This article originally appeared on the BeyeNETWORK.
Today, data governance is all the rage. With Sarbanes-Oxley and Basel II, corporations must face the fact that corporate data is no longer a luxury, but a necessity. It is ironic that the corporations that had embraced the concept of an enterprise data warehouse (EDW) had the least difficulty in adapting to the new regulatory requirements for data. Conversely, organizations that had rejected and obfuscated the EDW for years had (and are still having) the most problems with Sarbanes-Oxley and Basel II.
It is with this background that the need for data governance has been recognized.
There are many facets to data governance. One facet is the day-to-day administration of the data. This includes the tasks of making sure data is loaded properly, making sure that the mechanisms for ensuring transaction and data update are in place and are followed, making sure that data is accessible when it needs to be accessible, making sure that data is available during the promised hours of availability, and so forth.
These activities are necessary to the proper usage and operation of the database environment. For years, these activities have been tended to by the database administrator. But with governance comes another awareness. That awareness is that there is another dimension to the tending and care of data outside of the duties of the database administrator.
This set of duties belongs to the data steward. The data steward is the person that the organization goes to if a value of data is found to be incorrect. In addition, it is the data steward that defines the source of data and the transformation algorithms for the placement of the data in the data warehouse. The data administrator takes care of the physical form of the data. The data steward takes care of the actual content of the data.
How is the data steward chosen? As a rule, the data steward comes from the user community, not the IT community. Only occasionally is the data steward an application user, rather than a business user. The selection of the data steward has several interesting aspects. One of those aspects is power. The data steward has the power to decide how data is to be used and interpreted all over the organization. In that regard, the data steward is a representative of the entire organization, not just an application.
But there is another side to the job of the data steward. And that side is to be a firefighter who is on call every time there is a problem with the data. There are then both privileges and responsibilities that come from being a data steward.
Rarely is there one data steward for the corporation. A large complex organization will have lots of stewards, each managing a different part of the corporate data.
Another interesting question is: Who should be selected in the organization as a data steward? A good rule of thumb for selecting the data steward is to find someone who feels pain if a value of data is incorrect. Find the person or organization that suffers whenever a unit of data is recorded incorrectly.
This leads to another interesting question. That question is – who should be the steward of data if no one feels any pain when the data is incorrect. From a philosophical standpoint, perhaps the corporation shouldn’t even have the data if no one feels any pain.
Unfortunately, the modern corporation probably has so much mundane data that if the data is incorrect, no one has a problem.