Most team decisions don't stall because the options are bad — they stall because the team is comparing options on different, unstated criteria. One person is optimizing for cost, another for speed, a third for risk, and the meeting goes in circles because nobody agrees on what's actually being measured.
A decision matrix fixes this by forcing the criteria into the open before anyone scores the options.
Step 1 — agree on criteria before looking at options
This order matters more than the matrix itself. If you pick criteria after you already have a favorite option in mind, you'll unconsciously pick criteria that favor it. Write the criteria down first, in a room (or a doc) where nobody has stated a preference yet.
Three to five criteria is usually the right range. More than that and the matrix becomes its own debate.
Step 2 — weight the criteria, don't pretend they're equal
Cost and risk are rarely equally important. Give each criterion a weight (1–5), and be explicit about why. This is where most of the actual disagreement should happen — not in scoring the options, but in agreeing how much each criterion matters. That's a healthier argument to have, because it's about values and constraints, not about which option someone already likes.
Step 3 — score each option against each criterion, independently
Score one criterion across all options before moving to the next, rather than scoring one option fully before moving on. This avoids the halo effect, where an option that looks good on the first criterion gets generously scored on everything else too.
Step 4 — multiply, sum, and treat the result as a starting point
Weighted score = criterion weight × option score, summed across criteria. The winning number isn't gospel — it's a structured summary of a conversation you already had. If the top two options land within a few points of each other, that's useful information too: it tells you the decision is closer than it felt, and the tie-breaker should be something the matrix doesn't capture (team conviction, reversibility, how painful it is to be wrong).
Where this breaks down
A decision matrix doesn't work well live in a meeting on a whiteboard that gets erased — the value is in being able to revisit it later when someone asks "why did we choose this," and in updating it when a criterion's weight changes. Keeping it as a board you can edit and reference, rather than a one-time printout, is what makes it a decision record, not just a decision moment.