The main challenge with being information driven is the governance of the data. Let’s start with a simple example that illustrates this point. You have contact data in several systems in your enterprise. Consider three groups, sales, marketing and support, all who are constantly updating customer data. Can anyone update address? If an address is updated by two systems, which one takes precedence? Is it simply a timestamp or the role of the person who did the update? What happens if a customer updates their address on a portal, should they take priority? What happens if they use a home address and not a business address? Who can update the contact name? What name do you use to address the contact? How do you define what a unique contact is? Is it based on email address?

Many of these mundane questions are extremely difficult to get consensus in an enterprise. They tend to be political issues and power struggles within organizations. And to think, these questions just scratch the surface. The questions you need to answer run deep and wide. The point of data governance (DG) is to bring sanity to your data and create a structure for consistency and agreement for your data. Ultimately, data governance must be addressed before you can become information driven. As a close friend once told me, technology issues are easy, people issues are hard. This is the hard people issue about data.

Sponsorship and Ownership
I’ve had several discussions with groups on what it takes to be a successful CIO. One of the primary skills required is “selling.” It’s the job of a CIO to influence the appropriate senior executives of the need for DG. It would be extremely rare for an executive outside of IT to recommend the need for DG. With the example questions above, it’s the CIO having the conversations with the people who run sales, marketing and support.

The goal of these conversations is to get one senior resource committed from each group who is ultimately responsible for the data in that function, i.e. the sponsor. This should be, at a minimum, a VP or director from sales, marketing and support. To be most effective, sponsorship must be included in people’s job description and compensation metrics.

Operating Model
Within IT, I suggest your data architect run this group, provided this person has strong facilitation skills. Otherwise, the data architect must be augmented by a resource who can manage facilitation. In essence, a consortium type of model is being followed. IT doesn’t own the data, but facilitates the ownership. The business sponsors own the data and need help from IT to know what questions to answer. This model requires consistent, ongoing meetings and participation from all members of the DG group.

There is no prescriptive model for the frequency and length of meetings. This will all depend on the unique needs of your situation.

Objectives
The objective of this team is to agree on accountability (ownership), the processes, policies, standards, organization, and technologies required to manage the information.

Definitions:
Processes – These are the rules of how systems will update each other and the human processes that will be followed to monitor the quality of the data and address issues with the data when they arise.
Policies – These are the commitments each team makes in order to keep the information viable, e.g. all data that errors will be resolved within 24 hours.
Standards – This is agreement on topics such as naming standards, data definitions (dictionary) and formats that everyone will follow.
Technologies – The business team does not define the technology to be used, but the requirements of the technology.

Metrics
Key to any data governance program is defining the metrics of success. To do this, the data in question must be profiled (baselined) for availability, accessibility, quality, consistency, auditability, and security. The team must determine acceptable measures for each of these metrics and manage the program to the metrics. It is critical that metrics are tracked on an ongoing basis to ensure the solution remains vibrant and does not atrophy.

This tracking can be accomplished in a straightforward manner, as simple dashboards can be created to show the data against the metrics that have been set.

In the end, you can read books that detail elaborate models for data governance. As a practitioner, I suggest taking a pragmatic approach that will work within your organization, as opposed to implementing a “boil the ocean” type of initiative which rarely works.