In a recent model, a client was using population, growth, and historical performance as decision criteria. We suggested thinking of these criterion as "objectives" -- there is a fine line between criteria and objective, but there is a difference.
For example, "size" could be a criteria when selecting a new car. But the problem with using just "size" to describe what you want is that it does not tell you if "big" size or "small" size is desired.
Alternatively, thinking of the problem in terms of "objectives" you might say, "I need a car that is easy to park in the city..." Perhaps my "objective" then is "easy to park?" There may be larger cars with tighter turning radius or maybe I really do need a "short car" to be able to fit into small parking spaces--a Smart Car. Thinking in terms of "objectives that I want to achieve" forces further analysis and tends to generate questions that could refine the decision criteria.
So with regard to population and historical performance, we asked "what objectives are you trying to achieve with these?" Engage where population is greater than x? Where historical performance exceeds y? ...and so on. These questions lead to better understanding of what needs to be achieved by the alternatives in the model.