Build a weighted decision matrix tool that explains its answer

How a weighted decision matrix works

1. List the options. Three to six is manageable. Two options is a pros and cons list; more than eight and scoring gets careless.
2. List the criteria. What actually matters: salary, commute, growth and team for a job; price, range, boot space and running costs for a car.
3. Weight each criterion. On a 1 to 5 scale, or by sharing 100 points. Weights are the most honest part of the exercise: they force you to admit that commute matters more than the logo on the building.
4. Score each option against each criterion, say 1 to 10.
5. Multiply and add. Each score times its weight, summed per option. Highest total wins.

B and C are nearly tied, which is the most useful thing the matrix can tell you: the decision is close, and your choice should rest on something the criteria didn't capture, or on firmer scores for growth and commute.

Explain the result, don't just announce it

A total alone hides why an option won. A good decision making tool adds three things:

Contribution bars

For each option, a stacked bar showing how much each criterion added. You see at a glance that C wins on growth and B on commute.

A sentence

"C leads B by 2 points. C's advantage is growth (+12) and salary (+5); B's is commute (+12) and team (+3)." Written by the app's own logic from the numbers, not by AI.

Sensitivity

"If commute were weighted 5 instead of 4, B would win." Show the smallest change to one weight or score that would flip the result. If it's tiny, the matrix is a coin toss.

When the result feels wrong

That reaction is data. Often a criterion is missing ("I'd hate living in that city") or a weight is dishonest. Add a "gut feeling" criterion if you must, but say so; the point is to make the reasoning visible, not to outsource the decision.

Describing it to Anticode

"A decision matrix tool. Add options and criteria freely. Weights 1 to 5 per criterion, scores 1 to 10 per cell, entered on a grid with keyboard navigation. Results: weighted totals, a ranked list, stacked contribution bars, a plain-English explanation of why the leader leads, and a sensitivity line showing the smallest change that would flip first place. Notes per cell for the reasoning behind a score. Several saved decisions in the browser. Calm, neutral, readable; no gamified confetti."

Anticode's model writes it as one app that runs in the browser. Test it with a case you can check by hand, like the table above, and try the awkward cases: a tie, a criterion with weight zero, an option with blank scores.

Saving, sharing and limits

Decisions are saved in that browser on that device. Good for personal choices; less so for group ones.

• No shared scoring. A hiring panel or a committee can't each score on their own devices and combine results; there's no network or shared data. One person can enter everyone's scores on one screen in a meeting, or ask for an average-of-scorers mode filled in on one device.
• No outside data. It can't fetch prices, reviews or specifications for options; you enter what you know.
• No AI advice. The explanation is arithmetic, which is the point.
• Share by export. Ask for a printable summary or an export file someone else can import into their own copy.

If the decision is about money over time, a break-even calculator or a rent vs buy calculator may answer it more directly. For a side-by-side of products for your customers, see the product comparison tool.

Questions people ask

How many criteria should a decision matrix tool have?

Four to eight is usual. Beyond that, small criteria add noise and the weights of important ones get diluted.

Should I use a pros and cons list instead?

For two options with a few factors, a pros and cons list is fine. Weights help once options differ in many ways.

Is my decision private?

It stays in your browser on your device and isn't sent anywhere.