Complete firstDecision frame
Options, decision owner, time horizon, and minimum requirements are stated.
A changing scope makes the matrix compare different questions.
Method guideDecision Making
Decision Matrix
A practical method guide: clarify prerequisites and materials, understand the complete process, and look up outputs, pitfalls, and stop criteria.
ComplexityMedium
Time45-90 min
Participants2-8
FormatWorkshop + async
MaturityEstablished
01
Prerequisite
What needs to be finished first
Complete firstRateable criteria
Criteria distinguish options and can be applied consistently across all of them.
Vague or overlapping criteria create false precision.
02
Preparation
What needs to be ready before start
Materials
Decision question, real alternatives, criteria and weights, defined rating scale, evidence sources, and a blank comparison table.
People / roles
The decision owner approves criteria and weights. Subject experts rate evidence and uncertainty. Facilitation records divergent assessments.
Pre-read
Describe options, clarify hard requirements, gather evidence for relevant criteria, and involve the decision owner.
Time needed
45–90 minutes for a prepared comparison. Missing market, cost, or technical research takes additional work.
Setup
Write criteria so they are distinct and decision-relevant. Normalize weights to 100 percent total. Define a shared scale with observable anchors. For a weighted sum, total score = sum of weight × rating; ratings must share a scale. Low costs or risks need a scale direction that reflects desirability. Keep evidence, assumptions, and uncertainty visible in each cell.
03
Core question
The one question this method answers
Which option best fits the decision against the same evidenced criteria and disclosed uncertainty?
04
Flow
Phases with goal, actions, output and pitfall
A Decision Matrix places alternatives beside decision-relevant criteria. In a weighted version, criterion weights and ratings use a shared, defined scale. The sum makes assumptions traceable; it is not objective truth. Uncertainty and close scores must shape the recommendation.
Weighted comparison of two services
Fictional teaching example
Fictional ratings on an anchored 1–5 scale with a narrow result shown openly.
| Criterion | Criterion | Weight | Service A | Service B | Fictional evidence |
|---|---|---|---|---|---|
| Criterion 1Rating 1–5 × weight; higher means better fit. | Understandable to users | 30% | 4 × 30% = 1.20 | 3 × 30% = 0.90 | Scripted comprehension check |
| Criterion 2Rating 1–5 × weight; higher means better fit. | Access protection | 35% | 3 × 35% = 1.05 | 5 × 35% = 1.75 | Check against pilot requirements |
| Criterion 3Rating 1–5 × weight; higher means better fit. | Workflow fit | 25% | 5 × 25% = 1.25 | 4 × 25% = 1.00 | Fictional workflow prototype |
| Criterion 4The scale is anchored so lower implementation effort receives a higher rating. | Implementation effort | 10% | 4 × 10% = 0.40 | 3 × 10% = 0.30 | Fictional team estimate |
| Weighted totalSum weight × rating; maximum total is 5. | Total score | 100% | 3.90 | 3.95 | Service B leads by 0.05; check robustness. |
1. Criterion
- Criterion 1
- Understandable to users
- Criterion 2
- Access protection
- Criterion 3
- Workflow fit
- Criterion 4
- Implementation effort
- Weighted total
- Total score
2. Weight
- Criterion 1
- 30%
- Criterion 2
- 35%
- Criterion 3
- 25%
- Criterion 4
- 10%
- Weighted total
- 100%
3. Service A
- Criterion 1
- 4 × 30% = 1.20
- Criterion 2
- 3 × 35% = 1.05
- Criterion 3
- 5 × 25% = 1.25
- Criterion 4
- 4 × 10% = 0.40
- Weighted total
- 3.90
4. Service B
- Criterion 1
- 3 × 30% = 0.90
- Criterion 2
- 5 × 35% = 1.75
- Criterion 3
- 4 × 25% = 1.00
- Criterion 4
- 3 × 10% = 0.30
- Weighted total
- 3.95
5. Fictional evidence
- Criterion 1
- Scripted comprehension check
- Criterion 2
- Check against pilot requirements
- Criterion 3
- Fictional workflow prototype
- Criterion 4
- Fictional team estimate
- Weighted total
- Service B leads by 0.05; check robustness.
Formula: weighted total = sum of weight (as a decimal) × rating. Weights total 1 or 100%. Ratings always increase with suitability. A plausible change from 30/35/25/10 to 35/30/25/10 reverses the ranking: A 3.95, B 3.85. Check weights and minimum requirements before recommending.
- 1
Decision
Set the decision frame
Goal
Everyone evaluates the same decision.
Inputs for this step
- Decision need and context
Actions
- Confirm options and decision authority
- mark hard eligibility gates.
Output
Bounded decision
Scoring starts after must-haves are clarified.
Quality check
The matrix is expected to invent options that do not exist yet.
Pitfall
The decision is not yet sufficiently bounded. Develop options first and check minimum requirements.
Timing: 8–10 minutes
- 2
Criteria
Separate and anchor criteria
Goal
Each criterion can be rated consistently across options.
