 MANAGING
A DATA WAREHOUSE |
Gaining senior management commitment is an essential first
step. A data warehouse initiative needs an executive sponsor
willing to support the effort by mobilising the necessary
resources and facilitating the resolution of business issues
as they arise.
The following tips reflect widely accepted best practices:
- Use a pilot to discover real requirements, and to help
choose tools.
- Refine and expand the warehouse iteratively, in small
stages. This approach is sometimes called spiral development.
- Aim to create a data
warehousing department, not just a project team – the
skills needed to build the initial warehouse will also be
needed for ongoing maintenance and future expansion.
- Use experienced people in the initial stages, but make
sure they transfer skills to nominated permanent staff from
the outset.
- Involve key users, and let the business lead.
- Don’t just reproduce current reports - experiment with
ad hoc
analysis in parallel.
- Defer major investments until they can be supported by
a clear business case.
Before embarking on development, it is essential to plan.
Good project management practices can and should be applied
to any data warehousing initiative, but these need to be combined
with an understanding of the technologies and processes involved.
Because of this, a data warehousing department is often set
up as a specialist management information unit within IT.
The following sections offers guidelines for managing a data
warehouse, and identify some of the key issues that need to
be addressed:
- Planning
- outlines the main steps needed to build and maintain a
data warehouse;
- Pitfalls
– highlights the most dangerous traps that may catch the
inexperienced, and suggests ways to avoid them.
For a more in-depth discussion of data warehouse management,
see the books by Inmon,
Kimball
and Adelman.
|
