How do I build a weighted comparison table for a big decision?

A weighted table is not a machine that decides for you. It is a device for making your own preferences visible, so that a decision stops being an argument between two vague impressions and becomes an argument about two specific numbers. Most of its value arrives before the total is calculated, at the moment you have to say out loud how much each thing really matters.

Step by step

1. Write the criteria down the left-hand side.Use the must-haves you already screened on as filters, not rows — anything on the table should already have passed them. The rows are the things you are genuinely trading off.
2. Weight each criterion from 1 to 5 before you look at any option.Weighting after you have seen the options is how a table gets reverse-engineered to produce the answer you already wanted. Do the weights first, in a separate sitting if you can.
3. Put the options across the top, three to five of them.More than five and the scoring becomes guesswork, because you cannot hold that many things in mind consistently enough to score them on the same scale.
4. Score each option 1 to 5 on each criterion.Score against a fixed meaning, not against the other options: 1 is poor, 3 is adequate for your purpose, 5 is as good as this realistically gets. Anchoring to the other options is what makes the table drift.
5. Multiply, add, and read the totals.Weight times score for each cell, sum each column. Write the totals down. A gap of less than about 10 per cent between the top two is not a real difference given how rough the scoring is.
6. Run the sanity check.Look at the winner and ask whether you are disappointed. If you are, do not override the table silently — go back and find the weight that is wrong, change it, and see what happens. That is usually a criterion you care about more than you admitted.
7. Test the result against a change of weights.Push the one or two weights you were least sure about up and down by a point. If the winner does not change, you have a robust answer. If it flips, the decision genuinely hinges on that one judgement, and that is where the remaining thought should go.

Tips

• Do the weights and the scores in different sittings, or with different people, so the two are not contaminated by each other.
• Add a row for total cost of ownership rather than purchase price, and weight it honestly. Price is almost never a 5 and almost never a 1.
• If you and a partner are deciding, each fill in your own weights and compare the two weight columns first. The disagreement is nearly always about weights, not scores.
• A criterion everything scores the same on is doing no work. Delete the row and the table gets clearer.

Common mistakes

• Adding a criterion to justify the option you like — It converts the table from a tool into a decoration. If a new criterion appears mid-scoring, it either belonged there from the start or it does not belong at all.
• Using a 1 to 10 scale — People cannot reliably distinguish ten levels of anything. Five gives you enough resolution and forces genuine judgement rather than false precision.
• Treating a two-point total gap as decisive — The scoring is rough by construction. Close totals mean the options are close, which is a useful answer in itself.

Questions people ask