Inputs for this step
- Requirements and subject knowledge
Actions
- Remove duplicates
- set weights totaling 100%
- give scale points observable anchors.
Output
Criteria and rating rule
Weights reflect importance in this decision, not universal truth.
Quality check
Criteria are vague, duplicated, or reverse-engineered for a favored option.
Pitfall
Ratings rely on vague or overlapping criteria. Rewrite criteria with the decision owner.
Timing: 12–15 minutes
- 3
Evidence
Rate options with evidence
Goal
Every matrix cell is traceable.
Inputs for this step
- Option descriptions and criteria
Actions
- Record sources per cell
- combine different units only with clear normalization
- mark uncertainty.
Output
Rated matrix
A number without evidence and scale anchors is only an opinion with a decimal place.
Quality check
Ratings use different time periods or definitions.
Pitfall
Ratings lack comparable evidence or scales. Align the rating method and obtain missing data.
Timing: 15–25 minutes
- 4
Ranking
Check totals and margins
Goal
The ranking is arithmetically correct and meaningful.
Inputs for this step
- Complete matrix and agreed weights
Actions
- Sum weight times rating
- explain the result and units
- mark close options.
Output
Provisional ranking
A total score cannot compensate for a critical safety or compliance risk.
Quality check
The highest total is declared the only correct choice.
Pitfall
A narrow total is presented as a certain choice. Show minimum requirements and critical disadvantages next to the total.
Timing: 8–12 minutes
- 5
Robustness
Record sensitivity and decision
Goal
The decision owner understands how robust the matrix is.
Inputs for this step
- Ranking, assumptions, and uncertainty
Actions
- Find values that reverse the ranking
- show close options as close
- document rationale for recommendation and decision.
Output
Reasoned decision brief
A fragile ranking calls for more investigation or deliberate risk acceptance.
Quality check
A narrow lead is presented as robust.
Pitfall
Uncertain values are not checked for their effect. Report sensitivity and reduce the most relevant evidence gap.
Timing: 10–15 minutes
05
Artifact
What comes out at the end
Sketch: what the artifact looks like
Form
Versioned Decision Matrix with question, options, eligibility requirements, weighted criteria, scale anchors, evidence, sensitivity, recommendation, and final decision.
Versioning / ownership
Save date, scope, criteria, weights, scale, and sources. Record changes and recalculate the ranking. Keep the recommendation separate from the formal decision.
Tool alternatives
- Spreadsheet with explicit scale anchors and weight total
- Workshop board to agree criteria before ratings
- Prototype, offer data, or technical investigation for uncertain cells
trade-off-matrix-markdown.md
Markdown matrix for comparing options across competing criteria and accepted trade-offs.
Trade-off Matrix: Decision
| Criterion | Weight | Option A | Option B | Option C |
|---|---|---|---|---|
| Benefit | 30 | |||
| Cost | 20 | |||
| Risk | 20 | |||
| Time-to-Learn | 15 | |||
| Reversibility | 15 |
Accepted trade-offs
- Option A accepts:
- Option B accepts:
- Option C accepts:
Recommendation
Which option is recommended and why?
Sensitivity
Which small change in weighting or assumption would flip the recommendation?
decision-matrix-working-template.md
Plan a bounded decision with comparable options, a shared rating scale, and evidenced criteria.
Decision Matrix: Worksheet
Question
To be confirmed
Desired outcome
To be confirmed
Scope
Run a criteria comparison Time: 45–90 minutes for joint scoring; evidence gathering is additional
Complete this worksheet on paper or in your own document during the work.
Preparation for this scope
Method setup
- Materials: Decision question, real alternatives, criteria and weights, defined rating scale, evidence sources, and a blank comparison table.
- Roles: The decision owner approves criteria and weights. Subject experts rate evidence and uncertainty. Facilitation records divergent assessments.
- Advance information: Describe options, clarify hard requirements, gather evidence for relevant criteria, and involve the decision owner.
- Overall time needed: 45–90 minutes for a prepared comparison. Missing market, cost, or technical research takes additional work.
- Setup: Write criteria so they are distinct and decision-relevant. Normalize weights to 100 percent total. Define a shared scale with observable anchors. For a weighted sum, total score = sum of weight × rating; ratings must share a scale. Low costs or risks need a scale direction that reflects desirability. Keep evidence, assumptions, and uncertainty visible in each cell.
Prepare for the work steps
Decision: Set the decision frame
- Decision need and context
Criteria: Separate and anchor criteria
- Requirements and subject knowledge
Evidence: Rate options with evidence
- Option descriptions and criteria
Ranking: Check totals and margins
- Complete matrix and agreed weights
Robustness: Record sensitivity and decision
- Ranking, assumptions, and uncertainty
Weighted comparison of two services
| Criterion | Criterion | Weight in percent | Option A: rating and weighted value | Option B: rating and weighted value | Evidence, source, or uncertainty |
|---|---|---|---|---|---|
| Criterion 1 | |||||
| Criterion 2 | |||||
| Criterion 3 | |||||
| Criterion 4 | |||||
| Weighted total |
- Criterion 1: Rating 1–5 × weight; higher means better fit.
