In relational databases, normalization is a process that eliminates redundancy, organizes data efficiently, reduces the potential for anomalies during data operations and improves data consistency. The formal classifications used for quantifying "how normalized" a relational database is are called normal forms (abbrev. NF).

A non-normalized database is vulnerable to data anomalies because it stores data redundantly. If data is stored in two locations, but later is updated in only one of the locations, then the data is inconsistent; this is referred to as an "update anomaly". A normalized database stores non-primary key data in only one location.

Normalized databases have a design that reflects the true dependencies between tracked quantities, allowing quick updates to data with little risk of introducing inconsistencies. Instead of attempting to lump all information into one table, data is spread out logically into many tables.