Database Reference
In-Depth Information
involved. Make sure to monitor the solution's progress and publish these results. We have used basic web pages,
bulletin boards, and wiki pages in the past. All of these are simple, inexpensive, and effective.
How Will We Build It?
Building a BI solution can be approached in two ways: a top-down approach or bottom-up approach.
In the top-down approach, the needs of the company as a whole are determined, including all the data that a
company uses, with the ultimate goal of creating a solution that will cover all the needs of all users as soon as the
entire BI solution is published. In some cases, this is necessary. In these cases, decision makers need all available
information through a single viewpoint to make decisions. Interim reports and partial updates are insufficient.
Looking at a half-dozen viewpoints is not feasible either. The information from the BI solution is not just an
additional tool used in the decision-making process; it is the primary tool. Here, all of the components must be
present before the solution is of use.
In the bottom-up approach, the focus is on defining a particular business process and building a solution
around this process. The idea is that you can add onto the solution by focusing on a different process in the next
version. You continually refine the solution by analyzing the process and incorporating it in the current solution.
Your BI solution is only one tool used in the decision-making process, and the solution can be functional even
when it is only partially complete.
■
The top-down versus bottom-up approaches have been defined as the two paths to choose from. Each is
commonly associated with two different developers, Bill Inmon and Ralph Kimball, who are both early (and often com-
petitive) authors on the subject of data warehousing. In the end, both are correct approaches, just not for all solutions.
Tip
Over the years, the top-down approach has lost many of its adherents. This is mainly because the approach
is often extremely costly and impractical. Therefore, the recommended practice is to use the bottom-up approach
whenever possible. Conversely, when the bottom-up approach is impractical, switch to the top-down approach.
The trick is asking the question of what is necessary for this particular BI solution. As a consultant, you may not
initially know the business you are building the solution for well enough to predetermine what is necessary, so
make sure to conduct interviews with this question in mind.
As you may have guessed, in this topic we focus on the bottom-up approach, using techniques associated
with RAD. (We first mentioned RAD when discussing how to determine the hours it will take to build your
solution.) With RAD development, the focus is on supplying the top 80% of the requested requirements. Quite
often, this 80% can be accomplished in about 20% of the time it would take to do a full implementation of all
requirements. Although this is not always the case, you might be surprised by how often it is accurate. When it is,
you are able to supply your customers with a satisfactory BI solution at a fraction of the cost.
Who Will We Get to Build It?
Finding good people to work on a solution is easier said than done. It is a wonderful fantasy to imagine you have
all the experience and all the knowledge and all the free time to perform every task the solution requires within a
timely manner; however, it is unrealistic to believe that you are capable of accomplishing all of this on your own.
What can be done to mitigate the reality of human frailty? In a word: teamwork!
Putting together a team of developers with compatible skills and personalities is a rewarding and lucrative
endeavor. In addition to the developers, a team should consist of competent managers, testers, and interested
stakeholders. We highly recommend looking for compatible personalities first and sufficient skills second. Although
others may argue that we have this reversed, experience has shown that skilled team members who can work
together will achieve much more than a group of even better skilled individuals who refuse to cooperate with one
another. Failure to build a compatible team costs money and time, two commodities that are often in short supply.
