A relational database allows you to ‘normalise’ your data. Normalising data means that you only have to store repeated information in one place.
For example, if you are storing invoice data then each invoice for the same Customer does not need to store the address of the customer, this is held in a Customer table and only a reference to the customer’s record in the Customer table is stored in the Invoice table.
This not only cuts down on the amount of data that needs to be stored, but also means that if the customer changes their address then you only need to change the address once, in the Customer table instead of for every record of theirs in the Invoice table.
Normalisation can also cut down on the number of columns that you would otherwise use in a spreadsheet.
Another use of normalization is to alleviate the need for creating endless columns in a table.
For a really good explanation of this goto Allen Browne’s website