I’m often asked how many data stewards a company needs.

A rule of thumb I’ve used is “one data steward per $1 billion in annual revenue”, but of course that can’t be a hard and fast rule.

Depending on the degree of complexity in the company, the degree of automation, what data quality tools are available, and the major processes around the enterprise, that number is going to vary.

Here are some thoughts to keep in mind as you build your data governance organization:

1. Build the business case and secure your funding. Don’t forget this is a multi-year effort. Several times, I’ve seen data governance teams and MDM initiatives that never thought beyond their first year’s funding, and who were surprised when funding for Year 2 and beyond wasn’t forthcoming.
2. You’ll need a multi-layer, cross-functional approach. Start with a high level group that sets data policy, usually called a Data Governance Steering Committee. Your executive sponsor(s) should drive that. Then you’ll probably a Data Governance Office to act as “the glue” driving your program, and have a group of data stewards which handles day-to-day data management, which reports to the Data Governance Office.
3. Focus on your data governance processes, including relevant metrics that show the impact to the business, and on communicating those metrics and their impact to the rest of the enterprise. Create a communications and education strategy to get the word out.
4. Look for IT people who “speak business” and business people who “speak IT”. These liaison-type people are going to be crucial to your data governance organization. Although I believe that data governance should be driven by the business, there’s no question that you’re going to need a lot of support from IT. So look for the “bi-lingual” folks in your company – they’re out there and they can be incredibly valuable in understanding the business impact of data issues, and in providing the technology support that the data stewards will need.
5. Don’t forget about metadata. In your travels as a data governance organization, you’re going to run into a lot of metadata (information about information). Develop a disciplined approach to cataloging and maintaining the metadata you come across. In many ways, that will probably be one of the most useful outputs of your data governance initiative.

We’ll have more to say on data governance in this space in the future, but let me leave you with one idea. Instead of describing your MDM initiative as (hopefully) including data governance, think of it as a data governance initiative that has an MDM component.

You’ll be doing data governance in your enterprise long after your current MDM hub implementation is over.