Putting the prefix meta in front of a word means essentially that you are abstracting information in some way. A meta-analysis is an analysis of a set of analyses. Metadata is a term used to denote the description of data—data about data. This often means the data's structure, its column names essentially, but it can mean any abstraction of the data, such as purpose, date of collection, source, or standard.

Metadata is the key to bridging the gulf between the raw data sitting in databases and data warehouses in what are necessarily complex data structures optimized for performance and the everyday user of this data, irrespective of how familiar the user is with the data content. Metadata should be seen as an essential communication protocol for data analysis, particularly where a self-service approach is sought. It can be a common language through which user requirements can be expressed and understood by BI consumers, developers, and applications.