Data Warehouse Ownership
Who is the owner of Data Warehouse? It is very important issue but hard to answer. Here are several examples from the practice of how to apply Data Warehouse ownership.
First, there can be two levels of data ownership: production data level and cube/theme data level. Ownership on production data level is clear to define because users who want data from productions systems have to ask production system owner for permission and access rights. Production system owner allows data cube, data theme or report initiator to use data from production. Data organized in new environment like cube or theme mean new information. It is so because situation is more complicated, cubes or themes are made from usually made from several production sources and data is worthless without good context. Who is the owner of the cube? It is the person who initiated cube building, request who started building of the cube. Owner of cube or theme is second level owner. Anybody who would like to use data from reporting cube, reports or themes must ask second level data owner for access rights. Owner of the cube asks for data access from source systems but potential users of the cube ask cube owner.
Making fragmented ownership by dimensions and measures ownership is too complicated and messes complete procedure.
Who is Data Warehouse owner? If the ownerships has to belong to only one department it is a little bit risky decision. For example if IT is owner, they are just service providers without in deep knowledge what to do exactly with content. Perhaps better variant is to give ownership to strategy or to finance, to be more precise to controlling. Reason for this decision is general view over definitions. Definitions are necessary for solid external and management reporting. Controlling knows best what kind of products will be made upon Data Warehouse for official reporting. If this solution is still not adequate perhaps best solution would be to establish committee of experts from all business areas with lead from controlling and to give ownership to this body.
Without ownership there will be no strategic development of Data Warehouse system. It will only run for coverage of current needs without any pro active approach. Data quality management will be very doubtful and at the end there will be no one to decide how to handle vast quantities of needed actions.
There is a strong need to declare officially Data Warehouse ownership and to establish data security procedures as soon as possible.