Proving Impact and ROI for Data Governance

Getting business leaders on board with data governance is easier with the right plans and metrics in place to tell the story about its impact and ROI.

This article originally published in the Data Leadership Collaborative.

One of the reasons that data governance may intimidate business leaders, data analysts, and even Chief Data Officers is that the definition often includes the complete suite of capabilities needed to manage data well. Governance is used to apply to metadata, data literacy, data lineage, data security, data quality, master data, and other topics. To me, the broader term for all these activities should be data management. Data governance as a practice is more about setting the standards and policies for good data use and access: how it’s gathered, stored, processed, and disposed of. That’s a lot less intimidating.  

Another reason data governance puts people off is that they hear the term and think overhead, bureaucracy platforms, councils, committees, establishing metrics for data governance and ROI, and other things that will slow them down because they’re add-ons to what they need to do.

In fact, data governance should be embedded into how you operate as a company. If you’re putting data management practices into place properly, governance should just be part of how you run your business. The point is, when it’s not baked into how you run your company and pull information to do so, data governance does feel like a burden.

Data governance is a necessity

I’ve spoken to executives who have told me they’re glad they don’t have to do “the governance thing” anymore and can move into cooler data-related areas like AI. But my point—and the reason I’m publishing a book on data governance—is that if you don’t do data governance up front, you can’t do the cool things like data science, advanced analytics, customer experience work, large language models, and data automation projects well. 

Instead, you end up taking your data scientists and prompt engineers off of their projects, which forces them to spend time preparing and cleansing data, measuring and improving data quality, and joining disparate data sets. This is why data governance too often feels like hygiene. It’s really work that should be done by the governance team. No, it’s not all pleasant, but then you can’t eat ice cream all day and expect to enjoy good health. Governance is like eating your broccoli.

Getting your data governance bearings

If you’re onboarding into a new data governance position, it’s important to identify the most burning problem you could easily fix now as well as the consistent themes you need to start investing in now to fix later. That’s a balance you need to strike to prove your impact. 

One common problem is inconsistent data governance metrics at the executive level. Sales shows up to an earnings meeting with their own numbers and says, “I’ve got this,” but Marketing says, “No, I contributed X number to that pipeline and I want credit for it.” One thing you can do as a data leader is to play referee. Volunteer to dissect the metrics for data governance, look at the calculations and the sourcing, dig into the data lineage, and assess the quality. In most cases, you’re likely to realize that neither Sales nor Marketing was correct. However, this can be a great place to start because you’re showing the value you can add in some very common situations. 

A common longer-term problem is around consistent customer identification. Solving this may take a combination of new architecture, customer master data management, good lineage and quality, along with governance to make it right. As crazy as it sounds, I was once at an organization where it took 18 months to agree on what a customer was. My point is, weaving data governance into meaningful business problems and use cases is the way to show value, versus saying “I’m going to fix master data” and disappearing for two years.

When it comes to training staff on how to deal with data and be good shepherds of governance, creativity is important. For example, how can you effectively incorporate data education into other types of training? If you’re a bank, how do you incorporate data into credit risk training, or operations training, or data engineering training? This approach will always go over better than announcing that you’re going to do data literacy training, which can feel both insulting (executives don’t like to feel illiterate) and like an experience that is being foisted upon your colleagues.

Baking governance into the natural ethos of the organization is what yields a data culture; that is, a culture where information is used to make better decisions that drive impact.  

The data governance ROI question

Every leader, regardless of discipline, should have a way to determine how they have positively influenced the business. ROI is a way to do that, and data governance leaders struggle with it. In part that’s because they often don’t understand the business well enough.

To be fair, we can’t fault business leadership if we want to celebrate the fact that we’ve mastered 2 million customer records but their response is essentially, “So what?” Laddering down from our achievement, we need to ask ourselves questions about the ROI of our data governance program such as:

  • What did this work do for the business to improve user experience?
  • How did this translate into sales? 
  • How did this empower the sales teams to make better decisions? 
  • How did this help revenue operations to process orders better? 
  • What happened as a result of our work that the business can feel and measure? 

Too often, we stop at the 2 million master records or the data metric and don’t actually take it to the business side. Ultimately that’s who needs to say, “Yes, I experienced this value thanks to what the data team did for me.” Having the right data governance metrics in place can help you develop a data-driven story of your successes. The difference between data leaders that succeed and fail is often because of this ability to tell a value story about their impact.

In information we trust

There is a lot of work we still need to do as data leaders to position data governance as an enablement function. I would argue we need to be very practical about how we do this, because the outcome is trust: helping our companies trust how we use information, how we’re embedding it into the way we run our organizations, and that we can deliver it consistently. 

During the process of writing my book, I often thought back to how hard it was to find good information to help me be successful when I was starting out. I made a lot of mistakes in that journey. That’s why it was important to me to take some of those learnings and give them back to the next generation of leaders coming up behind me, to help them find their way more easily.  

Now that I’m in my second tour of duty as a CDO, always with a hand on the tiller of data governance, I would recommend a few best practices for a new data leader: 

Remember that every situation is unique.

You can’t come in with an attitude that an approach that worked well in your last job can simply be laid over a new situation. No two challenges are ever the same, especially when it comes to data governance.

Stay curious.

Think about the problem that the stakeholder is experiencing and ask them probing questions. Why is X or Y the case? What led the organization to this moment? What have you tried and what haven’t you tried? 

  • Remember that every situation is unique. You can’t come in with an attitude that an approach that worked well in your last job can simply be laid over a new situation. No two challenges are ever the same, especially when it comes to data governance.
  • Stay curious. Think about the problem that the stakeholder is experiencing and ask them probing questions. Why is X or Y the case? What led the organization to this moment? What have you tried and what haven’t you tried? 
  • Don’t try to be the smartest person in every room. Always assume that the people around you know a lot of things that you don’t.
  • Listen deeply. One of the best ways a data professional can be successful is to listen to the people that they’re trying to serve – namely, the business. Remember, it’s our job to make things better for those around us. 
  • Work as a partner. Work with your stakeholders to co-create a solution. If you try to do all of this on your own, I don’t think it’s possible to be successful. You’ll also build more trust by collaborating to weave data governance into business processes rather than imposing it.

As a final note, consider that data governance can be much more than applying controls and churning out metadata. Done well, with the right perspective on outcomes and ROI, governance is really helping to design the future of your organization alongside your colleagues. With the velocity business moves at today, having the right information to make the company run better and create impact in the market will always draw allies to your side.