Build a Decision Matrix for a Complex Choice

This prompt walks you through building a weighted decision matrix for a hard choice between three options — it proposes evaluation criteria, has you weight them, scores each option, computes weighted totals, and then sanity-checks whether the math-driven winner actually feels right. The final bias check is what separates it from a spreadsheet; it forces you to confront cases where the 'logical' answer conflicts with your gut. Use it for consequential, multi-factor decisions where you're stuck going in circles.

What you can expect back

Suggested criteria (weight these 1-5):
1. Income stability
2. Learning & growth
3. Autonomy
4. Long-term upside
5. Stress / work-life balance
6. Alignment with partner's risk comfort

Tell me your weights and I'll build the full matrix. As a starting illustration, using equal weights:

| Criterion (w) | Startup | Corporate | Independent |
| Income stability | 2 | 5 | 2 |
| Learning & growth | 5 | 2 | 4 |
| Autonomy | 3 | 2 | 5 |
| Long-term upside | 5 | 2 | 4 |
| Balance | 2 | 4 | 3 |
| Partner comfort | 2 | 5 | 2 |
Weighted totals: Startup 19 | Corporate 20 | Independent 20

Sanity check: Corporate and Independent tie, but you said you value learning and autonomy most — if you weight those higher, Independent likely pulls ahead. Possible hidden bias: you may be over-weighting income stability out of your partner's caution rather than your own. Send your real weights and we'll resolve it.

Illustrative example — your results will vary by tool and inputs.

Get sharper results

• 01Do the weighting step before you see any scores — assigning weights after you know which option is winning lets confirmation bias creep in.
• 02Add a 'regret' criterion ('which option would I most regret not trying?') since pure spreadsheets miss the emotional cost of the path not taken.
• 03When two options finish within a point of each other, treat it as a near-tie and let the bias check and your gut break it, not the rounding.
• 04Ask it to run a quick sensitivity test — 'if I raise the weight on autonomy by one, does the winner change?' — to see how fragile the result is.
• 05Be honest in your context about who else is affected; the model can only catch hidden biases you actually surface to it.

Adapt it for your case