4 practical lessons from data governance leaders at Dropbox, General Motors, and Patagonia

“We have to slow down a bit in order to speed up.”

May 30, 2024

I think anyone working in data today would agree that governance is tough. I talked recently about why it fails and my thoughts about how to fix it, but I’m far from the only person thinking about this. That’s why I’m excited to open up today’s newsletter to three new voices who are leading the discussion about building better data governance.

At our recent Re:Govern conference, my co-author Austin Kronz hosted a fascinating conversation with three amazing data governance leaders: Cortney Worthy, Data Governance Lead at Dropbox; Sherri Adame, Data Governance Lead at General Motors; and Lisa Harrison, Director of Data Products and Enablement at Patagonia.

Today’s newsletter features just a few of the countless insights and learnings from this session. Keep reading to learn four invaluable lessons on modern data governance and practical strategies that anyone can use to excel in their governance journey!

Watch the recording

✨ 4 practical lessons from real data governance leaders

1. Don’t move forward without thorough discovery

“I didn't want to make data governance priority decisions in a vacuum. I wanted to be able to understand fully what I was working with and the culture that I was in.” – Lisa Harrison

Getting started is often one of the hardest parts of data governance. Rarely is a company starting from zero — instead, when it decides to improve its governance processes, the data landscape is littered with outdated or incomplete frameworks, programs, and processes across different departments. As Sherri put it, “At every large organization, there's been an attempt at governance. There are pockets of teams that do it really well, but it's really hard to thread it across an enterprise.”

This is why it’s important for any data governance program to start with thorough discovery. For Lisa, this involved nearly 50 one-to-one conversations with folks across the organization, trying to understand Patagonia’s existing pain points, expectations, and processes. These learnings will set the foundation for a tailored governance strategy that aligns with your company’s specific needs and culture.

Here are some of the things that Cortney, Lisa, and Sherri did as part of their discovery processes:

Meet with leadership to set the right expectations up front
Engage with stakeholders to learn what’s working and what isn’t
Collaborate with data platform and observability teams to align on technical aspects
Map out areas of opportunity and assess their feasibility and impact

2. Pivot, rather than persuade, to find quick wins

“When you start from scratch, it's pretty much an open landscape. I wanted to see places where there was high feasibility and high impact… We wanted to capitalize on where there was appetite and zeal for the work.” – Lisa Harrison

Often we think of data governance as a linear process — first do Step One, then Step Two, and so on. If you are met with resistance at one of those steps, plow through until you succeed. However, all three of our data governance leaders talked about the value of finding the first steps, or quick wins, that are right for your company.

Lisa explained that this was a big learning in her career: “I definitely wasted a lot of time in the beginning feeling like it was my responsibility to push through resistance… Along the way, I picked up what I think is a beautiful tool, the ability to pivot. Sometimes that’s even faster — going back to the drawing board, seeing where else we can get some momentum around governance.” Starting with quick wins can help demonstrate the value of governance initiatives early on and build support for future efforts.

Lisa’s quick win at Patagonia was bringing together the people who were most excited about data governance into a cross-functional data protection committee. "This included folks from Legal and HR, folks from Cybersecurity, and folks who were dealing with the data on a day-to-day basis. That committee got together to help set some of the priorities and work on some of the data governance initiatives to start.” Creating this committee allowed her to capitalize on where there was already “appetite and zeal” for governance work, rather than trying to rope in uninterested stakeholders.

At Dropbox, Cortney’s quick win was developing a data certification framework. “We already had Atlan, so what we wanted to do was increase observability… We started working with our data teams to actually develop a certification framework, and that’s where we saw really quick wins. People were interested in the data results they were provided, but we were also able to get those most important assets at the top of the list in Atlan.”

Sherri’s quick win at General Motors also started with ensuring data quality. “I immediately engaged with the development teams… I said, I know in your development specifications, you're already specifying what that data is and what the appropriate metadata is for that. Otherwise, this feature or function won't work. So let's put that in a standard format so that, as you deliver, I can automatically collect metadata and eventually get it pumped into a data catalog.”

3. Lay a solid foundation before scaling

“We have to slow down a bit in order to speed up.” – Cortney Worthy

Turning quick wins into sustained momentum involves expanding early programs across the entire data lifecycle. For Sherri, it was about “taking that end-to-end process that we put in and tacking a little bit more on both ends”. After starting by governing the data that was captured in General Motors’ cloud infrastructure, Sherri focused on creating consistent governance from earlier data ingestion to later data consumption. This involved extending governance to data sources and automating data contracts for users.

Lisa added that an important part of building momentum is setting metrics for success up front. If everyone agrees with how you’ll measure success, you can use these benchmarks to drive future initiatives and gather support in the future. “You just pick up stakeholders along the way, you pick up goodwill along the way, and you're able to move a little bit faster with each successive initiative.”

It’s also important to set an expectation that progress will be gradual — a marathon rather than a race. Cortney emphasized the importance of starting with pilot projects and iterating based on feedback. "As you start doing those pilots, add in more competencies and more capabilities to the framework," she noted. This iterative approach allows organizations to refine their governance practices and expand their scope gradually, ensuring sustainable progress.

4. Build the right narrative — it might not focus on “governance”

“It’s not a revolution. It’s not governance coming in saying, ‘Hey, you have to govern this way’. It’s an evolution to a place where everything we’re doing with data is optimized and governed.” – Sherri Adame

Creating a compelling narrative around data governance is crucial for gaining buy-in from stakeholders. At Patagonia, Lisa recognized the importance of aligning governance with your organization’s ethos. "In the industry, there’s a lot of talk about risk and opportunity and those being drivers behind governance…. I realized that, as a mission-driven organization, there was a third kind of driver that was bubbling up — responsibility and ethics and doing the right thing by the planet and by our customers."

By framing governance as a responsibility that aligns with the company's broader mission and values, Lisa was able to build a stronger case for its importance. This narrative not only helped in securing support from stakeholders quicker but also in embedding governance into the organizational culture.

Cortney also highlighted the importance of framing governance as an enabler rather than a hindrance in your narrative. Governance has often used the traditional “stick” approach, which relies on enforcing rules through compliance and audits. She explained, “Years ago, it was ‘You have to do this’... It was more of using that stick to get people to do what you wanted them to do. And I’m sure we all failed in that regard.”

However, Cortney found more success with a “carrot” strategy, which focuses on finding incentives and support and narratives that will resonate with different data people. “The carrots won’t be the same, going to the finance team or the customer experience team. That’s where knowing your customer and being able to relay what benefit will resonate with them is very important.”

This shift helps to foster a collaborative environment where teams are more willing to engage with and uphold data governance standards, ultimately leading to better data quality and utilization.

Watch the recording

📚 More from my reading list

Switching costs by Sean Byrnes
Mapping the mind of a large language model by Anthropic
Is your data lifting you up or letting you down? by Elliott Stam
The ABCs of data products by Yordan Ivanov
What is it like to dislike data? by Michiko Wolcott
Treat AI news like a river, not a bucket by Alberto Romero
How the Guinness Brewery invented the most important statistical method in science by Jack Murtagh

Metadata Weekly

Discussion about this post