Use when you have a written implementation plan to execute in a separate session with review checkpoints
npx skills add miles-knowbl/orchestrator --skill "priority-matrix"
Install specific skill from multi-skill repository
# Description
Rank opportunities and work threads across multiple prioritization dimensions. Produces actionable priority matrices with calibrated scores, bias-checked rationale, and stakeholder-aligned rankings. Creates the prioritized backlog that feeds into proposal building.
# SKILL.md
name: priority-matrix
description: "Rank opportunities and work threads across multiple prioritization dimensions. Produces actionable priority matrices with calibrated scores, bias-checked rationale, and stakeholder-aligned rankings. Creates the prioritized backlog that feeds into proposal building."
phase: IMPLEMENT
category: specialized
version: "2.0.0"
depends_on: ["context-cultivation"]
tags: [planning, prioritization, analysis, decision-making, frameworks]
Priority Matrix
Rank opportunities and work threads across multiple prioritization dimensions.
When to Use
- Multiple competing opportunities -- Several options identified, need systematic ranking
- Resource allocation decisions -- Limited capacity, must choose what to pursue first
- Stakeholder alignment -- Different stakeholders advocate for different priorities
- Post-cultivation synthesis -- Context cultivation surfaced many threads, need to converge
- Strategic planning -- Quarterly or initiative-level prioritization across a portfolio
- When you say: "prioritize these", "what should we focus on", "rank the options", "which one first?"
Reference Requirements
MUST read before applying this skill:
| Reference | Why Required |
|---|---|
prioritization-frameworks.md |
Catalog of frameworks (RICE, MoSCoW, ICE, Eisenhower) to select from |
scoring-calibration.md |
Techniques for consistent, unbiased scoring across options |
Read if applicable:
| Reference | When Needed |
|---|---|
bias-mitigation.md |
When scoring involves subjective judgment or political dynamics |
stakeholder-alignment.md |
When multiple stakeholders must agree on priorities |
effort-estimation.md |
When effort dimension requires detailed estimation |
Verification: Ensure at least one scored matrix is produced with documented rationale for every score.
Required Deliverables
| Deliverable | Location | Condition |
|---|---|---|
PRIORITIES.md |
Project root | Always -- ranked list with rationale |
MATRIX.md |
Project root | Always -- full scoring matrix with dimension breakdowns |
CALIBRATION-LOG.md |
Project root | When 5+ options are scored -- documents calibration decisions |
Core Concept
Priority Matrix answers: "What should we do first, and why?"
Prioritization decisions are:
- Systematic -- Use structured frameworks, not gut feeling
- Transparent -- Every score has documented rationale
- Calibrated -- Scores are consistent across options and dimensions
- Stakeholder-aligned -- Methodology agreed upon before scoring begins
- Actionable -- Results in a clear, ordered backlog with next steps
Prioritization is NOT:
- Voting or popularity contests (that is opinion gathering)
- One-time decisions set in stone (priorities evolve; re-score periodically)
- A substitute for strategy (strategy defines goals; prioritization orders the path)
- Detailed project planning (that is spec and scaffold)
The Priority Matrix Process
+---------------------------------------------------------------+
| PRIORITY MATRIX PROCESS |
| |
| 1. OPTION ENUMERATION |
| +-> Collect and define all candidate work threads |
| |
| 2. FRAMEWORK SELECTION |
| +-> Choose the right prioritization framework |
| |
| 3. DIMENSION DEFINITION |
| +-> Define and weight scoring dimensions |
| |
| 4. CALIBRATION |
| +-> Establish scoring anchors and reference points |
| |
| 5. SCORING |
| +-> Score each option on each dimension with rationale |
| |
| 6. AGGREGATION |
| +-> Calculate weighted scores and produce rankings |
| |
| 7. BIAS CHECK |
| +-> Review for cognitive biases and political distortion |
| |
| 8. VALIDATION |
| +-> Sanity check results against intuition and strategy |
+---------------------------------------------------------------+
Step 1: Option Enumeration
Before scoring anything, build the complete list of candidates.
