Hi community! We are in the process of establishing a more formal data governance program and one question that came up is which team is the best to own and drive such an initiative. We were debating where this may sit in companies like LinkedIn. Is this driven by an analytics team, product, engineering, finance?

At the company I am at the engineering team introduced the idea to help tackle technical debt (duplicated code/pipelines, sold as save money and time). Then the product analytics team who curate metadata the most jumped at it the most to get rid of their other tools once they saw the value. More can be read about that here: https://blog.datahubproject.io/datahub-to-kickstart-spring-data-cleaning-cfd7fe28e6a3 I think like starting a fire you need a few embers started but need others to pitch in and grab the rest of the wood to really get it going. Pitch it well and find key allies in other teams that align with the new vision. Just saying hey we found this tool, could you please completely figure it out will be too daunting of an ask if they weren't the original team with the idea.

What do you want to achieve or solve through this? Largest challenges with governance tend to be how much coordination it requires and how easily "governance" becomes ceremonial and detached from actual business outcomes. Whatever you do, I would avoid creating a separate or independent governance team, and rather keep it more directly connected to business operations, like product or analytics. And finding people to run it who are motivated to help others do more with data and are great communicators probably matters more than whether they are analytics/product/engineering.

Super helpful! Thanks both! I have pitched the idea of a tool-aided (left it open which one) solution for data governance and discovery but left open which tool for now. I wanted to test the waters with various business units and initial feedback is actually very positive. It turned out that analytics already had this on their strategy radar. So I now have allies in legal, engineering, product, and analytics. They all have different objectives and I need to come up with a more refined business case to get enough buy-in and not make this die a quick death 🙂

That's a great approach! I often think that if the initial reaction is negative, it's almost better not to even try yet, because it'll be an uphill battle and maybe not the highest value thing you can be doing at the time. But if early reception is positive, then it can provide a lot of easy wins 🙂

While some companies prefer to keep data governance “more directly connected to business operations”, having a dedicated data governance team can offer many benefits, especially in large companies. Here are some reasons why: • Expertise: A dedicated data governance team has a deep understanding of data governance principles and best practices. This expertise can provide guidance to the rest of the organisation on how to manage data effectively. • Consistency: A dedicated data governance team can ensure that data governance policies and processes are consistent across the organisation. This minimises confusion and ensures that everyone is following the same guidelines. • Accountability: A dedicated data governance team can hold individuals and teams accountable for following data governance policies and processes. This promotes responsible and compliant management of data. • Scalability: As organisations grow and data becomes more complex, a dedicated data governance team can help manage this complexity and ensure that data is used responsibly and compliantly. While I must disclose that I am part of the dedicated data governance team, and therefore my opinion may be biased, I strongly believe that defining processes and standards across a large organisation would be challenging without such a team. Of course, in smaller organisations with fewer employees, this may not be an issue 🙂 regarding ownership: The data and all related assets should not be owned by the data governance team, but rather driven and controlled by the relevant business units. The data governance team can provide guidance, establish policies and procedures, and ensure compliance with data-related regulations and best practices, but ultimate responsibility for the data should rest with the business units that create it (data owners) or use it (data consumers)

At LinkedIn, we had a large Data team that rolled up to Engineering and the VP of the Data Org was also the CDO. Governance wasn’t set up as a separate team, but was baked into the functions of the individual teams through “horizontal initiatives” that were sponsored by execs and was part of the engineering DNA (set aside 20% bandwidth for horizontal initiatives) established pretty early. GDPR itself was unique in that it was run like a separate program reporting status daily to the CDO (with sponsorship from the privacy legal team, and head of security).

From my experience (working with various companies in Europe as a consultant) there is one big mistake to avoid: As soon as Data Governance is seen mainly as an IT project you have a major problem. You need support from the business units because they have to provide the data documentation, glossar, tag the data, etc. On the other hand, business units usually don't have the resources (or at least they tell you they don't have) to do additional tasks beside the daily business. Therefore you need some kind of dedicated team to get things moving at the beginning. As soon as business see the value and start adopting governance (usually takes 12-18 month, sometimes longer) you can start scaling down the central team. But also in the long run there should be some central entity who coordinate and keeps track of the activities of the data owners and data stewards.