There are three often-asked questions of any project in the IT business: How quickly can you build it, how good will it be and how easy will it be to build? These three attributes of your project can be displayed as the three sides of a pyramid as shown in Figure 1.

Figure 1: The Good, the Quick and the Easy Pyramid

You may be familiar with the old adage that any project can have only two of these three attributes ­ that is, quick and easy but not good, good and quick but not easy, or good and easy but not quick. This adage often holds true in data warehousing projects as well. If you don't have an architecture as the underpinning to your business intelligence BI environment, you may have only two of the three attributes.

However, with a solid, well-planned architecture as your starting place, you can actually achieve all three of these important characteristics in your BI project. Let's look at each one in detail.

The Good. What are the characteristics that constitute a "good" project? Most people would agree that it is good if it is on time, on budget and if it meets or exceeds the business community's expectations. The project is also good if it is maintainable, flexible and easily enhanced.

The Quick. A quick project is one that is created in a timely manner. For business intelligence, this usually means that you have the application up and running in three to six months or less.

The Easy. An easy project is one in which minimal difficulties are encountered. For BI, this means minimal extraction and transformation problems, quickly resolvable integration issues and a well-documented environment.

Sounds like nirvana, doesn't it? Not too many of us can say that we have an environment that supports any two of these attributes, much less all three! Following are two common scenarios I encounter with IT projects on a regular basis.

Quick and Easy But Not Good. Unfortunately, I encounter a lot of companies that have taken the fast and easy approach to building their BI environments. This is also known as the quick and dirty approach. Typically, this situation arises from a focus on only the data mart application to the exclusion of the overall architectural need. This approach leads to data mart chaos fairly rapidly (see Figure 2). The appeal is obvious ­ you meet the established time line and encounter minimal difficulties because there is no need for enterprise-oriented integrated or consistent data.

Figure 2: Non- Architected Marts

What should be equally obvious is that this shortsighted approach leads to extremely unsatisfactory results down the road. Often the results are so abysmal that the entire effort must be abandoned and declared a failure.

Good But Not Easy or Quick. Building a sustainable architecture for your BI environment requires significant commitment for the initial project. Much of the effort to build the first "in- architecture" data mart is spent on infrastructure issues rather than actually satisfying the business community's analytical needs. Taking a broader view of the long-term environment means that this first project must:

  • Solve the very difficult extraction and integration issues from an enterprise perspective,
  • Document the transformation rules, definitions, calculations and data quality audits and controls in a solid meta data environment,
  • Build a proper data delivery architecture using a data warehouse as the source of data,
  • Choose a tool suite that is scalable and integrates well (ETL, data delivery, access, data cleansing, etc.), and
  • Create a consistent user interface infrastructure to the application (e.g., Web interface).

These steps have little or nothing to do with the ultimate end-user application. They are infrastructure components that are reusable and lead
toward a consistent and maintainable environment. The first project, unfortunately, will not see the benefit of this thoughtful, long-term planning which is what makes this approach so very difficult to achieve. It is the succeeding projects that benefit from this foresight.

Good, Quick and Easy!

Can all three attributes ever be accomplished in a single project? Yes, but only if you have taken the time and committed the resources to build your architecture first. Let's look at the second project in our architected environment described earlier.

  • The next project is quick. We are working with a well- documented environment using nicely integrated tools. Getting new data into a well-documented environment is much faster and easier than starting from scratch again.
  • The next project is easy. We have a clean, documented source (data warehouse) of data for our next BI application. Many of the data integration problems have been solved for us, and those remaining may be handled faster and easier with the established procedures. We can quickly extract data to build the next application.
  • The project is good. We have consistent results, definitions, calculations, etc., and the application fits nicely into the existing user interface. The application is maintainable, flexible and easily enhanced.

The second project may not proceed as smoothly as I have described, but it certainly will be significantly faster, easier and, therefore, better than the alternative chaos. With each successive BI project building upon this architecture, you really begin to receive the benefits ­ each project rolling out is good and quick and easy. Yes, you can achieve all three!

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access