Building the Enterprise data warehouse is off course not a goal in itself. By definition the business owns the business case that leads to the strategy where you actually need an Enterprise data warehouse. Let's assume the business made a pretty convincing case that leads to building an EDW (lets abbreviate it). This is by the way often a huge undertaking; to get a generally accepted (by the CxO) business case!!!
The trouble now is; your not there yet. Let's use the infamous 'I want a car - metaphor'. The business case is set; u need a car to make money. But what kind of car? A VW? Mercedes? Toyota? A frenchie? or maybe a chinese one......
How do you decide on the type of 'car'? Not that difficult I think, you will need to state your ambitions. Do you want to impress your neighbours with a big engine? Do you need to travel with your kids now and then? whats the luggage space you need? What's your budget? etc...
So - building the EDW is not that different compared to buying a car. You need to state your amibitions first. These amibitions guide you in determining the right architectural choices all the way. These ambitions need to be accepted by your sponsor in full!!!
In the next paragraph I will sum up some of those ambitions. In the coming weeks I will go through them one by one in more detail. Let me emphasize; these are possible ambitions, it may not be your ambitions. It's highly dependent on the type of industry your in!!!
The ambitions mentioned here are not sorted from high to low prio or something.
1. Seperate your data from your source systems
2. Complete system of record - all history is recorded
3. Source systems are interfaced/extracted once and distributed many times for many different functions (one of those functions is ofc BI, but there are more!!)
4. Data compliance; complete accountability of data
5. Data without (business) definitions; not allowed
6. Delivery of a data quality process, not delivery of good data quality (which does not mean you do not need a full blown data quality program - YOU all need it!)
7. Data can be confronted with each other
8. An infrastructure designed for performance (e.g. technical and functional scalabiltiy) and user friendliness
9.Increasing time to market of new informational products (in other words; data faster to the needy one)
10. An agile architecture - forsee the unforseeing and manage continues change coming from the source systems as well as from the end user.
11.Support and complement the data centric and master data management strategies
12. Delivering front-end services to the business (not only BI!!!)
13. The EDW must support operational, tactical and strategic processes