A weighted score can help a team talk clearly about a choice, but the decimal result is not independent evidence that the winner is right. People choose the criteria, weights and ratings that generate it. The useful question is whether those inputs represent the decision and whether the ranking survives reasonable disagreement. Build the matrix to expose the assumptions, then test them. This guide uses a small invented comparison so you can see exactly how score direction and weight changes affect the result without hiding behind a large spreadsheet.
Eliminate infeasible options first
Write down any requirement that must be met regardless of a high average score. An option that cannot fit the available space or meet a required compatibility condition may need to be removed before scoring. Do not let excellent performance on unrelated criteria silently compensate for a non-negotiable requirement. Workday Kit does not enforce such gates. Its decision matrix ranks the rows you provide, so responsibility for deciding which alternatives belong in the comparison remains with the people using it.
ASQ describes a decision matrix as a way to compare alternatives against chosen weighted criteria. That framework is useful because it requires the discussion to become explicit. For the invented example, compare Studio A and Studio B using Quality and Speed. The labels are intentionally simple; they are not a claim that two factors are enough for every supplier selection. A real decision may also require evidence about scope, availability, service, ownership or other relevant constraints before those alternatives can be meaningfully scored.
Make every high score point in the desirable direction
Workday Kit expects ratings between zero and ten, with higher always better. For Quality and Speed, specify what a high score actually means. A speed rating might represent desirability of a verified delivery proposal rather than the raw number of days. If you enter delivery days directly, a slower option receives a larger number and can be rewarded by mistake. Explain the conversion between observed evidence and ratings instead of assuming the tool understands the units behind a number.
Cost creates a similar trap. Use a criterion such as Affordability, with a stated desirable direction, rather than entering a currency amount into a score column. ASQ specifically highlights consistent score direction in its matrix guidance. Avoid double-counting closely related criteria: Delivery speed and Fast turnaround may reward the same advantage twice unless they genuinely measure different things. More columns do not necessarily make the model more balanced. Each criterion should add an identifiable part of the decision, supported by information you can discuss.
Calculate a transparent initial result
Give Quality a weight of sixty and Speed a weight of forty. Studio A receives ratings eight and six; Studio B receives six and nine. A's weighted total is eight times sixty plus six times forty, or 720. B's total is six times sixty plus nine times forty, also 720. Divide each by the total weight 100 and both score 7.20. This is a tie under the supplied assumptions. The tool identifies both highest-scoring options rather than inventing a distinction.
Weights need not be percentages. Six and four produce the same normalized scores as sixty and forty. All-zero weights are rejected because there is no meaningful weighting basis to normalize. Criterion names must be unique and the identity column name Option is reserved. Match the score-table headers to the criterion names and their order. These input rules protect the structure of the calculation, but they do not validate the truth of a rating or prove that the selected criteria capture everything important.
Change one assumption and watch the ranking
Now increase Quality's weight to seventy and reduce Speed's to thirty while keeping the ratings unchanged. A scores 7.40 and B scores 6.90. If the team instead weights Quality at fifty and Speed at fifty, A scores 7.00 and B scores 7.50. The preferred alternative changes because the priorities changed, not because new performance evidence appeared. ASQ's discussion of weighted decision sensitivity provides background for asking whether small input changes alter a conclusion. The specific scores here are independently constructed demonstration arithmetic.
Treat that sensitivity as useful information. It may show that the team needs to agree how quality and speed trade off before debating which studio is better. Alternatively, the ratings themselves may be uncertain. Ask what evidence would justify moving a rating from six to seven, and whether different reviewers are using the same interpretation. Do not average disagreement automatically when it reflects incompatible definitions. Clarify the question first so any later numbers mean approximately the same thing to everyone involved.
Distinguish selection from task prioritization
Save the matrix with a short decision note: which alternatives were eligible, what evidence informed the ratings, why the weights were chosen and which sensitivity checks were tried. If the final choice differs from the top score, explain the additional consideration instead of secretly changing weights to make the preferred option win. A useful matrix makes judgment more visible. Its strongest outcome is a reviewable conversation and an explicit basis for action, not a display of mathematical certainty that the underlying evidence cannot support.