Enterprise Software Solutions 

  Jump to: DW University, Employment

To estimate a project, follow these steps:

1. High-level business requirements: Identify some basic key metrics, with some dimensional data
2. Identify Source systems (include key facts, grain, and dimensions)
3. Determine how many months of data are required to be available
4. Develop a high level logical data warehouse model of dimensions and very key facts, including the grain of these facts, based on source systems and requirements
5. Identify which dimensions of those identified in the logical model need to be reported on as slowly changing vs. fixed (current only) dimensions
6. Identify basic metadata requirements
7. Develop a data mart strategy (mart subjects with dimensionality, grain of metrics, and time range for each)
8. Based on what is learned from the steps listed above, identify an ETL tool
9. Based on all of the information above allow a few experienced professionals in the field of data warehousing estimate the project and use the statistical mean.

Using what is learned from the analysis described above, a basic idea of the scope and complexity of the data warehouse can be understood and , therefore, cost estimated. The lack of any of the above information will certainly be devastating to the accuracy of the project estimate. The number and complexity of reports will give no meaningful estimate for a DW project because the complexity of the underlying data is key, not the number of times that data is formatted into a report.

What is important to understanding the complexity of a data warehouse. The number of reports and any level of complexity is meaningless information and should not be used in a DW project estimate. This approach will only work in OLTP system reports.

[ Home ] [ Up ]

Send mail to Webmaster@ESScorporation.com with questions or comments about this web site.
Copyright © 2004 Enterprise Software Solutions, Inc.
Last modified: July 29, 2004