Why coarse beats precise
A Pugh matrix refuses to let you say seven out of ten. That looks like a limitation and is why the method works.
Ask only what people can answer
People cannot reliably score seven concepts out of ten across eight criteria. The numbers drift, anchor on whatever was scored first, and change on a second pass. It feels rigorous and is not.
The same people can reliably say whether something is better than what they have now. A Pugh matrix asks only that, and gets a more honest result by refusing precision nobody possesses.
Everything compares to the baseline
Concepts are never rated against each other, only against the current approach. That keeps each judgement simple and independent, and it means adding a concept later does not disturb the existing ratings.
It also means the baseline matters. A strawman baseline makes every concept look good; a strong one makes them all look marginal. Use the current approach honestly — the question is whether change is worth it, and a rigged comparison answers something else.
The ranking is a by-product
The usual next step is not to pick the winner. It is to take the strongest concept and attack its minuses: can this be changed so its weaknesses disappear, borrowing from whichever concept scored well there?
That produces a hybrid better than anything originally on the list, and it is the reason the method exists. A Pugh matrix filled in once and filed has been used as a scoring sheet rather than a design tool, which wastes most of its value.
When to add weights
Unweighted, every criterion counts equally, and a concept that is better on five trivial criteria beats one better on two critical ones. Weighting fixes that and reintroduces some of the judgement the method was avoiding.
A reasonable middle path is to run it unweighted first. If the answer is clear, weights would not have changed it. If it is close, add weights and watch what moves — the movement itself is informative about which criteria are carrying the decision.