| Aspect | Questions |
|---|---|
| Completeness | Have we captured every thread from context cultivation? |
| Granularity | Are options at a comparable level of scope? |
| Independence | Can each option be pursued independently? |
| Clarity | Is each option defined well enough to score? |
| Duplicates | Have we merged overlapping or redundant options? |
Option Definition Template
For each candidate, capture enough context to score it:
### Option: [Name]
**Description:** [One-paragraph summary of what this work thread entails]
**Origin:** [Where this option came from -- cultivation theme, stakeholder request, gap analysis]
**Scope estimate:** [T-shirt size: XS / S / M / L / XL]
**Dependencies:** [Other options or external factors this depends on]
**Stakeholder champion:** [Who is advocating for this option, if anyone]
Enumeration Checklist
- [ ] All context cultivation themes represented
- [ ] All stakeholder requests captured
- [ ] All gap analysis items included
- [ ] Options are at comparable granularity
- [ ] Duplicates and overlaps merged
- [ ] Each option has a clear one-line description
- [ ] Total option count is manageable (3-15 is ideal)
If you have more than 15 options: Group them into categories first, prioritize categories, then prioritize within the top 2-3 categories. Avoid scoring 20+ items in a single matrix -- cognitive load degrades scoring quality.
Step 2: Framework Selection
Choose the prioritization framework that fits your context. No single framework is universally best.
Framework Comparison
| Framework | Best For | Dimensions | Complexity |
|---|---|---|---|
| Weighted Scoring | General-purpose, customizable | Custom (3-6) | Medium |
| RICE | Product features, growth teams | Reach, Impact, Confidence, Effort | Medium |
| ICE | Quick scoring, startups | Impact, Confidence, Ease | Low |
| MoSCoW | Scope negotiation, fixed timelines | Must/Should/Could/Won't | Low |
| Eisenhower | Personal or team task triage | Urgency, Importance | Low |
| Cost of Delay | Economic optimization | Value, Time criticality, Risk reduction | High |
Framework Selection Guide
Use Weighted Scoring when:
- You need full customization of dimensions
- Stakeholders need to see transparent methodology
- Options span diverse categories (technical, business, operational)
Use RICE when:
- Prioritizing product features or growth initiatives
- You have data on reach (users affected) and can estimate confidence
- The team is familiar with product management practices
Use ICE when:
- Speed matters more than precision
- Early-stage exploration with high uncertainty
- You need a quick first pass before deeper analysis
Use MoSCoW when:
- Working with a fixed deadline or scope boundary
- Stakeholders think in terms of "must have" vs "nice to have"
- Negotiating scope for an MVP or release
Use Eisenhower when:
- Triaging a mix of urgent and important work
- Team is overwhelmed and needs to cut non-essential work
- Distinguishing between reactive and proactive priorities
Use Cost of Delay when:
- Options have different time sensitivities
- Delay has measurable economic impact
- You want to optimize sequencing, not just ranking
Framework Decision Record
### Framework Choice: [Name]
**Selected:** [Framework name]
**Reason:** [Why this framework fits]
**Alternatives considered:** [What else was considered and why rejected]
**Adaptations:** [Any modifications to the standard framework]
Step 3: Dimension Definition
Define the scoring dimensions and their weights. Dimensions should be:
- Independent -- Avoid double-counting the same factor
- Measurable -- Each dimension must be scorable (even if subjective)
- Relevant -- Every dimension should matter for this specific decision
- Balanced -- No single dimension should dominate unless intentional
Default Weighted Scoring Dimensions
| Dimension | Weight | Description | Score Guide |
|---|---|---|---|
| Impact | 30% | Business value delivered if completed | 1=negligible, 5=moderate, 10=transformative |
| Effort | 25% | Resources required (inverse: high effort = low score) | 1=massive, 5=moderate, 10=trivial |
| Urgency | 25% | Time sensitivity and cost of delay | 1=can wait indefinitely, 5=this quarter, 10=this week |
| Alignment | 20% | Fit with strategy, goals, and capabilities | 1=off-strategy, 5=partially aligned, 10=core to strategy |
RICE Dimensions
| Dimension | Description | Scale |
|---|---|---|
| Reach | How many users/customers affected per quarter | Absolute number (e.g., 500, 5000) |
| Impact | How much each person is affected | 0.25=minimal, 0.5=low, 1=medium, 2=high, 3=massive |
| Confidence | How sure are we about the estimates | 50%=low, 80%=medium, 100%=high |
| Effort | Person-months of work required | Absolute number (lower = better) |
RICE score = (Reach x Impact x Confidence) / Effort
ICE Dimensions
| Dimension | Description | Scale |
|---|---|---|
| Impact | How much will this move the needle | 1-10 |
| Confidence | How sure are we this will work | 1-10 |
| Ease | How easy is this to implement | 1-10 |
ICE score = Impact x Confidence x Ease
Custom Dimension Template
### Dimension: [Name]
**Weight:** [Percentage]
**Description:** [What this dimension measures]
**Scoring guide:**
- 1-2: [Low anchor description]
- 3-4: [Below average description]
- 5-6: [Average description]
- 7-8: [Above average description]
- 9-10: [High anchor description]
**Example scores:**
- [Concrete example] = [Score] because [rationale]
- [Concrete example] = [Score] because [rationale]
Weight Allocation Checklist
- [ ] Weights sum to 100%
- [ ] No single dimension exceeds 40% (unless deliberately chosen)
- [ ] Stakeholders agree on dimension definitions
- [ ] Stakeholders agree on weight distribution
- [ ] Score guides are documented before scoring begins
Step 4: Calibration
Calibration ensures scores are consistent and meaningful. This is the most overlooked step and the most important for producing trustworthy results.
