decision-analysis（决策分析）

Decision Analysis

Model and analyze complex decisions using structured techniques: decision tables, weighted scoring matrices, and decision trees. Creates clear, defensible decision frameworks with traceable rationale.

What is Decision Analysis?

Decision Analysis is a systematic approach to evaluating complex choices by breaking them down into components: objectives, alternatives, criteria, and trade-offs. It transforms subjective judgment into structured, transparent reasoning.

Technique 1: Decision Tables

What is a Decision Table?

A decision table captures complex conditional logic in a compact grid format. It lists all combinations of conditions and their corresponding actions.

Decision Table Workflow

Step 1: Identify Conditions and Actions

## Decision Context

**Decision:** [What are we deciding?]
**Trigger:** [When is this decision made?]

### Conditions (Inputs)
| # | Condition | Possible Values |
|---|-----------|-----------------|
| C1 | [Condition 1] | [Value A / Value B / ...] |
| C2 | [Condition 2] | [Yes / No] |
| C3 | [Condition 3] | [Low / Medium / High] |

### Actions (Outputs)
| # | Action | Description |
|---|--------|-------------|
| A1 | [Action 1] | [What happens] |
| A2 | [Action 2] | [What happens] |

Step 2: Build the Decision Table

## Decision Table: [Name]

| Rule | C1 | C2 | C3 | A1 | A2 |
|------|----|----|----|----|----|
| R1 | Premium | Yes | High | X | - |
| R2 | Premium | Yes | Low | X | X |
| R3 | Standard | Yes | - | - | X |
| R4 | Standard | No | High | - | - |
| R5 | - | No | Low | - | X |

**Legend:** X = Execute action, - = Skip, [blank] = Any value

Step 3: Validate Completeness

Decision Table Template

## Decision Table: [Decision Name]

**Context:** [Business context]
**Owner:** [Decision owner]
**Last Updated:** [ISO date]

### Conditions
| ID | Condition | Values |
|----|-----------|--------|
| C1 | | |
| C2 | | |

### Actions
| ID | Action | Description |
|----|--------|-------------|
| A1 | | |
| A2 | | |

### Rules
| Rule | C1 | C2 | → | A1 | A2 | Notes |
|------|----|----|---|----|----|-------|
| R1 | | | | | | |
| R2 | | | | | | |

### Validation
- [ ] All combinations covered
- [ ] No contradictions
- [ ] Rules simplified

Technique 2: Weighted Scoring Matrix

What is a Weighted Scoring Matrix?

A weighted scoring matrix (also called decision matrix or Pugh matrix) evaluates multiple alternatives against weighted criteria to produce a ranked list.

Weighted Scoring Workflow

Step 1: Define the Decision

## Decision Context

**Decision:** [What are we choosing?]
**Objective:** [What outcome do we want?]
**Constraints:** [Non-negotiable requirements]
**Timeline:** [When must we decide?]

Step 2: Identify Alternatives

## Alternatives

| # | Alternative | Description | Source |
|---|-------------|-------------|--------|
| A | [Option A] | [Brief description] | [How identified] |
| B | [Option B] | [Brief description] | [How identified] |
| C | [Option C] | [Brief description] | [How identified] |

Step 3: Define and Weight Criteria

## Criteria

| # | Criterion | Description | Weight | Rationale |
|---|-----------|-------------|--------|-----------|
| 1 | [Criterion 1] | [What it measures] | 30% | [Why this weight] |
| 2 | [Criterion 2] | [What it measures] | 25% | [Why this weight] |
| 3 | [Criterion 3] | [What it measures] | 25% | [Why this weight] |
| 4 | [Criterion 4] | [What it measures] | 20% | [Why this weight] |
|   | **Total** | | **100%** | |

Weighting Methods:

Step 4: Score Alternatives

## Scoring Scale

| Score | Meaning |
|-------|---------|
| 5 | Excellent - Fully meets/exceeds criterion |
| 4 | Good - Mostly meets criterion |
| 3 | Adequate - Partially meets criterion |
| 2 | Poor - Minimally meets criterion |
| 1 | Unacceptable - Does not meet criterion |

Step 5: Calculate Weighted Scores

## Decision Matrix

| Criterion | Weight | Alt A | Alt B | Alt C |
|-----------|--------|-------|-------|-------|
| Criterion 1 | 30% | 4 | 3 | 5 |
| Criterion 2 | 25% | 3 | 5 | 4 |
| Criterion 3 | 25% | 5 | 4 | 3 |
| Criterion 4 | 20% | 4 | 4 | 4 |
| **Weighted Score** | | **3.95** | **3.95** | **4.05** |
| **Rank** | | 2 | 3 | 1 |

