SSAS - How to deal with Aggregation design?

Question

I'm working on a SSAS project to build a demo cube.

I created few dimensions (Category, Time, Store). I would like to allow users to browse Sales, so I configured a fact table Sales (linked to my previous dimensions).

I would like display a lot of information (such as quantity, sales amount, prices, costs and more). However I don't know what aggregate behavior to use depending on my case of study.

For example, if I want to display the total ($) of sales for a specified store / month / category, it's seems I need to use a SUM aggregation. However, If I need to display a data specificly to a row such as the sale price of a unit, or the quantity of items by sales. Should I use a NONE aggregation because it doesn't have sens to sum the quantity of different products by month.

Can someone explain me how to do this ?

Answer 1

I would add calculated measures to derive those values. If the user/query is down at the most granular level they will calculate for that specific row. At higher levels they will provide useful analytic value, e.g. compare sale price by store or region.