Calibration Techniques
| Technique | Description | When to Use |
|---|---|---|
| Anchor setting | Score 2-3 reference options first as benchmarks | Always (5+ options) |
| End-point definition | Define concrete examples for 1, 5, and 10 | Always |
| Independent scoring | Multiple people score independently, then compare | When reducing individual bias |
| Relative ranking first | Rank options within each dimension before assigning numbers | When absolute scoring feels arbitrary |
| Score normalization | Adjust scores to use the full range (avoid clustering at 7-8) | When scores bunch together |
Anchor Setting Process
- Choose 2-3 well-understood options as anchors
- Score anchors first on all dimensions
- Use anchors as reference points when scoring remaining options
- If a new score feels wrong relative to an anchor, adjust
### Calibration Anchors
| Anchor Option | Dimension | Score | Rationale |
|---------------|-----------|-------|-----------|
| [Well-known option A] | Impact | 8 | [Why this is an 8] |
| [Well-known option B] | Impact | 3 | [Why this is a 3] |
| [Well-known option A] | Effort | 5 | [Why this is a 5] |
| [Well-known option B] | Effort | 9 | [Why this is a 9] |
Score Distribution Health Check
After scoring, verify the distribution looks reasonable:
- [ ] Scores use at least 60% of the 1-10 range (not all clustered at 6-8)
- [ ] At least one option scores below 5 on some dimension
- [ ] At least one option scores above 8 on some dimension
- [ ] No dimension has all identical scores (that dimension is useless)
- [ ] Anchor options still feel correctly placed relative to others
Step 5: Scoring
Score every option on every dimension. Document rationale for every score.
Scoring Template
## Scores: [Option Name]
| Dimension | Score | Rationale |
|-----------|-------|-----------|
| Impact | [1-10] | [Why this score -- specific evidence or reasoning] |
| Effort | [1-10] | [Why this score -- reference comparable past work] |
| Urgency | [1-10] | [Why this score -- what happens if we delay] |
| Alignment | [1-10] | [Why this score -- which strategic goals does it serve] |
Scoring Discipline Rules
- Score one dimension at a time across all options -- Not one option at a time. This improves consistency.
- Write rationale before moving to the next score -- Forces deliberate thinking.
- Reference anchors frequently -- "Is this more or less impactful than Anchor A?"
- Flag low-confidence scores -- Mark scores where you are guessing with a
(?)suffix. - Separate fact from opinion -- Note when a score is data-driven vs. judgment-based.
Handling Uncertainty
| Confidence Level | Approach |
|---|---|
| High (data available) | Score directly, cite the data |
| Medium (informed judgment) | Score with rationale, flag as estimate |
| Low (guessing) | Use range (e.g., 4-7), take midpoint, flag for validation |
| Unknown | Assign neutral score (5), heavily flag, seek more information |
Step 6: Aggregation
Calculate weighted scores and produce the final ranking.
Weighted Score Calculation
For each option: Final Score = SUM(dimension_score x dimension_weight)
Full Matrix Template
# Priority Matrix
**Framework:** [Framework name]
**Date:** [Date scored]
**Scored by:** [Who participated]
## Dimension Weights
| Dimension | Weight |
|-----------|--------|
| Impact | 30% |
| Effort | 25% |
| Urgency | 25% |
| Alignment | 20% |
## Scoring Matrix
| # | Option | Impact | Effort | Urgency | Alignment | Weighted Score | Rank |
|---|--------|--------|--------|---------|-----------|----------------|------|
| 1 | Option Alpha | 9 | 7 | 8 | 9 | **8.25** | 1 |
| 2 | Option Beta | 8 | 6 | 9 | 7 | **7.65** | 2 |
| 3 | Option Gamma | 7 | 9 | 5 | 8 | **7.10** | 3 |
| 4 | Option Delta | 6 | 8 | 4 | 6 | **5.90** | 4 |
## Score Breakdown
[Include rationale tables from Step 5 for each option]
RICE Score Calculation
| Option | Reach | Impact | Confidence | Effort | RICE Score | Rank |
|--------|-------|--------|------------|--------|------------|------|
| Alpha | 5000 | 2 | 80% | 3 | 2667 | 1 |
| Beta | 2000 | 3 | 100% | 2 | 3000 | -- |
Note: RICE produces absolute scores, not 1-10 normalized scores. Rankings compare relative positions.
