Databases Reference
In-Depth Information
expand as the data warehouse grows. It is worth building a strong foundation
so that other data warehouse projects can leverage what has been built. If you
are adding to an existing data warehouse or building a new data mart, the
overall cost may be significantly less, if you can leverage the technology that
is already in place. In this case, most of the investment will be centered on
the data and delivery into the business community. Remember to highlight
the costs that are an investment for the entire enterprise so that these can be
isolated from the project-specific costs. Project-specific costs should be used to
determine the return on investment for this project.
Building a data warehouse is not like ordering a laptop computer. Unfortu-
nately, there is no central website where you can make a series of choices and
see the price build as you go. While basic hardware and software costs can be
estimated, however, some of the biggest costs result from the people required
to do the work. The most costly part of the project is during the construction
phase. Many resources are needed for extracting, transforming (or preparing),
and loading the data for analysis.
Data warehouse budgets range from tens of thousands into the millions of
dollars. These costs are influenced by the following:
Size of the data
Complexity of the business itself
Overall size of the organization
Some small companies have extremely large volumes of data and high levels
of complexity. Some large companies have a lot of data, but the analyses are
straightforward.
For new data warehouse efforts, it is advisable to break the project up into
smaller pieces. It can be extremely difficult to estimate the effort required to
design and build a data warehouse before full requirements are gathered.
After all, you don't really know what you are going to build! That is why you
must set up an initial project to gather requirements (Chapter 6), define the
overall data architecture (Chapter 9), and then design the solution (Chapter 7).
Once the data model is designed, you know what will be built. Then it will
be reasonable to estimate what it will cost to build and deploy the data
warehouse. For subsequent work, it is still advisable to break things up into
more manageable projects with shorter timeframes to deliver results.
The most accurate estimates of effort are created with the involvement of
the people who will actually be doing the work. The data modeler should be
consulted prior to estimating the time needed to develop the data model. In
addition to the estimates to complete specific tasks, the overall effort must
include project management and administrative overhead. The task estimates
and overhead can be combined to provide a realistic estimate of the effort
needed. This can be translated into costs by applying the organization's
