Business Guidelines for Successful Data Warehousing
The list of points below should provide a set of guidelines for the design, construction and management of a successful data warehouse in any environment.
Whilst the end result may well be a compromise based on affordability, if the general guidelines are not followed, success is not assured.
Because you need it
- Your data warehouse should be built on the premise of needing it rather than be an imperative driven by an Information Technology department
- Data warehouses need heavy resourcing initially especially in the analysis phase so business buy-in is essential
Designed to get data out of
- Traditional database environments are designed to get data into
- Your data warehouse should be designed to get data out of
Should be able to change at the same speed as your business
- Your marketing and business plans change rapidly - the systems that support those decisions should be able to adapt quickly as well
Should be able to grow rapidly ⁄ exponentially
A successful data warehouse will grow to fill the needs of all who see the benefit of using it. You must expect it to grow.
Therefore, the decision that was taken to build it on a PC initially may not be a long term solution.
Do not limit yourself to a single operating system or a language or database platform that only works in one operating system.
Should be a best of breed solution
A successful data warehouse should take advantage of ALL the tools in the market place available.
Just because a software company builds a good database they do not necessarily build the best BI tool. And so on...
Good "glue" can make them all work together, so pick the best of each.
Should NOT need a cast of thousands to operate
A successful data warehouse can (and should) be run by a small team with the right tools and education.