20 KiB
Prioritization: Advanced Methodologies
Table of Contents
- Advanced Scoring Frameworks
- Alternative Prioritization Models
- Stakeholder Alignment Techniques
- Data-Driven Prioritization
- Roadmap Optimization
- Common Pitfalls and Solutions
1. Advanced Scoring Frameworks
RICE Score (Reach × Impact × Confidence / Effort)
Formula: Score = (Reach × Impact × Confidence) / Effort
When to use: Product backlogs where user reach and confidence matter more than simple impact
Components:
- Reach: Users/customers affected per time period (e.g., 1000 users/quarter)
- Impact: Value per user (1=minimal, 2=low, 3=medium, 4=high, 5=massive)
- Confidence: How certain are we? (100%=high, 80%=medium, 50%=low)
- Effort: Person-months (e.g., 2 = 2 engineers for 1 month, or 1 engineer for 2 months)
Example:
- Feature A: (5000 users/qtr × 3 impact × 100% confidence) / 2 effort = 7500 score
- Feature B: (500 users/qtr × 5 impact × 50% confidence) / 1 effort = 1250 score
- Decision: Feature A scores 6× higher despite lower per-user impact
Advantages: More nuanced than 2x2 matrix, accounts for uncertainty Disadvantages: Requires estimation of reach (hard for new features)
ICE Score (Impact × Confidence × Ease)
Formula: Score = Impact × Confidence × Ease
When to use: Growth experiments, marketing campaigns, quick prioritization
Components:
- Impact: Potential value (1-10 scale)
- Confidence: How sure are we this will work? (1-10 scale)
- Ease: How easy to implement? (1-10 scale, inverse of effort)
Example:
- Experiment A: 8 impact × 9 confidence × 7 ease = 504 score
- Experiment B: 10 impact × 3 confidence × 5 ease = 150 score
- Decision: A scores higher due to confidence, even with lower max impact
Advantages: Simpler than RICE, no time period needed Disadvantages: Multiplicative scoring can exaggerate differences
Weighted Scoring (Custom Criteria)
When to use: Complex decisions with multiple evaluation dimensions beyond effort/impact
Process:
- Define criteria (e.g., Revenue, Strategic Fit, User Pain, Complexity, Risk)
- Weight each criterion (sum to 100%)
- Score each item on each criterion (1-5)
- Calculate weighted score:
Score = Σ(criterion_score × criterion_weight)
Example:
| Criteria | Weight | Feature A Score | Feature A Weighted | Feature B Score | Feature B Weighted |
|---|---|---|---|---|---|
| Revenue | 40% | 4 | 1.6 | 2 | 0.8 |
| Strategic | 30% | 5 | 1.5 | 4 | 1.2 |
| User Pain | 20% | 3 | 0.6 | 5 | 1.0 |
| Complexity | 10% | 2 (low) | 0.2 (penalty) | 4 (high) | 0.4 (penalty) |
| Total | 100% | - | 3.9 | - | 3.0 |
Decision: Feature A scores higher (3.9 vs 3.0) due to revenue and strategic fit
Advantages: Transparent, customizable to organization's values Disadvantages: Can become analysis paralysis with too many criteria
Kano Model (Customer Satisfaction)
When to use: Understanding which features delight vs must-have vs don't-matter
Categories:
- Must-Be (Basic): Absence causes dissatisfaction, presence is expected (e.g., product works, no security bugs)
- Performance (Linear): More is better, satisfaction increases linearly (e.g., speed, reliability)
- Delighters (Excitement): Unexpected features that wow users (e.g., dark mode before it was common)
- Indifferent: Users don't care either way (e.g., internal metrics dashboard for end users)
- Reverse: Some users actively dislike (e.g., forced tutorials, animations)
Survey technique:
- Ask functional question: "How would you feel if we had feature X?" (Satisfied / Expected / Neutral / Tolerate / Dissatisfied)
- Ask dysfunctional question: "How would you feel if we did NOT have feature X?"
