The cube is not that complicated. Nway aggregation, shallow time dim (only months) the most complicated hierarchy is 4 level internal hierarchy tree, simple calculated members (ratios etc) . Most of the work is done by the ETL beforehand, creating the "slices" ( each slice contains a customer,base month, new month, the data contains his movement between hierarchies' members , i.e. John was premium in 03/2014, nowdays not-so-premium. The cube shows movement totals, i.e. comparing mar 2014 to 2017 10 clients moved to premium, 15 clients moved from premium to ordinary, 10 clients are new, 20 left, 150 stayed in premium,1000 stayed in ordinary class
... View more