Getting All the Cow Folk to Giddy Up in the Same Direction

A conversation with Kim Weiss, Chief Data Officer of North Dakota

For those who love to read, this blog post summarizes episode 50 of the Data Culture Podcast; that being said, we definitely recommend listening to the episode!

We Built It. They Didn’t Come.

North Dakota’s data journey didn’t start recently. When Governor Doug Burgum was elected in 2016, his technology-forward agenda created real excitement for those working in IT across the state. The early vision took the form of a unified data platform, and when the pandemic hit, all the ingredients seemed to align: urgent need, prioritization, funding, and momentum to implement a state-wide data platform.

So they built it. And then they waited for people to come.

They didn’t. Not in the way anyone hoped. Kim Weiss, North Dakota’s Chief Data Officer, is candid about what went wrong. The vision was too centered on technology, on the solution. It was lacking the people, the process, and critically, it missed addressing culture and habits. There was awareness that those things mattered, but the urgency of the pandemic pushed the team toward shipping the platform rather than building the cultural foundation underneath it.

The result was a data lake that existed but didn’t stick. Instead of agencies flocking to use it, the team found themselves pulling and convincing. The momentum that the pandemic had created faded without the cultural infrastructure to sustain it.

Swipe One, Swipe Two

Kim frames North Dakota’s journey in swipes, iterations rather than clean breaks. Swipe one was the technology-first approach. It made progress, but it couldn’t hold. Swipe two, which began in earnest around 2023, asked a different question: where and why did we fall off the path, and what do we need to do differently?

By 2023, more of the ingredients were in place and the right sponsorship existed across multiple fronts, not just technology. Data literacy had started showing up in conversations across agencies, partly because people had lived through the pandemic and seen firsthand the struggles that come with poorly managed data. The awareness was on people’s lips in a way it hadn’t been before.

The formal vision they landed on: to harness the value of data to help North Dakota thrive. But unlike swipe one, the strategy behind that vision was built on four distinct work streams.

The first is data governance and stewardship, instituting a statewide framework and helping agencies implement their own localized programs. The second is infrastructure and architecture, and the third is data literacy, which focused on building the muscle and thought leadership agencies need to think about data differently. The fourth was unique and new: service offerings, or to the point, how NDIT continued to develop and enhance its product and service delivery to agencies. This last one was an important pivot from where NDIT was before as it pushed IT to think about adapting themselves to what their customers, the agencies, needed and continually adapt to those needs.

Governance and literacy are the two that get at culture most directly, because they represent the biggest mindset shift: moving from treating data as a byproduct of transactions to treating it as a strategic asset.

Change Management: The Missing Ingredient

Kim used to think, and she’s honest about this, that a sound vision, a sound goal, and a plan were really all you needed. Clearly articulate improved outcomes to stakeholders and success would naturally follow.

It doesn’t, at least, not without managing the change.

Swipe one had zero change management built in. There was no strategy, no strategic plan around data usage and analytics adoption at all, rather a vision and a technology investment and an assumption that it was enough to move forward. Swipe two made change management a priority from the very beginning. When NDIT went out with an RFP for a vendor partner to support the effort, change management was written into the requirements. When they started building deliverables, change management became baked into everything they did.

One of the first recommendations from the vendor partner was to invest in a strategic communication and change management practitioner before anything else. The logic: if you want this to be successful, you have to get out front, manage the change, communicate to stakeholders, and keep them actively engaged. Without that, you can build things, but even the most eager champions have day jobs and current modes of working that must be reminded and nudged in a new direction.

Kim hasn’t been able to resource it exactly as she’d like as government is lean by design, but it’s a huge priority. NDIT’s communications team has been pulled in early and the support has been significant.

Balancing Foundation with Visible Value

One of the tensions Kim navigates constantly is the balance between laying groundwork and delivering tangible value. You can build policies, processes, a beautiful plan, an implementation roadmap, but without some visible, immediate value to agencies, you lose momentum before you’re even out of the gate.

