Data governance initiatives have a notorious reputation for being difficult to do, and for most of their history this reputation has been well-earned, but in today’s world there are few vital considerations that, when factored into your strategy, can transform data governance into a force for accelerating and scaling data-driven outcomes.

As you consider either launching or revisiting your data governance program, here are a few considerations to keep yours on the right track.

1. Our multi-cloud world makes data governance a top priority. Many companies are fully committing to cloud or reimagining how they want to manage their data in the cloud, and data governance needs to be a core part of that game plan. You could say that our multi-cloud world makes data governance a whole new ballgame, especially for large enterprises that may have built up decades of legacy systems and face needle-in-the-haystack problems whenever they try to manage data. Even if you can quickly access your data in the cloud, knowing whether it's accurate and updated or stale and misleading can be a tough problem. Data governance is essential to good data practice in these environments. 

2. There is no data democratization and no data sharing without data governance. Data democratization and sharing may be the holy grail for many enterprises, but sound data governance is its engine. People throughout your organization should be empowered to use data on a day-to-day basis for transformation, but if the business doesn’t have trust in that data you’ll be hard-pressed to achieve your full potential for any data and analytics initiative. 

3. AI is a core component of delivering data governance at scale. We are well past the point where the scale of data can be managed by the brute force of humans, or even in human time. That’s why bringing the techniques of AI and machine learning into data management is a powerful strategy. If a machine can do part of the work of figuring out what data fits together, for example, master data management moves a lot faster. Imagine being up and running on a master data management initiative and seeing the first waves of output in 30 days. AI-powered data management tools also can step in to both filter and distill options and learn over time as data analysts get feedback from users on what’s relevant, what works, and what doesn’t.

Where this approach really starts to take off is using AI to help map your data management policies (about customer data or supplier information, for example) to your business view of the data. AI allows you to locate and link business semantic elements across multiple systems that share these elements, so as you move or deliver data you can automatically bring along the policies that accompany it.  

4. Every technology strategy needs a people strategy. The right technology foundations are essential, but technology alone can’t solve data governance. In addition to needing the appropriate technical foundation, you also need to be focused on the people and process. Your data technology has to serve each set of stakeholders in the data governance process in a way that makes sense for how they need to interact. The data governance framework for a data steward will look different for a data engineer and different again for the average business user. Of course, successful change management also requires actively engaging people with the technology versus laying out ultimatums. 


Solving the last mile problem
As you consider all these factors, it’s crucial to keep the foundation you’re building on in mind. You want something that empowers multiple business units to have data impact on a daily basis. You want to hone people’s instincts that the first action they take when they’re making decisions is to go look at data, versus consulting primal knowledge or gut feel. This gets more important every day as consumer-side demands for instantly relevant information spill over into enterprise response time expectations. 

The reality is, if you can deliver a sound, consolidated data management foundation, overseen by a governance organization where people are truly responsible for different domains of data, you can create a working data marketplace, where users are granted access to the right data at the right time for the right reasons. If your data is reasonably easy to access and subject to the right governance workflows, your employees will adopt and scale with it. If you can’t, they are likely to go rogue and create more data silos. And shadow IT these days is probably the last thing you want.