Problem Solving & Quality · Solutions

Weighted Decision Matrix

Agree criteria and weights first, then score each option against each criterion: the choice becomes a visible calculation instead of a debate of ranks.

Time45 min
FormatTeam
StageSolutions

Weighted Decision Matrix: what it is and why it works

A weighted decision matrix compares options against a set of agreed criteria, each with an importance weight. Options sit in rows, criteria in columns. Each option is scored against each criterion, typically from 1 to 5, the score is multiplied by the criterion weight, and the products are summed into a total. Before any scoring, options that fail a must-have constraint, such as a safety requirement, a legal limit or a hard budget ceiling, are eliminated. What remains is a ranking whose logic anyone can inspect and challenge.

The method works because it separates two discussions that usually get tangled: what matters, and how well each option delivers it. Agreeing weights first reduces the temptation to tune criteria to favor a preferred solution. Scoring one criterion at a time across all options, with a written reason for each score, limits the halo effect. A sensitivity check then shows whether the ranking is robust or hinges on one debatable weight. Compared with a Pugh matrix, which compares concepts relative to a reference and aims to improve them, the weighted matrix is better for a final choice between well-defined options. A risk-benefit grid can then sequence the chosen actions.

What you need

A shortlist of well-defined options, typically three to eight
Must-have constraints that any option has to satisfy
Candidate criteria covering effectiveness, cost, time, safety and maintainability
Data or expert estimates to support scores, such as quotes, lead times and test results
The people who will implement and maintain the chosen solution

What you get

Criteria with agreed weights summing to 100
A scored matrix with a one-line reason for every score
A ranked list of options, with any eliminated ones and why
A sensitivity note showing which weights could change the decision

When to use it

When options are compared by the seniority of whoever defends them.

How to do it, step by step

List the shortlisted options in rows.
Agree the criteria in columns — effectiveness on the cause, cost, lead time, safety, ease of maintenance — before looking at scores.
Give each criterion a weight that sums to 100 and eliminate any option that fails a must-have constraint.
Score each option from 1 to 5 on each criterion, with a one-line justification per score.
Multiply, add up, and test sensitivity: if a small change of weight reverses the ranking, discuss that criterion openly.

Worked example: Choosing a fix for repeat seal failures on cooling water pumps

Illustrative scenario — figures are realistic but not from a real company.

A power plant's auxiliary cooling water system has four horizontal pumps. Mechanical seal failures averaged one every six weeks, each costing around $6,500 in parts and labor. Root cause analysis showed that suspended solids in the water were damaging the seal faces. The team shortlisted five options.

Options: cartridge seal with harder faces, external clean flush, a cyclone separator on a flush line, packing with a lantern ring, and upstream strainer upgrade. The must-have constraint of no visible leakage onto the floor eliminated packing.
Criteria and weights: effectiveness on the cause 35, lifecycle cost 25, lead time 15, maintainability 15, operator safety 10. Weights were fixed before anyone scored.
Scoring was done column by column with a reason per score. The external flush scored 5 on effectiveness but 2 on cost, because it required a new supply line. The cyclone separator scored 4 and 4.
Totals: cyclone separator 400, external flush 355, hard-faced cartridge seal 320, strainer upgrade 290. Shifting 10 points from cost to effectiveness still left the cyclone first, so the ranking was considered robust.

Result. Cyclone separators were fitted to all four pumps over two months. In the following year, two seal failures occurred instead of the eight or nine expected, both traced to a start-up procedure issue. The team kept the matrix with the decision record, which saved time when the same question came up for the fire water pumps.

Open it in the FieldKit Problem Solving app → Get the Problem Solving template pack — join the waitlist

Common pitfalls and how to avoid them

Setting weights after seeing the scores, consciously or not, to favor a preferred option.Agree and freeze weights before scoring begins, and record them.
Using criteria that overlap, such as "cost" and "payback", so one factor is counted twice.Check that each criterion measures something distinct; merge or remove duplicates.
Scoring without justification, so the matrix looks precise but reflects opinion.Write a one-line reason or data point for every score, and flag scores based on guesses.
Treating a narrow win as decisive, such as 312 against 305.Run a sensitivity check; if the ranking flips easily, discuss the critical criterion openly or gather more data.

Frequently asked questions

How do you create a weighted decision matrix?

List the options in rows and the criteria in columns. Remove options that fail any must-have constraint. Assign each criterion a weight so that the weights sum to 100. Score each option from 1 to 5 on each criterion, working one criterion at a time. Multiply each score by its weight, add the results for each option and compare totals, then test how sensitive the ranking is to the weights.

What is the difference between a weighted matrix and a Pugh matrix?

A Pugh matrix compares each concept against a reference as better, same or worse, without weights, and is used early to improve and combine concepts. A weighted decision matrix scores well-defined options against weighted criteria to select one. A common sequence is to use Pugh to shape and merge concepts, then a weighted matrix to make the final, documented choice.

How many criteria should a decision matrix have?

Usually four to eight. Fewer than four often misses an important dimension such as maintainability or safety. More than eight dilutes the weights, so that each criterion carries little influence and the totals converge. If you have many criteria, group related ones under a heading or turn some into must-have constraints instead of scored criteria.

Origin

Weighted scoring — multi-criteria decision practice; cf. Kepner–Tregoe decision analysis (musts and wants), 1965.

Used in these playbooks

One-page improvement cycle 3 weeks

Three weeks for a local problem owned by a team leader: see it on site, tell it on one A3, choose the countermeasure openly, pilot it and make it the new standard.

Related methods

More in “Solutions”

Generate countermeasures that act on the cause and choose between them openly.