This problem is very similar to other areas of education and assessment. Assessment by nature is tied to measurement. What you measure is usually an output of a process or a defined task. What you value though may be only tangentially tied to that process or task. When you look at "excellence" you add to the problem by now setting an external value to what has been assessed, may only be partially tied to what you value, and often is tied to inappropriate methods of measurement.

Take the word "command". If I ask you to define the word command you will make basic assumptions on what to do. You will likely pick one of a few desired paths:
1) You reach for a doctrinal publication and find the word command and report back on that as the gospel or definitive source.
2) Perhaps you do something similar and "google" or find a dictionary definition of the term.
3) Maybe you express yourself and how you perceive the word and what it means to you.

But, notice I didn't say what I was assessing and provided no rubric for that assessment. For all intents in purposes I could be assessing your thinking strategies and ability to think outside the box. I could be evaluating your adherence to certain traits I obliquely am testing in military matters. Perhaps I could care less about the word command, but I want to assess your ability to express the traits and qualities that a rubric or other method of assessment values.

The superlative of "outstanding" though can have no reasonable metric or value applied to it without having a real principle of measurement to base the resulting valuation. As a result it is meaningless regardless of what it is measuring unless the strategy of evaluation has been determined. This though can lead to the second problem which is measuring everything with no plan or strategy in place for what to do with those measurements. So you send a lot of people around measuring a bunch of things but there is no real standard to meet and no assessment mechanism for that standard. Just a bunch of metrics. Metrics are not measurement. 50 guys in the company are a size 10 boot. So what. To be valid there must be associated data (even if just a normalized curve).

So.. the principle of "outstanding" is less than an outstanding result of lazy leadership. Or, so IMHO it must be.