Multi-Criteria Decision Analysis

Overview

Systematic procedure for making decisions involving multiple criteria, tradeoffs, and stakeholders

Steps

Step 1: Frame the decision

Define the decision clearly and establish scope:

1. Write a clear decision statement
2. Identify decision type (binary, selection, prioritization, allocation)
3. Determine timeline and urgency
4. List key stakeholders and their interests
5. Document known constraints and non-negotiables

Step 2: Generate and screen options

Identify all viable alternatives:

1. Brainstorm all possible options without filtering
2. Include “do nothing” / status quo as baseline
3. Screen against hard constraints (disqualify non-viable)
4. Aim for 3-7 meaningfully different options
5. Write brief description of each option

Step 3: Define evaluation criteria

Identify and define the factors for evaluation:

1. List all factors that matter for this decision
2. Group similar factors to avoid redundancy
3. Ensure completeness (financial, time, quality, risk, strategic)
4. Write clear definition for each criterion
5. Define scoring scale for each (what does 1 vs 5 mean?)

Step 4: Assign weights to criteria

Determine relative importance of each criterion:

1. Choose weighting method (direct assignment, pairwise, swing)
2. Assign weights BEFORE seeing any scores
3. Validate weights sum to 100%
4. Get stakeholder alignment on weights
5. Document rationale for weight choices

Step 5: Score options on criteria

Rate each option on each criterion:

1. Gather data and evidence for each option-criterion pair
2. Apply scoring rubric consistently
3. Score all options on one criterion before moving to next
4. Document evidence/rationale for each score
5. Get multiple perspectives where possible

Step 6: Calculate weighted totals

Compute overall scores and rank options:

1. For each option: sum(weight x score) across all criteria
2. Rank options by weighted total
3. Note the margin between top options
4. Identify which criteria most differentiate options

Step 7: Perform sensitivity analysis

Test robustness of the recommendation:

1. Vary each weight by +/-20% and recalculate
2. Find threshold weights where winner changes
3. Assess if threshold weights are plausible
4. Test optimistic/pessimistic score scenarios
5. Document how stable the recommendation is

When to Use

• Comparing multiple options across several evaluation dimensions
• Making decisions where stakeholders have different priorities
• Need to justify a decision to others with clear rationale
• Tradeoffs between criteria are unclear or contested
• Want to reduce bias from first impressions or anchoring
• Decision has significant impact and warrants structured analysis
• Multiple valid options exist and intuition is insufficient
• Need to build consensus among team members with differing views

Verification

• Options are mutually exclusive and collectively exhaustive
• Criteria are non-redundant (no double-counting)
• Weights were set before scoring (no gaming)
• Each score has documented evidence/rationale
• Sensitivity analysis performed on key weights
• Intangible factors explicitly considered
• Dissenting views captured and addressed
• Decision rationale documented for future reference

Input: $ARGUMENTS

Apply this procedure to the input provided.