Tie-Breaking Rules
When two options have the same or very close scores (within 5%):
- Impact wins -- Higher impact option ranks higher
- Effort wins -- If impact is equal, lower effort option ranks higher
- Confidence wins -- If still tied, higher confidence option ranks higher
- Stakeholder tiebreak -- Escalate to the decision maker
Step 7: Bias Check
After scoring and ranking, explicitly check for common cognitive biases that distort prioritization.
Bias Checklist
| Bias | Description | Check Question |
|---|---|---|
| Anchoring | Over-weighting the first information received | Did the order we scored options affect results? |
| Recency | Favoring recently discussed options | Are older but valid options ranked too low? |
| Sunk cost | Favoring options with prior investment | Are we scoring future value, not past spend? |
| Champion bias | Favoring options pushed by senior stakeholders | Would the ranking change if options were anonymous? |
| Availability | Overweighting vivid or memorable options | Are less flashy options getting fair scores? |
| Groupthink | Conforming to perceived team consensus | Did anyone score independently before group discussion? |
| Optimism | Underestimating effort, overestimating impact | Are effort scores realistic? Compare to past projects. |
| Status quo | Favoring options that preserve current state | Are transformative options getting fair impact scores? |
Bias Mitigation Techniques
- [ ] Score independently before group discussion
- [ ] Randomize the order options are presented
- [ ] Assign a "devil's advocate" for top-ranked options
- [ ] Compare effort estimates to actual past effort on similar work
- [ ] Ask: "Would we still rank this #1 if [champion] were not advocating?"
- [ ] Ask: "What would need to be true for the bottom option to be #1?"
- [ ] Re-score any flagged low-confidence scores after bias review
See
references/bias-mitigation.md
Step 8: Validation
The matrix produces a ranking. Validation ensures the ranking makes sense.
Validation Questions
- [ ] Does the #1 option feel right? If not, what is the matrix missing?
- [ ] Would a reasonable stakeholder object to the top 3? Why?
- [ ] Do the bottom 3 genuinely deserve to be deprioritized?
- [ ] Does the ranking align with stated strategy and goals?
- [ ] Are there dependencies that make the ranking impractical?
- [ ] Is there an obvious "do first" option the matrix ranked low?
- [ ] Has the methodology been explained and accepted by stakeholders?
The Gut Check Test
If the matrix ranking contradicts strong intuition, do not simply override it. Instead:
- Identify the gap -- Which dimension is causing the surprising result?
- Check scoring -- Was that dimension scored correctly for the surprising option?
- Check weighting -- Should that dimension's weight be adjusted?
- Resolve explicitly -- Either fix the model or document why intuition is wrong
### Gut Check: [Surprising Result]
**Expected rank:** [What you expected]
**Actual rank:** [What the matrix produced]
**Root cause:** [Which dimension/weight caused the surprise]
**Resolution:** [Adjusted score/weight OR accepted the matrix result because...]
Dependency Validation
Even if Option A outranks Option B, if A depends on B, then B must come first. After ranking:
- [ ] Check for dependency cycles (A needs B, B needs A)
- [ ] Adjust execution order for hard dependencies
- [ ] Note where parallel execution is possible
- [ ] Document any re-ordering and rationale
Output Formats
Quick Priority List (3-5 Options)
# Priorities: [Context]
**Date:** [Date]
**Framework:** ICE (quick scoring)
## Priority Stack
1. **[Option Name]** (Score: 8.5)
- Why first: [One-line rationale]
- Next action: [Concrete next step]
2. **[Option Name]** (Score: 7.2)
- Why second: [One-line rationale]
- Next action: [Concrete next step]
3. **[Option Name]** (Score: 6.8)
- Why third: [One-line rationale]
- Next action: [Concrete next step]
## Deprioritized
- **[Option Name]** (Score: 4.1) -- [Why not now]
- **[Option Name]** (Score: 3.5) -- [Why not now]
## Methodology
[Brief description of framework and dimensions used]
Full Priority Matrix (6+ Options)
# Priority Matrix: [Context]
**Date:** [Date]
**Framework:** [Framework name]
**Scored by:** [Participants]
**Confidence:** [Overall confidence level]
## Executive Summary
[Top 3 priorities in one paragraph. Key trade-offs and rationale.]
