Decision Matrix

Frequently asked questions

The questions people actually ask about this document.

What is a weighted decision matrix?

A grid that scores every option against every criterion, then weights each criterion by how much it matters. The total for an option is the sum of weight × score. It turns an argument about a choice into an argument about weights and scores, which is a far more productive argument to have.

How do I choose the weights?

Start by ranking the criteria rather than numbering them: which one would you refuse to compromise on? That one gets the highest weight. Then place the others relative to it. Do this before you score anything, because weights set after the scores are visible are weights chosen to produce a result you have already picked.

Should I score out of 5 or out of 10?

Five when the judgements are coarse and you want to avoid false precision, which covers most team decisions. Ten when you genuinely can tell an 8 from a 7, which usually means you have data rather than opinions. The scale does not change the ranking, only how much room you have to express small differences.

How do I handle cost, risk, or anything where low is good?

Mark the criterion “lower is better” with the arrow next to its name, then score the raw figure: the cheapest option gets the lowest score and the tool flips it when it counts. The alternative, scoring “how good is the cost”, works until two people interpret it in opposite directions, which they will.

Can I not just pick weights that give the answer I want?

Yes, and that is the most useful thing the tool does. If you find yourself adjusting weights until a particular option wins, you have discovered that you already made the decision, and the weights you reached for tell you what you actually care about. Write those down and justify them, or admit the choice was made on something not in the matrix.

What if two options come out almost level?

Then the matrix has told you something real: on the criteria you named, there is nothing to choose between them. Do not break the tie by nudging a score. Either find the criterion you left out (there usually is one) or accept that the decision is cheap to get wrong and pick the one that is easier to reverse.

When is a decision matrix the wrong tool?

When one criterion is a veto rather than a weight: if an option that fails security is out regardless of its total, that is a filter, not a row. Apply it first and score what survives. It is also wrong for decisions with two options and one obvious dimension, where the grid is ceremony around a choice you could make in a sentence.

How is this different from a Pugh matrix?

A Pugh matrix scores everything relative to one baseline option as better, same or worse, which is fast and deliberately rough. A weighted matrix scores on an absolute scale and multiplies by weights, which is slower and more precise. Pugh is for narrowing a long list of concepts; this is for deciding between the few that survive.

Related tools

• RACI Matrix Builder

  .raci

  Assign who is Responsible, Accountable, Consulted and Informed for every task, and catch the gaps.

  Available

• Eisenhower Matrix

  .eisen

  Sort work by urgency and importance into four quadrants you can act on.

  Available

• Decision Tree Maker

  .dtree

  Map choices, outcomes and probabilities into a tree that calculates expected value.

  Available