- Criterion 2: Rating 1–5 × weight; higher means better fit.
- Criterion 3: Rating 1–5 × weight; higher means better fit.
- Criterion 4: The scale is anchored so lower implementation effort receives a higher rating.
- Weighted total: Sum weight × rating; maximum total is 5.
Blank worksheet: record the decision, minimum requirements, rating scale, and weight total. Enter evidence and ratings yourself. Example values are not carried over.
NASA Systems Engineering Handbook: Decision Analysis
Decision: Set the decision frame
Expected artifact: Bounded decision
Scoring starts after must-haves are clarified.
Entry:
...
- The matrix is expected to invent options that do not exist yet.
Criteria: Separate and anchor criteria
Expected artifact: Criteria and rating rule
Weights reflect importance in this decision, not universal truth.
Entry:
...
- Criteria are vague, duplicated, or reverse-engineered for a favored option.
Evidence: Rate options with evidence
Expected artifact: Rated matrix
A number without evidence and scale anchors is only an opinion with a decimal place.
Entry:
...
- Ratings use different time periods or definitions.
Ranking: Check totals and margins
Expected artifact: Provisional ranking
A total score cannot compensate for a critical safety or compliance risk.
Entry:
...
- The highest total is declared the only correct choice.
Robustness: Record sensitivity and decision
Expected artifact: Reasoned decision brief
A fragile ranking calls for more investigation or deliberate risk acceptance.
Entry:
...
- A narrow lead is presented as robust.
Open questions and next steps
...
Method guide: https://methodatlas.meierhoff-systems.de/en/methods/decision-matrix/run-sheet
06
Example output
Concrete filled scenario, fictional example
Fictional example for orientation. Counts and timings describe this case and are not universal requirements.
decision-matrix-beispiel.md
Concrete filled scenario, fictional example
Fictional teaching example: choosing a notification service
All options, ratings, and test results are invented. A small product team compares two services on the same 1–5 scale. A rating of 1 clearly misses the requirement, 3 meets the minimum need, and 5 is a strong fit. Higher ratings mean greater suitability for every criterion. Weights total 100 percent.
| Criterion | Weight | Service A | Service B | Fictional evidence |
|---|---|---|---|---|
| Understandable to users | 30% | 4 × 30% = 1.20 | 3 × 30% = 0.90 | Scripted comprehension check |
| Access protection | 35% | 3 × 35% = 1.05 | 5 × 35% = 1.75 | Check against pilot requirements |
| Workflow fit | 25% | 5 × 25% = 1.25 | 4 × 25% = 1.00 | Fictional workflow prototype |
| Implementation effort | 10% | 4 × 10% = 0.40 | 3 × 10% = 0.30 | Team effort estimate |
| Weighted total on a scale up to 5 | 100% | 3.90 | 3.95 | Service B leads by 0.05. |
The small difference does not prove a winner. A sensitivity check shifts five percentage points from access protection to comprehension, changing the weights to 35/30/25/10. Service A rises to 3.95 and Service B falls to 3.85, reversing the ranking. The team checks whether access protection is secured as a hard minimum and whether comprehension should truly receive more weight. The decision owner records the selection, caveats, and any evidence still needed.
07
Pitfalls
Recognize symptoms and steer against them
Quick overview, the full text is in the matching phase in the flow
- Set the decision frameThe decision is not yet sufficiently bounded.View in phase
- Separate and anchor criteriaRatings rely on vague or overlapping criteria.View in phase
- Rate options with evidenceRatings lack comparable evidence or scales.View in phase
- Check totals and marginsA narrow total is presented as a certain choice.View in phase
- Record sensitivity and decisionUncertain values are not checked for their effect.View in phase
Trap
Set the decision frame
Symptom
The decision is not yet sufficiently bounded.
What to do
Develop options first and check minimum requirements.
Trap
Separate and anchor criteria
Symptom
Ratings rely on vague or overlapping criteria.
What to do
Rewrite criteria with the decision owner.
Trap
Rate options with evidence
Symptom
Ratings lack comparable evidence or scales.
What to do
Align the rating method and obtain missing data.
Trap
Check totals and margins
Symptom
A narrow total is presented as a certain choice.
What to do
Show minimum requirements and critical disadvantages next to the total.
Trap
Record sensitivity and decision
Symptom
Uncertain values are not checked for their effect.
What to do
Report sensitivity and reduce the most relevant evidence gap.
08
Stop criteria
When to pause, resolve prerequisites, or choose another approach
Options or minimum requirements are unclear: stabilize the decision frame before scoring.
Criteria measure the same effect more than once or omit important perspectives: revise the criteria.
Evidence is too uncertain to support the ranking: research or test instead of adding decimal places.
The ranking changes with small plausible adjustments: show options as close and reduce uncertainty deliberately.
A decision owner wants the matrix to decide automatically: take explicit responsibility for the decision and rationale.
Finished the method guide?
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.