## Priority Stack
### Tier 1: Do Now
| Rank | Option | Score | Key Rationale |
|------|--------|-------|---------------|
| 1 | [Name] | [Score] | [Why] |
| 2 | [Name] | [Score] | [Why] |
### Tier 2: Do Next
| Rank | Option | Score | Key Rationale |
|------|--------|-------|---------------|
| 3 | [Name] | [Score] | [Why] |
| 4 | [Name] | [Score] | [Why] |
### Tier 3: Do Later
| Rank | Option | Score | Key Rationale |
|------|--------|-------|---------------|
| 5 | [Name] | [Score] | [Why] |
| 6 | [Name] | [Score] | [Why] |
### Not Now
| Option | Score | Why Deprioritized |
|--------|-------|-------------------|
| [Name] | [Score] | [Reason] |
## Full Scoring Matrix
[Include the complete matrix table from Step 6]
## Dimension Definitions and Weights
[Include the dimension table from Step 3]
## Calibration Notes
[Include anchors and calibration decisions from Step 4]
## Score Rationale
[Include per-option rationale tables from Step 5]
## Bias Review
[Include completed bias checklist from Step 7]
## Validation Notes
[Include completed validation checklist from Step 8]
## Execution Dependencies
[Dependency map and any re-ordering from validation]
## Next Steps
1. [Concrete action for priority #1]
2. [Concrete action for priority #2]
3. [When to re-score -- trigger conditions]
Common Patterns
The Portfolio Prioritization
Score strategic initiatives across business dimensions: revenue impact, customer satisfaction, operational efficiency, competitive positioning.
Use when: Annual or quarterly planning, allocating a fixed budget across initiatives.
The Feature Backlog Triage
Apply RICE scoring to a product feature backlog. Emphasize reach and confidence to avoid building low-impact, high-uncertainty features.
Use when: Sprint planning, product roadmap creation, feature grooming sessions.
The Technical Debt Matrix
Score technical debt items on: blast radius (what breaks if ignored), fix effort, frequency of pain, and dependency blocking. Weight blast radius and dependency blocking heavily.
Use when: Deciding which tech debt to address during a maintenance sprint or platform stabilization.
The Opportunity Assessment
Combine market analysis with feasibility scoring. Dimensions include: market size, competitive intensity, capability fit, time to value.
Use when: Evaluating new business opportunities, partnership proposals, or market entry decisions.
Relationship to Other Skills
| Skill | Relationship |
|---|---|
context-cultivation |
Cultivation outputs become the option list for prioritization |
context-ingestion |
Raw context feeds cultivation which feeds prioritization |
proposal-builder |
Priorities feed directly into proposal scope and sequencing |
architect |
Top-priority options may need architectural analysis |
spec |
After prioritization, top items move into specification |
implement |
Execution order follows the priority stack |
code-review |
Review effort can be prioritized using a lightweight matrix |
Key Principles
Methodology before scoring. Agree on framework, dimensions, and weights before anyone scores a single option. Changing methodology mid-scoring invalidates prior scores.
Rationale is the product. The ranked list is useful; the documented rationale is invaluable. Future prioritization decisions build on today's reasoning, not just today's numbers.
Calibrate relentlessly. Uncalibrated scores are meaningless. Two people giving "7 out of 10" may mean completely different things. Anchors and examples make scores comparable.
Separate scoring from advocacy. The person who proposed an option should not be the only person scoring it. Separation reduces champion bias and increases trust in results.
Prioritization is perishable. A priority matrix has a shelf life. Re-score when new information emerges, when market conditions change, or at regular intervals (monthly or quarterly).
Explicit is better than implicit. A mediocre framework applied transparently beats a brilliant decision made in someone's head. The matrix makes reasoning visible, challengeable, and improvable.
References
references/prioritization-frameworks.md: Detailed framework implementations (RICE, MoSCoW, ICE, Eisenhower, Cost of Delay)references/scoring-calibration.md: Calibration techniques, anchor setting, score normalizationreferences/bias-mitigation.md: Cognitive biases in prioritization and countermeasuresreferences/stakeholder-alignment.md: Techniques for building consensus on methodology and resultsreferences/effort-estimation.md: Effort estimation approaches for the effort/ease dimension
# Supported AI Coding Agents
This skill is compatible with the SKILL.md standard and works with all major AI coding agents:
Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.