This chapter aims to proposes the idea of usage of functional dependency as a criterion that database designers and administrators can use to decide whether to normalize transitive functional dependencies. It discusses a theoretical framework and method of denormalization to provide a rigorous tool for deciding whether to normalize or denormalize relations. The chapter describes a method for making this denormalization decision, based on the idea of functional dependencies. In practice of database systems design, a relational model is designed and normalized up to a certain normal form, and a normalized relation is subsequently denormalized ad hoc, as necessary to improve performance. Normalization theory was developed to reduce potential update anomalies by providing a systematic way of aggregating or disaggregating attributes. Because normalization of data usually improves data integrity, it also increases the effectiveness of a database system’s long-term maintenance. Normalization theory is based on functional dependencies in a relation.