In developing a data governance structure from the ground up, what are the first steps you would take?
Sort by:
I would widen the scope to D&A governance. Many organizations work on data governance, but Analytics Tools are a major consumer of data. Forgetting Analytics governance leads to issues. Gartner publishes on D&A governance, but unfortunately, most organizations spell it with a capital D and an "a" in a very small font.
Analytics governance starts with zero-touch version control, creating a metadata catalog of all analytics assets, and then building monitoring, reporting, and alerting.
Starting with the correct model is essential. You can read my blog on the comments I made to the Gartner Franchising Governance model for Analytics here: https://www.linkedin.com/pulse/governing-power-bi-why-gartners-latest-recommendation-edwin-6aqqc/?trackingId=6nGPRrvXSHCzs2Zstkh9XQ%3D%3D
Look for previous examples that can be learned from and built on top of. Organisations like Microsoft and McKinsey and others have examples in the public domain.
Great thoughts from Chris, Yvonne, and Rebecca. I would add that placing a strong focus on educating people across the enterprise on the importance of data, collaboration on data standards and practices, and data governance is critical. Taking the time to help everyone learn and align on both the vision and practices will reduce friction as you implement governance and encourage everyone to take a more active role in supporting the governance activities.
Some may assume everyone understands the importance of data, and often it is clear to individuals what specific data is critical to their role or function but this perspective may include limiting assumptions from a more siloed perspective. In today's settings many types of data are useful for both their initial intended purpose and for additional value generating areas. Think, for example, about how data generated from a logistics operating team for creating more efficient routes is also valued by the customer when it powers features that allow them to track delivery status. A great data strategy and data governance program can help you better plan for the future while protecting your business today.
Introducing data governance in a broad conversation and going beyond a declared mandate or a small group of data focused functions can help you set the stage for lasting success. Best wishes!
- Gain leadership support, sponsorship, and commitment to review results regularly e.g. monthly
- Identify the key data attributes you are measuring and holding others accountable for
- Ensure every key data attribute has an owner, someone you can hold to account
- Ensure you have the reporting mechanisms in place to regularly and accurately monitor data accuracy
- Ensure there are clear 'consequences' for data inaccuracy including visibility to the leadership sponsors
If you don't have leadership support & sponsorship, you don't know what you're measuring, you can't measure it, and there are no 'consequences' for inaccuracy, no governance structure is going to work.
Well said Chris. I think your point about ensuring that every key data attribute has an accountable owner is key for data accuracy. It's easy to compile data but I've found that the complexities & areas for confusion pop up down the line and without being able to understand the historical context or to confirm accuracy you can't maintain a clean environment.
Executive support and begin with a MVP or pilot approach to gain some wins.