (Admin dashboard) = 40

→ Tackle in order: A → B → C

Critical Insight:

Initiative B "appears most important" but ranks 2nd:

Reason: Effort is large at 3 person-months
→ ROI inferior to A
→ However, highest priority after A completion

Initiative C "strong request from administrators" but lowest:

Reason: Small Reach (200 people)
Reason: Extremely large Effort (8 person-months)
→ Using 8 person-months for other initiatives creates higher overall value

Phase 5: Results and Learning

Evaluation after 6 months:

Initiative A implemented:
- Reached predicted Reach
- Confirmed satisfaction improvement
- Effort also as estimated

Initiative B implemented:
- Retention improved 30% (better than expected)
- Impact of 3 was correct assessment

Initiative C:
- Still not implemented
- Completed 5 other high-score initiatives meanwhile
- Retrospectively correct decision

Team transformation:

Pre-introduction: 3-hour debates for each prioritization meeting
Post-introduction: 30-minute decisions with RICE calculation

Pre-introduction: Post-decision complaints "Why is this priority?"
Post-introduction: Scores provide rationale, improved conviction

Pre-introduction: Loudest person's opinion prevails
Post-introduction: Data makes final judgment, democratic process

Investigation Finding 2: Spotify's Application Case

Case Evidence (Strategic use of Confidence):

Challenge:

Developing new music recommendation algorithm
- Effect unknown
- Development cost certainly high
- Failure risk present

Traditional Approach (pre-RICE):

"Invest because it's innovative"
→ Uncertainty not considered
→ Accountability for failure unclear

RICE Evaluation:

Reach: 10,000,000 people/quarter
(Affects all users)

Impact: 3 (Massive)
(If successful, experience dramatically improves)

Confidence: 50% ← Key point here
(Technically feasible, effect unknown)

Effort: 20 person-months
(Entire ML team for 2 months)

RICE Score = (10,000,000 × 3 × 0.5) ÷ 20 = 750,000

Strategic Decision:

High score but emphasize Confidence=50%
→ MVP approach rather than full commitment

Implementation method:
Phase 1: Small-scale experiment (5 person-months)
- A/B test with 1,000 people
- Measure effectiveness
- Re-evaluate Confidence

Phase 1 results:
- Confirmed 20% engagement improvement
- Confidence: Updated from 50% → 90%

Phase 2: Full development (15 person-months)
- Roll out to all users
- Invest with high Confidence

Outcome:

Through staged Confidence updates: - Minimized initial investment - Reduced uncertainty with real data - Limited losses in case of failure - Full investment after success confirmation

This is fusion of MVP philosophy and Realization First Principle

The Power of RICE - Evidence Effectiveness

Power 1: Objectifying Subjective Judgment

Traditional problem:

"This feature is absolutely important!"
→ Why important?
→ "Gut feeling"
→ Rebuttal also "gut feeling"
→ Endless argument

RICE solution:

"Reach=100, Impact=3, Confidence=80%, Effort=10"
→ Score = 24

"Reach=10000, Impact=1, Confidence=100%, Effort=2"
→ Score = 5,000

Clear difference shown by numbers
→ Eliminates emotional reasoning

Power 2: Accelerating Organizational Consensus Building

Dropbox case:

Pre-introduction: Average 2 weeks for feature addition decisions
Reason: Inter-departmental interest coordination, endless meetings

Post-introduction: Decision-making shortened to average 2 days
Reason: RICE calculation becomes common language

Process:

1. Each department proposes initiatives (with RICE)
2. Automatically ranked by score
3. Implement top N items
4. If objections exist, debate "the numbers"

Power 3: Optimizing Resource Allocation

GitHub report:

1 year after RICE introduction:
- Team productivity improved 30%
- User satisfaction (NPS) improved +15 points
- Development time for "unimportant features" reduced 70%

Reason:
Concentrated investment in high-score initiatives
Courageous abandonment of low-score initiatives

Limitations and Cautions of RICE - Investigation Warnings

Limitation 1: Quantification Accuracy Issues

Reach estimation error:
Prediction: 1,000 people/month
Actual: 500 people/month
→ 50% error

Impact subjectivity:
Evaluator A: Impact=2
Evaluator B: Impact=1
→ 2x difference

Countermeasures:

• Continuously improve estimation accuracy with historical data
• Standardize Impact criteria with Baseline of Measurement (BOM)
• Regular score re-evaluation

