We are in the process of building a data warehouse for our General Ledger Business Process. What is the recommended level of detail for performing user
General Ledger analysis
? What should be our grain as it pertains to General Ledger? Should we keep it at a very generic level? Or, should we go for a classification of balances into revenue, expenses, profit and loss, cost of goods sold, etc?
In general (no pun intended), although the majority of
) queries are run against summary data, the majority of the ROI from a data warehouse comes from the detailed data. So, you may think that you are providing "most" of the requirements by providing only summary data, and you are, but you are not providing the most "value." You can always summarize up from the detailed data, but cannot go the other way. I would suggest starting your General Ledger data warehouse project on the right foot - by sourcing the most granular detail you can. It sounds like your users will want different cuts of that data. You may have to classify, summarize, etc. from there before turning users on to it, but that is how I would begin.