- Map responses to category
Prioritization strategy:
- Must-Be first: Fix broken basics (dissatisfiers)
- Delighters for differentiation: Quick wins that wow
- Performance for competitiveness: Match or exceed competitors
- Avoid indifferent/reverse: Don't waste time
Example:
- Must-Be: Fix crash on login (everyone expects app to work)
- Performance: Improve search speed from 2s to 0.5s
- Delighter: Add "undo send" for emails (unexpected delight)
- Indifferent: Add 50 new color themes (most users don't care)
Value vs Complexity (Alternative Labels)
When to use: Same as effort-impact, but emphasizes business value and technical complexity
Axes:
- X-axis: Complexity (technical difficulty, dependencies, unknowns)
- Y-axis: Value (business value, strategic value, user value)
Quadrants (same concept, different framing):
- High Value, Low Complexity = Quick Wins (same)
- High Value, High Complexity = Strategic Investments (same as Big Bets)
- Low Value, Low Complexity = Nice-to-Haves (same as Fill-Ins)
- Low Value, High Complexity = Money Pits (same as Time Sinks)
When to use this framing: Technical audiences respond to "complexity", business audiences to "value"
2. Alternative Prioritization Models
MoSCoW (Must/Should/Could/Won't)
When to use: Fixed-scope projects (e.g., product launch, migration) with deadline constraints
Categories:
- Must Have: Non-negotiable, project fails without (e.g., core functionality, security, legal requirements)
- Should Have: Important but not critical, defer if needed (e.g., nice UX, analytics)
- Could Have: Desirable if time/budget allows (e.g., polish, edge cases)
- Won't Have (this time): Out of scope, revisit later (e.g., advanced features, integrations)
Process:
- List all requirements
- Stakeholders categorize each (force hard choices: only 60% can be Must/Should)
- Build Must-Haves first, then Should-Haves, then Could-Haves if time remains
- Communicate Won't-Haves to set expectations
Example (product launch):
- Must: User registration, core workflow, payment processing, security
- Should: Email notifications, basic analytics, help docs
- Could: Dark mode, advanced filters, mobile app
- Won't: Integrations, API, white-labeling (v2 scope)
Advantages: Forces scope discipline, clear for deadline-driven work Disadvantages: Doesn't account for effort (may put high-effort items in Must)
Cost of Delay (CD3: Cost, Duration, Delay)
When to use: Time-sensitive decisions where delaying has quantifiable revenue/strategic cost
Formula: CD3 Score = Cost of Delay / Duration
- Cost of Delay: $/month of delaying this (revenue loss, market share, customer churn)
- Duration: Months to complete
Example:
- Feature A: $100K/mo delay cost, 2 months duration → $50K/mo score
- Feature B: $200K/mo delay cost, 5 months duration → $40K/mo score
- Decision: Feature A higher score (faster time-to-value despite lower total CoD)
When to use: Competitive markets, revenue-critical features, time-limited opportunities (e.g., seasonal)
Advantages: Explicitly values speed to market Disadvantages: Requires estimating revenue impact (often speculative)
Opportunity Scoring (Jobs-to-be-Done)
When to use: Understanding which user jobs are underserved (high importance, low satisfaction)
Survey:
- Ask users to rate each "job" on:
- Importance: How important is this outcome? (1-5)
- Satisfaction: How well does current solution satisfy? (1-5)
Opportunity Score: Importance + max(Importance - Satisfaction, 0)
- Score ranges 0-10
- >8 = High opportunity (important but poorly served)
- <5 = Low opportunity (either unimportant or well-served)
Example:
-
Job: "Quickly find relevant past conversations"
- Importance: 5 (very important)
- Satisfaction: 2 (very dissatisfied)
- Opportunity: 5 + (5-2) = 8 (high opportunity) → prioritize search improvements
-
Job: "Customize notification sounds"
- Importance: 2 (not important)
- Satisfaction: 3 (neutral)
- Opportunity: 2 + 0 = 2 (low opportunity) → deprioritize
Advantages: User-centric, identifies gaps between need and solution Disadvantages: Requires user research, doesn't account for effort
3. Stakeholder Alignment Techniques
Silent Voting (Avoid Anchoring Bias)
Problem: First person to score influences others (anchoring bias), HIPPO dominates Solution: Everyone scores independently, then discuss
Process:
- Each stakeholder writes scores on sticky notes (don't share yet)
- Reveal all scores simultaneously
- Discuss discrepancies (why did eng score Effort=5 but product scored Effort=2?)
