This article originally appeared on the BeyeNETWORK.
Recently I was in Toronto to discuss the state of data warehousing with some people from a large bank. They asked me to write an article on a subject that I thought was dead. But it seems that this subject just won't die.
In our discussions we talked about the reasons that data warehouses had been so vehemently resisted by their organization. Among other things, the people said, "Well, everyone knows that it takes a long time and a lot of resources to build a data warehouse. Everyone knows that you have to build a data warehouse as a large single project." I confess that I went a little ballistic at that point. I thought that dragon had been slain years ago.
In fact, the opposite is true. Data warehouses should be built in a fast, iterative manner where tangible results are immediately apparent with a minimum of investment. My friends - Claudia Imhoff, Joyce Montanari, Bob Terdeman, Dan Meers, Jonathan Geiger, John Ladley, Sid Adelman, Larissa Moss and others - have always advised that the way to build a data warehouse is to build it in small, fast increments, one iteration at a time. The knowledgeable people who have built successful data warehouses have NEVER said that a data warehouse should be built as a single, large project. And it has been this way from the beginning.
Let me quote from one of the first references to the data warehouse. In this book on data warehousing, BUILDING THE DATA WAREHOUSE, QED, 1989 in chapter 7, page 185 it says -
"Any architecture that must be implemented all at once, in a Big Bang, is doomed to failure in today's world. There simply is too much risk, too long a period to wait until there is a payback, and the unreality of trying to freeze changes to consider any path that is revolutionary, rather than evolutionary.
It is very good news indeed that migrating to the architected data warehouse environment is a step by step activity that is accomplished one finite deliverable at a time. The most successful implementations of the architected environment are those where the data warehouse has been built one iteration at a time. In doing so the data warehouse can be built with absolutely minimal disruption to the existing application environment."
The simple truth is that data warehousing - built properly - has never been a Big Bang proposition. And it is documented clearly and openly that that is the case. In fact, your data warehouse will continue to evolve one iteration at a time until your company ceases operation.
Bill is universally recognized as the father of the data warehouse. He has more than 36 years of database technology management experience and data warehouse design expertise. He has published more than 40 books and 1,000 articles on data warehousing and data management, and his books have been translated into nine languages. He is known globally for his data warehouse development seminars and has been a keynote speaker for many major computing associations. Bill can be reached at 303-681-6772.
Editor's Note: More articles, resources and events are available in Bill's BeyeNETWORK Expert Channel. Be sure to visit today!