Limitation 2: Insufficient Consideration of Strategic Importance

Initiative X: RICE Score=10
→ Low but "essential as future foundational technology"

Initiative Y: RICE Score=1000
→ High but "inconsistent with strategic direction"

Countermeasures:

RICE is "tactical level" prioritization
Strategic level requires separate judgment

Method:
1. First confirm strategically essential initiatives
2. Implement remaining resources in RICE score order

Limitation 3: Overlooking Qualitative Value

Brand value, team learning, technical debt resolution
→ Difficult to quantify with RICE
→ Tendency for scores to come out low

Countermeasures:

"RICE+α" judgment:
- RICE Score: 70% weight
- Qualitative value: 30% weight
- Adjust in final decision

Caution: Danger of Mechanical Application

❌ "Mechanically implement in score order"
✅ "Use scores as decision material, decide comprehensively"

RICE is a tool for "democratization," not "dictatorship"
Final judgment made by humans

Related Case Files - Connecting Ciphers

Joint Investigation 1: Integration with MVP

MVP determines "what to build" ↓ RICE determines "in what order to build" ↓ Agile Development executes "how to build"

Joint Investigation 2: Combination with Baseline of Measurement (BOM)

BOM clarifies Impact scale standards
→ Minimizes team Impact evaluation discrepancies
→ Improves RICE calculation accuracy

Joint Investigation 3: Effect Measurement with HEART Framework

RICE decides priorities
↓
Implementation and release
↓
HEART measures effectiveness
(Happiness, Engagement, Adoption, Retention, Task Success)
↓
Improve Reach, Impact, Effort estimation accuracy with actual data

Industry-Specific Implementation Patterns - Field-Based Investigation

SaaS Companies: Slack Case

Characteristic: Enormous feature requests
Challenge: Enterprise vs small business needs conflict

RICE utilization:
- Calculate Reach by segment
- Enterprise: 100 companies × average 1,000 users = 100,000 people
- Small business: 10,000 companies × average 10 users = 100,000 people
- → Reach equivalent, judge by combination with Impact

Result: Calm data-driven judgment, improved fairness between segments

E-commerce: Amazon's Application

Characteristic: Abundant A/B test data
Challenge: Small improvements vs big transformations prioritization

RICE utilization:
- Reach is measured value (data by user visit page)
- Impact evaluated by purchase rate change in A/B tests
- Confidence consistently above 90% (data-driven)

Result: Numerical justification for "1% improvement" accumulation strategy

Startups: Airbnb Founding Period

Characteristic: Minimal resources, can't fail
Challenge: What to prioritize with limited people

RICE utilization:
- Emphasize Effort most (1 person-month precious with small team)
- Prioritize initiatives with Effort=0.5 person-months or less
- Accumulate "Quick Wins" to build momentum

Result: 20 small improvements in 3 months, dramatic UX improvement

Investigation Summary - Essence of the Case

Final Analysis: The Fundamental Problem RICE Solves

Three diseases of organizational decision-making:

Disease 1: HiPPO disease (authoritarian decisions)
Disease 2: Endless debate disease (never-ending discussions)
Disease 3: Regret disease (post-hoc "should have done")

Treatment through RICE:

Treatment 1: Democratization through formulas
→ Position and loudness neutralized
→ Data is sole judgment criterion

Treatment 2: Establishing common language
→ Debate shifts from "feeling" to "numbers"
→ Consensus building dramatically accelerates

Treatment 3: Accountability through transparency
→ "Why we chose this" clear
→ Retrospective verification and learning possible

True Value: Not a Perfect Formula, but a Dialogue Protocol

The essence of RICE:

❌ "A magic formula that calculates perfectly correct priorities"
✅ "A structured dialogue method for teams to discuss in common language"

Numerical precision < Process transparency
Calculation accuracy < Consensus building speed

Detective's Final Conclusion:

The RICE Framework is a "democratization device for prioritization."

(Reach × Impact × Confidence) ÷ Effort

What this simple formula brings to organizations is not
"the correct answer" but
"a process to derive an answer everyone can accept."

Eliminating emotion, politics, and authority,
A culture of decision-making through data and logic.

That is the true reason this framework
continues to be adopted by organizations worldwide.

Case closed.
Your next prioritization will no longer be lost.

【🔏CLASSIFIED FILE END】