- Converge on consensus score
Tools: Planning poker (common in agile), online voting (Miro, Mural)
Forced Ranking (Avoid "Everything is High Priority")
Problem: Stakeholders rate everything 4-5, no differentiation Solution: Force stack ranking (only one item can be #1)
Process:
- List all items
- Stakeholders must rank 1, 2, 3, ..., N (no ties allowed)
- Convert ranks to scores (e.g., top 20% = 5, next 20% = 4, middle 20% = 3, etc.)
Variant: $100 Budget:
- Give stakeholders $100 to "invest" across all items
- They allocate dollars based on priority ($30 to A, $25 to B, $20 to C, ...)
- Items with most investment are highest priority
Weighted Stakeholder Input (Account for Expertise)
Problem: Not all opinions are equal (eng knows effort, sales knows customer pain) Solution: Weight scores by expertise domain
Example:
- Effort score = 100% from Engineering (they know effort best)
- Impact score = 40% Product + 30% Sales + 30% Customer Success (all know value)
Process:
- Define who estimates what (effort, user impact, revenue, etc.)
- Assign weights (e.g., 60% engineering + 40% product for effort)
- Calculate weighted average:
Score = Σ(stakeholder_score × weight)
Pre-Mortem for Controversial Items
Problem: Stakeholders disagree on whether item is Quick Win or Time Sink Solution: Run pre-mortem to surface hidden risks/assumptions
Process:
- Assume item failed spectacularly ("We spent 6 months and it failed")
- Each stakeholder writes down "what went wrong" (effort blowups, impact didn't materialize)
- Discuss: Are these risks real? Should we adjust scores or descope?
Example:
- Item: "Build mobile app" (scored Impact=5, Effort=3)
- Pre-mortem reveals: "App store approval took 3 months", "iOS/Android doubled effort", "Users didn't download"
- Revised score: Impact=3 (uncertain), Effort=5 (doubled for two platforms) → Time Sink, defer
4. Data-Driven Prioritization
Usage Analytics (Prioritize High-Traffic Areas)
When to use: Product improvements where usage data is available
Metrics:
- Page views / Feature usage: Improve high-traffic areas first (more users benefit)
- Conversion funnel: Fix drop-off points with biggest impact (e.g., 50% drop at checkout)
- Support tickets: High ticket volume = high user pain
Example:
- Dashboard page: 1M views/mo, 10% bounce rate → 100K frustrated users → high impact
- Settings page: 10K views/mo, 50% bounce rate → 5K frustrated users → lower impact
- Decision: Fix dashboard first (20× more impact)
Advantages: Objective, quantifies impact Disadvantages: Ignores new features (no usage data yet), low-traffic areas may be high-value for specific users
A/B Test Results (Validate Impact Assumptions)
When to use: Uncertain impact; run experiment to measure before committing
Process:
- Build minimal version of feature (1-2 weeks)
- A/B test with 10% of users
- Measure impact (conversion, retention, revenue, NPS)
- If impact validated, commit to full version; if not, deprioritize
Example:
- Hypothesis: "Adding social login will increase signups 20%"
- A/B test: 5% increase (not 20%)
- Decision: Impact=3 (not 5 as assumed), deprioritize vs other items
Advantages: Reduces uncertainty, validates assumptions before big bets Disadvantages: Requires experimentation culture, time to run tests
Customer Request Frequency (Vote Counting)
When to use: Feature requests from sales, support, customers
Metrics:
- Request count: Number of unique customers asking
- Revenue at risk: Total ARR of customers requesting (enterprise vs SMB)
- Churn risk: Customers threatening to leave without feature
Example:
- Feature A: 50 SMB customers requesting ($5K ARR each) = $250K ARR
- Feature B: 2 Enterprise customers requesting ($200K ARR each) = $400K ARR
- Decision: Feature B higher impact (more revenue at risk) despite fewer requests
Guardrail: Don't just count votes (10 vocal users ≠ real demand), weight by revenue/strategic value
NPS/CSAT Drivers Analysis
When to use: Understanding which improvements drive customer satisfaction most
Process:
- Collect NPS/CSAT scores
- Ask open-ended: "What's the one thing we could improve?"
