Studies show that the average organization fails at least twice in its efforts to form an effective data governance program. But the changing landscape of regulation has put organizations in a position where compliance dictates that failure is not an option.

Like many organizations, when faced with the mandate to create an effective data governance program, Oregon Child Welfare had to sort through the myriad of existing, siloed data efforts to clearly define roles and responsibilities between technical and business collaborators and demonstrate near-immediate business value to an executive class who, although supportive, had other things to think about.

In order to accomplish these objectives, we determined that it was necessary to create a highly adaptable framework of data governance, which could grow and evolve over time.

The key initial features of our program included:

• Clearly defined roles and responsibilities – in an organization where tech had long been the de facto data governance entity, it's necessary to create a clear delineation between architecture, data quality efforts, and data governance, and what roles were needed for each;
• Development of a self-directed, bottom-up data governance structure – like many organizations, executive churn threatens the survivability of agency-wide initiatives at ODHS. Our data governance structure needed to be responsive to executive needs while simultaneously functioning in a quasi-independent capacity;
• Implementation of a Communications Plan and clear branding that targets education and communication to specific areas of the organization, and leverages end-user engagement to identify formal and informal data stewards;
• Showing business value as soon as possible – we selected key data elements that have been known issues and began the process of data governance there. We developed and reported the metrics around these elements, including actionable fixes/changes that support higher data quality, and monetized the difference between maintaining the status quo and investing in system development to produce higher data quality.