**Calculation:** Weighted Score = Σ(Score × Weight)

Step 6: Sensitivity Analysis

Test how results change if weights shift:

## Sensitivity Analysis

| Scenario | Weight Change | Winner | Confidence |
|----------|---------------|--------|------------|
| Baseline | As defined | Alt C | - |
| Cost +10% | C1: 40%, others adjusted | Alt A | Low |
| Quality +10% | C2: 35%, others adjusted | Alt C | High |

**Robustness:** [Is the winner stable across scenarios?]

Pugh Matrix (Concept Selection)

A specialized scoring matrix comparing alternatives to a baseline:

## Pugh Matrix: [Decision]

**Baseline:** [Reference option - usually current state or simplest option]

| Criterion | Weight | Alt A vs Baseline | Alt B vs Baseline | Alt C vs Baseline |
|-----------|--------|-------------------|-------------------|-------------------|
| Criterion 1 | 30% | + | S | ++ |
| Criterion 2 | 25% | - | + | S |
| Criterion 3 | 25% | S | + | - |
| Criterion 4 | 20% | + | S | + |
| **Σ Plus** | | 2 | 2 | 2 |
| **Σ Minus** | | 1 | 0 | 1 |
| **Σ Same** | | 1 | 2 | 1 |
| **Net Score** | | +1 | +2 | +1 |

**Legend:** ++ Much better, + Better, S Same, - Worse, -- Much worse

Technique 3: Decision Trees

What is a Decision Tree?

A decision tree maps sequential decisions and uncertain events to visualize possible paths and outcomes. It's ideal for decisions with multiple stages or probabilistic outcomes.

Decision Tree Workflow

Step 1: Frame the Decision

## Decision Tree Context

**Decision:** [Primary decision]
**Objective:** [What we're optimizing - NPV, utility, etc.]
**Time Horizon:** [How far into future]
**Key Uncertainties:** [Major unknown factors]

Step 2: Identify Decision Points and Uncertainties

## Structure

### Decision Points
| # | Decision | Options |
|---|----------|---------|
| D1 | [First decision] | Option A, Option B |
| D2 | [Subsequent decision] | Option X, Option Y |

### Chance Events
| # | Event | Outcomes | Probabilities |
|---|-------|----------|---------------|
| E1 | [Uncertainty 1] | High, Low | 60%, 40% |
| E2 | [Uncertainty 2] | Success, Failure | 70%, 30% |

Step 3: Assign Values and Probabilities

## Outcomes

| Path | Sequence | Probability | Value | Expected Value |
|------|----------|-------------|-------|----------------|
| P1 | D1:A → E1:High → D2:X | 0.60 | $100K | $60K |
| P2 | D1:A → E1:High → D2:Y | 0.60 | $80K | $48K |
| P3 | D1:A → E1:Low | 0.40 | $20K | $8K |
| P4 | D1:B → E2:Success | 0.70 | $150K | $105K |
| P5 | D1:B → E2:Failure | 0.30 | -$50K | -$15K |

Step 4: Calculate Expected Values (Rollback)

Work backwards from end nodes:

## Rollback Analysis

### Chance Node E1 (after D1:A)
EV = (0.60 × max($100K, $80K)) + (0.40 × $20K)
EV = (0.60 × $100K) + $8K = $68K

### Chance Node E2 (after D1:B)
EV = (0.70 × $150K) + (0.30 × -$50K)
EV = $105K - $15K = $90K

### Decision Node D1
Choose B: EV = $90K > $68K

**Recommendation:** Choose Option B

Decision Tree Mermaid Diagram

flowchart TD
    D1{Decision 1<br/>Choose A or B?}

    D1 -->|A| E1((Event 1<br/>Market))
    D1 -->|B| E2((Event 2<br/>Tech))

    E1 -->|High 60%| D2{Decision 2}
    E1 -->|Low 40%| OUT1[/$20K/]

    D2 -->|X| OUT2[/$100K/]
    D2 -->|Y| OUT3[/$80K/]

    E2 -->|Success 70%| OUT4[/$150K/]
    E2 -->|Failure 30%| OUT5[/-$50K/]

    style D1 fill:#ffcc00
    style D2 fill:#ffcc00
    style E1 fill:#66ccff
    style E2 fill:#66ccff

DMN-Lite: Decision Model Notation

For simple, repeatable decisions, use a lightweight DMN approach:

## Decision: [Name]

**Decision ID:** DEC-001
**Business Context:** [When this decision is made]

### Input Data
| Input | Type | Source |
|-------|------|--------|
| Customer Segment | Text | CRM |
| Order Value | Currency | Order System |
| Credit Score | Number | Credit Bureau |

