qa-aman/decision-matrix

Overview

Based on Software Requirements by Karl Wiegers & Joy Beatty - the Wiegers Priority Matrix that scores options across 4 factors (relative benefit, relative penalty, relative cost, relative risk) to produce a data-driven ranking. Also draws on Business Analysis Techniques by James Cadle for Decision Tables, Decision Trees, and Force-Field Analysis, and Lean Business Analysis by Mark Sherrington for Real Options thinking - deferring irreversible decisions until the last responsible moment. The key insight from Wiegers: gut-feel prioritization always favors the loudest stakeholder. A structured matrix makes trade-offs visible and defensible.

Workflow

Step 1: Define the decision and options

DECISION: [What are we choosing between?]
CONTEXT: [Why this decision matters now]
OPTIONS:
  A: [option name and brief description]
  B: [option name and brief description]
  C: [option name and brief description]
DECISION MAKER: [who has final authority]
DEADLINE: [when must this be decided]

Step 2: Define evaluation criteria

List criteria that matter for this decision. Common categories:

• Value criteria: business benefit, user impact, strategic alignment, revenue potential
• Cost criteria: development effort, operational cost, opportunity cost
• Risk criteria: technical risk, adoption risk, compliance risk, reversibility

Step 3: Assign weights to criteria

Each criterion gets a weight reflecting its importance (must total 100%):

| Criterion | Weight | Rationale |
|-----------|--------|-----------|
| Business impact | 30% | Primary driver per sponsor |
| Implementation cost | 25% | Budget-constrained project |
| Time to deliver | 20% | Regulatory deadline in Q3 |
| Technical risk | 15% | New technology stack |
| Scalability | 10% | Growth expected in Year 2 |

Get stakeholder agreement on weights before scoring. The weights reflect organizational priorities.

Step 4: Score each option (Wiegers Priority Matrix)

Rate each option against each criterion on a consistent scale (1-5 or 1-10):

| Criterion (Weight) | Option A | Option B | Option C |
|---------------------|----------|----------|----------|
| Business impact (30%) | 8 | 6 | 9 |
| Implementation cost (25%) | 5 | 8 | 3 |
| Time to deliver (20%) | 7 | 9 | 4 |
| Technical risk (15%) | 6 | 7 | 5 |
| Scalability (10%) | 4 | 5 | 9 |

For the Wiegers 4-factor variant, score each option on:

• Relative Benefit (1-9): Value of including this option
• Relative Penalty (1-9): Cost of NOT including it
• Relative Cost (1-9): Effort to implement
• Relative Risk (1-9): Technical/delivery uncertainty

Priority = (Benefit% + Penalty%) / (Cost% + Risk%)

Step 5: Calculate weighted scores

For each option: Sum of (score x weight) across all criteria.

Option A: (8x0.30) + (5x0.25) + (7x0.20) + (6x0.15) + (4x0.10) = 6.45
Option B: (6x0.30) + (8x0.25) + (9x0.20) + (7x0.15) + (5x0.10) = 7.15
Option C: (9x0.30) + (3x0.25) + (4x0.20) + (5x0.15) + (9x0.10) = 5.90

Step 6: Apply Force-Field Analysis for close calls

When scores are close (within 10%), use Cadle's Force-Field Analysis:

DRIVING FORCES (for change)    |    RESTRAINING FORCES (against)
=============================== | ===============================
[force] ------>  strength: 4    | strength: 3  <------ [force]
[force] ------>  strength: 5    | strength: 2  <------ [force]

The option with stronger net driving forces is preferred.

Step 7: Document the recommendation

RECOMMENDATION: [Option X]
SCORE: [weighted score]
KEY TRADE-OFFS: [what you're giving up by choosing this option]
REVERSIBILITY: [can this decision be changed later? at what cost?]
DISSENTING VIEWS: [any stakeholder disagreements to note]

Anti-Patterns

1. Choosing criteria after seeing the scores Bad: Adding or removing criteria to justify a preferred option. Good: Define and weight criteria before scoring. Lock them with stakeholder agreement.

2. Equal weights on everything Bad: All criteria weighted 20% each - hides the real priorities. Good: Force-rank criteria. If everything is equally important, nothing is important.

3. Scoring without evidence Bad: "I feel like Option A is an 8 on scalability." Good: "Option A uses a horizontally scalable architecture tested to 10K concurrent users - score 8."

4. Ignoring reversibility (Sherrington's Real Options) Bad: Treating all decisions as permanent. Good: If a decision is easily reversible, it may not need a full matrix. Invest analysis time proportional to the cost of being wrong.

5. Matrix replaces judgment Bad: "The matrix says Option B, so we go with B." Good: The matrix informs the decision. If the result feels wrong, the weights or criteria may be wrong - revisit them.

Quality Checklist

• [ ] Decision and options clearly defined
• [ ] Criteria cover value, cost, and risk dimensions
• [ ] Weights assigned and agreed by stakeholders before scoring
• [ ] Weights total 100%
• [ ] Scores supported by evidence (not gut feel)
• [ ] Weighted scores calculated correctly
• [ ] Close results analyzed with Force-Field Analysis or sensitivity check
• [ ] Trade-offs of the recommended option are explicit
• [ ] Reversibility of the decision is assessed