Successful introduction of data governance is really key to making the overall project successful. The activity of data governance in practical terms is really not new to most organizations but the word is new.
Most organizations have created discipline around the rules governing who adds and changes data and data definitions in a single system. An order entry system has all kinds of rules and discipline about who is authorized to make changes in the data definitions and rules. Knowledge about how these disciplines work for these systems is key to making data governance work for the data warehouse. The difference with a data warehouse is that the data provided comes from multiple systems. The responsibilities must be created out of the existing organization.
Before we add complexity of conflicts to the data definitions, a great first step would be to find out who are the data stewards for the systems feed the warehouse. In many cases this responsibility has been defined and that person will likely be the identified data steward. I think an important first step would be to identify these people in a data governance data base.
Let us take an example from health care, wait time is big issue and the definition has provoked much discussion. The important first step is to identify who are the data stewards for the existing tracking systems for the various systems tracking wait time. At that point, we then can initiate some discussion with this group about rules and definitions. Without this dialog with people with knowledge and responsibility, no progress can be made. Often we get distracted by the tools we can use and spend too little time working with the key data stewards.
In the context of a longer term development of data governance, we need to be very careful to build on what exists within the current organization and acknowledging the roles of key people. What works in one organization will not work for another. We need to carefully design a process that fits with each organization.