Kim and NDIT embedded PoCs throughout the process, making sure agency partners could quickly see data assets, analytics, and other ideas come to life quickly and frequently. There’s no one secret sauce to picking the right agency, the right starting point, or pilot initiative, Kim says, but they’re making every effort to avoid the trap of building everything before anyone can do anything.

The governance subject is a perfect example of this tension between iterative progress and the ‘build everything’ mentality. Data governance is a loaded term with wildly varied expectations and definitions, so when agencies hear it, their reaction is often some combination of overwhelmed and bored. It can be akin to public safety officers monitoring and mandating what you can’t do, versus an enabler of better work. Rather than trying to build out an entire governance framework before letting anyone touch it, they started with something tangible: a tactical data governance playbook.

The playbook meets agencies where they are. If you don’t have anything in place right now, step one: do you even know what you have? Do you have an inventory of your data assets? Do you have a point of contact for each one who is the decision maker? It’s lightweight, it’s practical, and it gives people something to wrap their arms around rather than asking them to wait while the full framework gets built.

Perfect Gets in the Way of Progress

This is a theme Kim returns to throughout the conversation. As a central IT organization providing unified services to cabinet agencies, NDIT needs critical stakeholder input. But balancing the voice of the customer with forward progress, and moving beyond analysis paralysis, is genuinely challenging.

The collaboration that came with swipe two has been one of the biggest early wins. Connecting people who felt isolated before, or who didn’t know that others existed or that certain services were available, has yielded value that goes well beyond the data strategy itself. But collaboration also means more voices, more input, more opportunities to stall.

Kim’s approach: aim small, miss small. Implement early, test early, and if you fail, shift it into a win. The ability to iterate quickly changes the psychology of the whole effort. When people know the next batch of changes can come soon, they stop treating every request like their one annual trip to the store. Agencies no longer had to craft the perfect request, guess at what might be needed in a far future, and pivot to iterative delivery and progressive discovery.

The Legislature Is a Stakeholder Too

One of Kim’s sharpest lessons learned, and one that people coming from outside government often miss entirely, is that the legislative body is a critical stakeholder in the data journey. What NDIT is able to do and the speed at which they can do it hinges on the support and investment they receive. That support comes from legislators who have completely different concerns and contexts than the people doing the work day to day.

Sid compares it to pitching investors who aren’t in your space. You have to present wins in a way that resonates with an audience that doesn’t live your daily life, and you have to bring them along the journey, not just show up at budget time with a request.

This was another area where swipe one fell short. The importance of communicating progress to the legislature, in their language, on their terms, wasn’t prioritized early enough. Swipe two has that awareness built in. Communicating wins, even small ones, to the right audiences in the right way is part of the change management practice now.

Slowing Down to Speed Up

Kim acknowledges that government is still learning what change management really means in practice. The formal concept is still relatively new in state government, and when timelines and budgets get tight, change management is one of the first things to get cut. That pattern reinforces the problem: the muscle never gets built because it keeps getting deprioritized.

But the underlying insight is both simple and hard. Data professionals can see why something should be quicker, why it should be easy, but for agency partners whose core job isn’t data, this is genuinely new. It’s a different way of thinking about information they’ve always treated as transactional. As Kim puts it, sometimes you have to slow down to speed up. When you create awareness, bring people along, and help them become more data literate, things can really catch fire. But you can’t flip a switch and expect it to feel natural.

The next chapter for North Dakota involves continuing the execution of a strategy that spans 20-plus prioritized projects over three to five years, maintaining legislative support for that rollout, and beginning to figure out where AI and emerging technology fit into the picture. Getting the data AI-ready loops right back into everything they’re already working on: governance, literacy, culture, and the ability to introduce and sustain change.

The vision looks clearer now than it did five years ago, but not because it was right from the start, rather it’s that each iteration taught them what the vision actually needed to include.

This blog post is based on a conversation between Sid Atkinson, Lee Harper, and Kim Weiss on the Data Culture Podcast. Kim Weiss is the Chief Data Officer of North Dakota. Connect with her on LinkedIn.