Why do so many Data Warehouse Projects fail?
It started with MIS and DSS, changed to Data Warehosing and nowadays it evolved to Business Intelligence. Every few years there are new Buzz-Words, but the main problems seem to remain.
Since DWH-Systems should merge the data of all operational systems, communication between the people involved tends to be complicated.
On the technical level significant Data Quality issues in the operational data often are detected lately after the data was merged into the Data Warehouse.
Besides this, there are 2 typical reasons for failing when starting a Data Warehouse initiative:
- Too big expecations:
The idea, to build the perfect Data Warehouse for all users often overloads a project. The DWH should be "the single version of truth" in the end, but when you try to achieve this in one big step, the chance to fail is high.
This appraoch tends to take too much time before anything useful can be deliverd.
- Too small steps:
To avoid the problem described above, it makes sense to start small - with a specific area. There is the danger of losing the "big picture" in the following phases, and you might end up in a lot of not integrated data marts.