Do not start with the spreadsheet
Write the decision question, accountable owner, deadline and realistic alternatives first. Identify hard constraints before ranking preferences. A model cannot make an infeasible option acceptable.
Model selector
| Decision pattern | Starting model | Main question |
|---|---|---|
| Several competing criteria | Weighted matrix | Which option best balances our priorities? |
| Full life-cycle cost | TCO | What will each option cost over the same period? |
| Uncertain outcomes | Expected value | What is the probability-weighted outcome? |
| Choices in sequence | Decision tree | What should happen after each event? |
| Different possible futures | Scenario analysis | How does each option perform under coherent futures? |
| Uncertain assumptions | Sensitivity analysis | Which assumptions can reverse the choice? |
| Volume needed for viability | Break-even | How much must be sold or produced? |
| Ownership or sourcing pattern | Buy vs build / lease vs own | Which capability and commitment model fits? |
A five-question screen
- Can the major outcomes be expressed in one common unit?
- Do several objectives conflict?
- Can uncertain outcomes and probabilities be estimated?
- Do choices occur in sequence after new information arrives?
- Could a small change in one assumption reverse the result?
Use more than one lens when needed
A purchase decision may begin with TCO, use a weighted matrix for quality and support, and apply sensitivity analysis to useful life and downtime. Combining models is sensible when each answers a different question.
Finish with a decision record
Record the chosen option, rationale, accepted tradeoffs, major assumptions and review triggers. This converts analysis into an accountable decision.