This is a great question. I always find it interesting to read the different responses from the entire class but I have always found the common theme or understanding of normalizing is to reduce redundancy within systems. Over the years I have tried to emphasize to customers the importance of Normalization because I look at redundancy as a reports/queries worst nightmare. The reason that I state this is the fact that these redundancies can skew results which in turn can produce negative outcomes for data users. If Normalization does not occur then the consequences can be catastrophic to any organization. Once organizations understand the importance of Normalization it produces the desired outcomes that will not only put a smile on the executive’s faces but can potentially save an organization from data failure.
What comes to mind when I think about this question is the old concept that “One size does not fit all”. Now this may be true in some instances but not all. Normalization is a perfect example. The levels of Normalization that occur with a system can be vastly different between the respective systems. This is why when requirements are being gathered and conceptual designs are being designed this is when Normalization is determined and at what level. So when determining at what level Normalization must occur you must clearly understand the data and requirements. Below is a great article about Normalization. Take a read and let me know your thoughts. 🙂