How to find "good" aggregate candidates?
I'd like to know how one can find good aggregate-candidates - without hopping from bw-user to bw-user and ansking him "which characteristics do you use mainly"?
I know that the rsrt gives the possibility to show the aggregate that would have been used if it would have exist.
The hatch: Aggregates are on a "per cube" base - not really on a per query-base.
It would be nice if the BI-Statistics could provide information about the mainly used characteristics-combinations per Cube. But I did not find such an information yet.
So my question is: How to you determine good aggregate-candidates - without "guessing" and "user-hopping"?