We've found/learned that the most important part of creating an AHP (Analytic Hierarchy Process) decision model is creating the basic framework or structure of the model.
This includes defining the goal of the decision you are trying to reach, brainstorming and articulating the various alternatives that you have for the decision, and finally defining the criteria by which you will evaluate the various alternative options.
The decision criteria can also be called "objectives" or "outcomes" depending on the type of model you are building. For example if I was prioritizing a group of projects that had been assigned to my function I might use "criteria" to describe the factors that would influence the ranking of alternatives since I was not in a position to reject the projects, but rather I just need to prioritize them.
However, if I was one level above this decision and was trying to determine what projects should be in our portfolio I might use the term "objectives" to describe the what I am trying to achieve with the portfolio mix of projects with respect to my organization goal. Those projects that did not meet my business objectives would rank low and could be "killed" in order to reduce my budget and focus my resource allocation on the projects that would deliver the highest benefit (with respect to my objectives).
The structure of the model is the key to an effective model. We often have to iterate many times through the model before we settle on one that will work. This process is most effective in small focused groups of people, prior to larger group facilitated sessions. The focus of the group sessions should be on ranking and simulating the model, not on how it is structured.