Decision Matrix
How to Use
What this tool does
Several offers on the table, each good in a different way, and comparing them by feel tends to drop a criterion halfway through. A decision matrix puts options in columns and criteria in rows, asks for one score per cell and a weight per criterion, then ranks the options by weighted total. Every cell starts genuinely empty: the tool does not pre-fill fives and then present a suspicious three-way tie.
One worked example you can check by hand: three options, three criteria with weights 0.5, 0.3 and 0.2, and scores of 8, 6 and 7 gives 8×0.5 + 6×0.3 + 7×0.2 = 6.80. Change any cell or drag a weight slider and both the totals and the ordering recalculate, with the leader highlighted.
Options and criteria can be added or removed, and a CSV export carries the scores and weights into a spreadsheet for further tuning. If a single cell is left blank, the tool names it rather than quietly substituting zero and printing a tidy but meaningless ranking.
A matrix only helps if the inputs are honest, so the grid starts completely blank and the tool refuses to rank anything until every cell has a score. That rule prevents the most misleading failure mode, where an unscored cell is silently counted as zero and quietly sinks one option. Scores are expected to point the same way: a higher number means better on every criterion. Cost, commute time and similar measures therefore need inverting before entry, or the ranking will reward the most expensive option. Weights are raw numbers you set; they are not normalised behind your back, so if they do not add up to one the totals still reflect exactly the values on the sliders. Checking that sum is part of using the tool, not a hidden step. The weighted total for an option is the sum of score times weight across criteria, displayed to two decimals so it can be reproduced on paper. Changing one cell or one slider recalculates the entire table, which makes sensitivity testing straightforward: move a weight and see whether the leader changes, or drop a score and see how much ground an option loses. The CSV export carries option names, criteria, weights, scores and totals, so the same comparison can be continued in a spreadsheet with more rows. The output is a quantified version of the judgements you entered, not a recommendation that accounts for anything outside the table.
Key features
- Truly blank grid: Cells start empty with no pre-set score; no scores means no ranking is produced.
- Weight sliders: Each criterion has its own slider, and totals plus ordering update as it moves.
- Weighted totals: Scores are multiplied by weights and summed to two decimals, so the arithmetic can be checked manually.
- Live ranking: Options are ordered from highest to lowest total with the leader highlighted.
- Missing-cell check: An unfilled cell produces a prompt naming it instead of being counted as zero.
- CSV export: The grid, scores and weights export together for further analysis in a spreadsheet.
How to use it
- Add options: Type the alternatives you are comparing, such as two job offers, adding as many as needed.
- List criteria: Write the dimensions that matter, for example salary, commute and growth.
- Set weights and score: Drag each criterion weight, then fill every cell with that option's score.
- Press Calculate: The result area returns weighted totals and a ranking with the top option highlighted.
- Export or adjust: Export CSV for a record, or change a few scores to see how sensitive the ranking is.
Things to keep in mind
- Weights are yours: Weights express preference and are never pre-filled; if they do not sum to one the raw values are used, so decide deliberately.
- Keep scores directional: Every criterion is scored higher-is-better; invert cost-type measures before entering them.
- Blank is not zero: A missing cell raises a prompt rather than dragging an option down with a phantom zero.
- A ranking is not advice: The order quantifies the inputs only; factors outside the table still belong in the final choice.
- Nothing is uploaded: Names and scores live on the current page and disappear on refresh, so export anything worth keeping.