- Categorize feedback (performance, features, support, etc.)
- Correlate categories with NPS (which issues are mentioned by detractors most?)
Example:
- Detractors (NPS 0-6) mention "slow performance" 80% of time
- Passives (NPS 7-8) mention "missing integrations" 60%
- Decision: Fix performance first (bigger impact on promoter score)
5. Roadmap Optimization
Dependency Mapping (Critical Path)
Problem: Items with dependencies can't start until prerequisites complete Solution: Map dependency graph, identify critical path
Process:
- List all items with dependencies (A depends on B, C depends on A and D)
- Draw dependency graph (use tools: Miro, Mural, project management software)
- Identify critical path (longest chain of dependencies)
- Parallelize non-dependent work
Example:
- Quick Win A (2 weeks) → Big Bet B (8 weeks) → Quick Win C (1 week) = 11 week critical path
- Quick Win D (2 weeks, no dependencies) can run in parallel with A
- Optimized timeline: 11 weeks instead of 13 weeks (if sequential)
Team Velocity and Capacity Planning
Problem: Overcommitting to more work than team can deliver Solution: Use historical velocity to forecast capacity
Process:
- Measure past velocity (effort points completed per sprint/quarter)
- Estimate total capacity (team size × utilization × time period)
- Don't plan >70-80% of capacity (buffer for unknowns, support, bugs)
Example:
- Team completes 20 effort points/quarter historically
- Next quarter roadmap: 30 effort points planned
- Problem: 150% overcommitted
- Fix: Cut lowest-priority items (time sinks, fill-ins) to fit 16 effort points (80% of 20)
Guardrail: If you consistently complete <50% of roadmap, estimation is broken (not just overcommitted)
Incremental Delivery (Break Big Bets into Phases)
Problem: Big Bet takes 6 months; no value delivery until end Solution: Break into phases that deliver incremental value
Example:
- Original: "Rebuild reporting system" (6 months, Effort=5)
- Phased:
- Phase 1: Migrate 3 most-used reports (1 month, Effort=2, Impact=3)
- Phase 2: Add drill-down capability (1 month, Effort=2, Impact=4)
- Phase 3: Real-time data (2 months, Effort=3, Impact=5)
- Phase 4: Custom dashboards (2 months, Effort=3, Impact=3)
- Benefit: Ship value after 1 month (not 6), can adjust based on feedback
Advantages: Faster time-to-value, reduces risk, allows pivoting Disadvantages: Requires thoughtful phasing (some work can't be incrementalized)
Portfolio Balancing (Mix of Quick Wins and Big Bets)
Problem: Roadmap is all quick wins (no strategic depth) or all big bets (no momentum) Solution: Balance portfolio across quadrants and time horizons
Target distribution:
- 70% Quick Wins + Fill-Ins (short-term value, momentum)
- 30% Big Bets (long-term strategic positioning)
- OR by time: Now (0-3 months): 60%, Next (3-6 months): 30%, Later (6-12 months): 10%
Example:
- Q1: 5 quick wins + 1 big bet Phase 1
- Q2: 3 quick wins + 1 big bet Phase 2
- Q3: 4 quick wins + 1 new big bet start
- Result: Consistent value delivery (quick wins) + strategic progress (big bets)
6. Common Pitfalls and Solutions
Pitfall 1: Solo Scoring (No Stakeholder Input)
Problem: PM scores everything alone, misses engineering effort or sales context Solution: Multi-stakeholder scoring session (2-hour workshop)
Workshop agenda:
- 0-15min: Align on scoring scales (calibrate with examples)
