Decision Matrix · Weighted Scoring Method
Decision Matrix / Weighted Scoring
What is a decision matrix?
- A scoring table: criteria × weights × scores
- Fix the weights before looking at the options
- A documented rationale
When a meeting weighs several alternatives, the argument usually starts with “which is better?” and ends with “who argues louder.” The moment you bring out a matrix, the topic changes. Instead of fighting over the conclusion, people first debate what to judge by, and once the criteria are settled, a calculator does the summing. In practice, the hottest 30 minutes of such a meeting come not while filling in the scorecard but while deciding whether price gets a 40% or a 20% weight. That is why the order matters: lock the weights and write them into the minutes before the list of alternatives goes up on screen.
There are three main forms. The weighted-score type rates each criterion 1–5 and sums the weighted scores. The Pugh method sets the current option at 0 and compares alternatives using only +, 0 and −. The AHP type derives weights by comparing criteria in pairs. If you have three or four criteria and need a quick answer, Pugh fits; if you must submit your rationale to outsiders, the weighted-score type fits. The recommended practical range is 5–7 criteria, with each weight at least 5%. Beyond ten criteria, points get sliced so thin that total scores stop differing, and items under 5% cannot change the ranking yet still eat up scoring time. Also, do not put must-have conditions in the weighted items. Conditions where falling short means game over, such as security authentication or delivery deadlines, are handled up front as pass/fail gates rather than scores.
The most common mistake is working the numbers backward. When your gut-feel first choice comes out second, you say “this criterion’s weight was a bit low” and shift 5%, and at that moment the table stops being a decision tool and becomes paperwork that justifies a conclusion already reached. It is more honest to write out why you want to flip the result and add it as a new criterion. Overtrusting the total-score gap is also risky. On a 100-point scale, treat a gap within 3 points as a tie, and reread the rationale starting with any cell where raters differ by 2 points or more. Because a 3 means different things to different people, unless you add a one-line description for each scale point, the total ends up as an average of personal tastes.
John KimInstructor John Kim, author of B2B Marketing Design and CEO of Nanum Consulting
Ara LimSpecialist in global leadership, AI, and business communicationA basic technique in the multi-criteria decision-making (MCDM) family, refined through Stuart Pugh’s selection matrix and Saaty’s Analytic Hierarchy Process (AHP), among others.
Consider a hypothetical case. Company A, an auto parts maker with 320 employees, had to choose a vendor for a KRW 120 million (≈ US$89,800) team-leader leadership program. The previous year it had gone with a vendor an executive recommended, got a satisfaction score of only 3.1 out of 5, and even received an audit finding asking it to justify the re-order. This time it had to narrow three candidates down within two weeks.
- At the first meeting, the HR director set the vendor proposals aside entirely, declared, “Today we only set the criteria, and that’s it,” and wrote five on the whiteboard: instructor expertise, customized design, after-care, price and references.
- The purchasing manager argued for 35% on price, but the training manager pushed back: “We went cheap last year and got a 3.1.” In 40 minutes they settled on instructor expertise 30, customized design 25, price 20, after-care 15 and references 10.
- They first set two must-have conditions: agreeing to the company security pledge and running a pilot within 4 weeks. As a result, Company B could not meet the pilot schedule and dropped out of scoring.
- For the remaining two, three HR staff scored 1–5 separately and added a one-line rationale to each cell. Scores for customized design differed by 2 points, so they reopened the proposal’s pre-interview plan and settled on 3.
- Converted to a 100-point scale, the weighted totals came to 81 versus 76. The team lead said, “The gap isn’t big, so let’s reread the rationale first,” read the after-care comments aloud, and then chose Company C.
- The decision rationale was kept on a single A4 page and drew no audit finding that year. Second-half training satisfaction rose from 3.1 to 4.3, and the same template was reused the next year to select a groupware solution.