Databases Reference
In-Depth Information
Question: Do other companies really build these in
90 days?
The ability to deliver in 90 days is closely related to the previous
question — how long will it take? Again, the most important factor to consider
is what you are trying to build. If you are putting in place the foundation for
the future of the data warehouse, bringing in all new technology and starting
from scratch (or starting over) to meet business needs, then you are unlikely
to accomplish this in 90 days. However, if you are building new reports and
analyses that use data that is already loaded, this can often be done in 90 days
or less.
Answer
Making time estimates can be very deceptive if you only look at how good
the technology is. The tools can be set up fairly quickly and can move data
around faster than ever before, so why can't we build a data warehouse in 90
days? A good data warehouse is not about the technology. The value lies in the
effort made to define the data elements and develop the business rules needed
to clean and integrate the data. This requires a great deal of thought and
discussion among and between business groups and IT. Then decisions must
be made. Regardless of how fast technology can implement those decisions, it
still takes time to research and make them.
It is also important to understand what is actually being delivered from
a 90-day project. Often, it is not a complete robust environment but a quick
prototype. If the prototype is not quite what is needed, then another 90-day
effort is launched to change what is delivered. Many organizations iterate
many times, hoping to get it right. It is important to understand and agree
upon what must be in place for success. Number one on the list? The data
warehouse must be driven by well-defined business requirements.
The amount of time required to put a robust data warehouse in place can
vary greatly. Rather than approach this by demanding a completed project in
90 days, determine what results you want and then figure out what has to be
done to develop a sustainable, long-term solution.
Once you have a solid foundation in place, the project schedules can be
shortened. The work to enhance a dimension or to add another fact table is
much less than the work to start at the beginning. It is also much faster to
build new BI reports and perform analyses using data that is already loaded
into the data warehouse. Many of these subsequent projects can be done in
90 days, including the testing and release activities. Ultimately, investing up
front to get a solid foundation in place will save the organization a great deal
of time and money.