- 15-60min: Silent voting on effort/impact for all items
- 60-90min: Discuss discrepancies, converge on consensus
- 90-120min: Plot matrix, identify quadrants, sequence roadmap
Pitfall 2: Analysis Paralysis (Perfect Scores)
Problem: Spending days debating if item is Effort=3.2 or Effort=3.4 Solution: Good-enough > perfect; prioritization is iterative
Guardrail: Limit scoring session to 2 hours; if still uncertain, default to conservative (higher effort, lower impact)
Pitfall 3: Ignoring Dependencies
Problem: Quick Win scored Effort=2, but depends on Effort=5 migration completing first Solution: Score standalone effort AND prerequisite effort separately
Example:
- Item: "Add SSO login" (Effort=2 standalone)
- Depends on: "Migrate to new auth system" (Effort=5)
- True effort: 5 (for new roadmaps) or 2 (if migration already planned)
Pitfall 4: Strategic Override ("CEO Wants It")
Problem: Exec declares item "high priority" without scoring, bypasses process Solution: Make execs participate in scoring, apply same framework
Response: "Let's score this using our framework so we can compare to other priorities. If it's truly high impact and low effort, it'll naturally rise to the top."
Pitfall 5: Sunk Cost Fallacy (Continuing Time Sinks)
Problem: "We already spent 2 months on X, we can't stop now" (even if impact is low) Solution: Sunk costs are sunk; evaluate based on future effort/impact only
Decision rule: If you wouldn't start this project today knowing what you know, stop it now
Pitfall 6: Neglecting Maintenance (Assuming 100% Project Capacity)
Problem: Roadmap plans 100% of team time, ignoring support/bugs/tech debt/meetings Solution: Reduce capacity by 20-50% for non-project work
Realistic capacity:
- 100% time - 20% support/bugs - 10% meetings - 10% code review/pairing = 60% project capacity
- If team is 5 people × 40 hrs/wk × 12 weeks = 2400 hrs, only 1440 hrs available for roadmap
Pitfall 7: Ignoring User Research (Opinion-Based Scoring)
Problem: Impact scores based on team intuition, not user data Solution: Validate impact with user research (surveys, interviews, usage data)
Quick validation:
- Survey 50 users: "How important is feature X?" (1-5)
- If <50% say 4-5, impact is not as high as assumed
- Adjust scores based on data
Pitfall 8: Scope Creep During Execution
Problem: Quick Win (Effort=2) grows to Big Bet (Effort=5) during implementation Solution: Timebox quick wins; if effort exceeds estimate, cut scope or defer
Guardrail: "If this takes >1 week, we stop and re-evaluate whether it's still worth it"
Pitfall 9: Forgetting to Say "No"
Problem: Roadmap keeps growing (never remove items), becomes unexecutable Solution: Explicitly cut time sinks, communicate what you're NOT doing
Communication template: "We prioritized X, Y, Z based on impact/effort. This means we're NOT doing A, B, C this quarter because [reason]. We'll revisit in [timeframe]."
Pitfall 10: One-Time Prioritization (Never Re-Score)
Problem: Scores from 6 months ago are stale (context changed, new data available) Solution: Re-score quarterly, adjust roadmap based on new information
Triggers for re-scoring:
- Quarterly planning cycles
- Major context changes (new competitor, customer churn, team size change)
- Big bets complete (update dependent items' scores)
- User research reveals new insights