### Decision Logic

IF Customer Segment = "Premium" AND Order Value > 1000 THEN Discount = 20% ELSE IF Customer Segment = "Premium" THEN Discount = 10% ELSE IF Order Value > 5000 THEN Discount = 15% ELSE THEN Discount = 0%

### Output

| Output | Type | Range |
| --- | --- | --- |
| Discount | Percentage | 0% - 20% |

## Output Formats

### Narrative Summary

Decision Analysis Summary

Decision: [What was decided] Date: [ISO date] Analyst: decision-analyst

Context

[2-3 sentences on why this decision was needed]

Approach

• Technique Used: [Decision Table / Weighted Matrix / Decision Tree]
• Alternatives Considered: [Count and brief list]
• Criteria Applied: [Count and key criteria]

Recommendation

Recommended Option: [Name] Rationale: [Key reasons - 2-3 points] Confidence: High / Medium / Low

Key Trade-offs

Risks and Mitigations

Next Steps

1. [Immediate action]
2. [Follow-up action]

### Structured Data (YAML)

decision_analysis: version: "1.0" date: "2025-01-15" analyst: "decision-analyst"

context: decision: "Select project management tool" objective: "Maximize team productivity while minimizing cost" constraints: - "Budget under $500/month" - "Must integrate with team messaging platform" timeline: "Decision by end of Q1"

technique: "weighted_scoring_matrix"

alternatives: - id: A name: "Tool A (Enterprise)" description: "Enterprise-grade, feature-rich work item tracking" - id: B name: "Tool B (Collaborative)" description: "User-friendly, good collaboration features" - id: C name: "Tool C (Developer-Focused)" description: "Modern, developer-focused interface"

criteria: - id: C1 name: "Ease of Use" weight: 0.30 rationale: "Team adoption is critical" - id: C2 name: "Feature Set" weight: 0.25 rationale: "Must handle complex workflows" - id: C3 name: "Integration" weight: 0.25 rationale: "Slack integration required" - id: C4 name: "Cost" weight: 0.20 rationale: "Within budget constraint"

scores: - alternative: A scores: {C1: 3, C2: 5, C3: 4, C4: 3} weighted_total: 3.75 - alternative: B scores: {C1: 5, C2: 4, C3: 5, C4: 4} weighted_total: 4.50 - alternative: C scores: {C1: 4, C2: 4, C3: 3, C4: 5} weighted_total: 3.95

ranking: - rank: 1 alternative: B score: 4.50 - rank: 2 alternative: C score: 3.95 - rank: 3 alternative: A score: 3.75

sensitivity: - scenario: "Cost weight +10%" winner: C stable: false - scenario: "Ease of Use weight +10%" winner: B stable: true

recommendation: choice: B confidence: high rationale: - "Highest weighted score (4.50)" - "Stable across sensitivity scenarios" - "Best ease of use for team adoption"

risks: - description: "Asana pricing may increase" likelihood: medium impact: low mitigation: "Negotiate annual contract"

### Mermaid Decision Matrix Visualization

quadrantChart title Decision Matrix - Tool Selection x-axis Low Cost --> High Cost y-axis Low Features --> High Features quadrant-1 Premium quadrant-2 Best Value quadrant-3 Budget quadrant-4 Expensive Limited "Tool A (Enterprise)": [0.7, 0.9] "Tool B (Collaborative)": [0.5, 0.7] "Tool C (Developer)": [0.3, 0.6] "Tool D (Basic)": [0.2, 0.3]

## When to Use

| Scenario | Technique |
| --- | --- |
| Rule-based logic with many conditions | Decision Table |
| Comparing multiple options on criteria | Weighted Scoring Matrix |
| Sequential decisions with uncertainty | Decision Tree |
| Concept selection vs baseline | Pugh Matrix |
| Simple repeatable business rules | DMN-Lite |
| Quick relative comparison | Pugh Matrix |
| Need stakeholder buy-in | Weighted Scoring (transparent) |

## Integration

### Upstream

- **stakeholder-analysis** - Identify decision makers and criteria sources
- **root-cause-analysis** - Understand problem before deciding solution
- **swot-pestle-analysis** - Strategic context for decisions

### Downstream

- **Requirements** - Decision drives requirement priorities
- **Risk registers** - Capture decision risks
- **Implementation plans** - Execute chosen alternative

## Related Skills

- `prioritization` - MoSCoW, Kano for feature prioritization
- `risk-analysis` - Risk assessment for decision alternatives
- `root-cause-analysis` - Problem analysis before solution selection
- `business-model-canvas` - Strategic business decisions
- `stakeholder-analysis` - Decision maker identification

## Version History

- **v1.0.0** (2025-12-26): Initial release