There are a lot of different theories and practices as to how to measure human performance in an organization. I have found that the most effective way is to develop up front the metrics by which the team will be measured. These metrics should be related to the strategic outcomes desired for the project. So if the main goal is to optimize costs, focus some of the measures around costs since timeline, scope, etc. are of secondary performance.
I've also noticed that the triple constraints create what I refer to as "triangular tension". By focusing on all three, performance is driven toward optimizing all three. So in the absence of any specific stated goals by the project sponsor, you might seek to develop a situation involving some tension between the goals.
For example, I worked in an organization which had client satisfaction and six-sigma performance as a goal. So as a whole, the organization was measured on meeting six-sigma system up times (99.999% availability), no critical open bugs, and a client services report card filled out from client input. There was a direct correlation between the availability and bugs as measured against client satisfaction. So if bugs went up and availability went down, client satisfaction would also dip. Therefore everyone focused on meeting the availability metrics and making sure the products were high quality. These goals were factored into decision-making at all levels of the organization, including at the project level.