Implementation Challenges

[teccum]

Just wanted to find out the various data warehouse implementation challenges that others are facing out there. Please share your experiences.

thanks
teccum.

 

[johnherman]

History is often of different format or different level of data quality (or both) than current data.

-------------------------
The reasonable man adapts himself to the world. The unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man. - George Bernard Shaw

 

[MDXer]

1) Lack of input by the business line users.
2) Lack of Specs or unrealistic specs.
3) Improper hardware (Sorry a 600GB warehouse won't run efficiently on a 2 proc 2GB memory server)
4) Lack of improper Development Lifecycle.
5) Poor Data Quality
6) Lack of Data
8) Lack of Cooperation from clients IT Staff (for Consultants)
9) Scope creep. Once you have a working demo all of a sudden you get an I wanna list with things that where never discussed early on in development.
10) Cost. Sometimes you don't get enough funds to get the proper hardware or developers. Other times companies believe just throwing more servers at the problem will make it go away when really it is the design that is the problem.

Just to name a few


Shoot Me! Shoot Me NOW!!!
- Daffy Duck

 

[teccum]

Yes, U r right. There are more challenges than mentioned. Here are few more are.

1. Wrong Estimation for the total development cycle and even on every phase of the development cycle of the project and sometimes it is an irony that professionals who have not worked on data warehousing projects doing the estimation.

2. Frequent requirements changes by the business users for which there is no proper CMC.

3. No peer reviews or quality testing at every stage of the project. This is being ignored assuming what one does is 100% correct at the first time.

4. Not sharing or doing proper knowledge transfer to the consultant assuming he should know about the business and the expectation that the consultant should pick up the entire business knowledge in few days starting working on the project.

5. Trying to Work on ETL and Reporting without creating a detailed proper specification.

6. Data load problems into the data warehouse where there is no proper backup strategy.

7. Not creating any documentation at each stage of the project (Proper documentation would be immensely helpfull at the maintenance stage of the data warehouse and also when we asked to work on a existing data warehouse).

8. Improper estimation on the size of the server required for the data warehouse.

9. Business and Users expecting that all types of reports in the world should be accomplished with a single OLAP tool. My experience is, B.O, Cognos and MicroStrategy have their own pros and cons. At some point or the other certain features are not possible with these tools.

10. No proper training on the OLAP tool to whomsoever who is going to access the data warehouse.

thanks
teccum.

 

[johnherman]

As someone with experience in the matter, teccum's point #9 is very much true. And I might add, despite the desires of management, one single tool is NOT going to solve everyone's problems. BO, Impromptu and Microstrategy have been accused of being "too complicated" by some users (Marketing) where Brio is loved. Brio is accused of being too simplistic by other (Accounting, Operations) and they want one or more of the others.

-------------------------
The reasonable man adapts himself to the world. The unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man. - George Bernard Shaw