That perception stems from two things. First, many organizations truly haven’t done enough to prioritize data governance. It’s hard to get buy-in because data management is often seen as a cost center instead of a revenue center. It’s also hard work, with no easy shortcuts. Most organizations are behind where they want to be. However, the second reason is less avoidable — the governance landscape is evolving very quickly. Organizations that implemented governance systems even three years ago are probably now behind on state-of-the-art systems.
Using a Data Mesh to Enhance Insights
State CIOs also expressed some goals commonly adopted by other industries. In the survey, tech officials told NASCIO they are investing in data analytics by enhancing dashboards and meaningful reports (80 percent) and by making it easier to combine data from multiple sources (57 percent). These common themes drive much of the innovation around data management in all industries.
There are several conceptual movements in data management right now that are relevant when it comes to combining data from different domains and creating more meaningful reports and dashboards. One approach that has recently gained a lot of popularity is the data mesh: a cultural and organizational approach to data management that decentralizes ownership and responsibility for the development of data products among disparate domains.
Domain experts understand their data and their domain much more clearly than a centralized data management group would, so in a data mesh, creation ownership of many data products remains with the domain group. The centralized data management team becomes the hub of a federated system of data management instead of a monolithic central team in charge of all data development and management.
This approach is very applicable to government because of the significant expertise built up in different domains, such as health care and education. Centralized state management teams still need to have authority and governance over all data assets. This standardization in tooling and approaches is more efficient overall. But the responsibility for producing valuable data assets could often be distributed to the domain experts who know much more about the nuances in domains like healthcare or education than a central data team without domain expertise. This federated governance model can result in higher quality data products.
Another benefit of data mesh thinking and culture is that it promotes more collaboration across domains. As cross-domain data use cases come up, teams are encouraged to have conversations with subject matter experts in different domains instead of going back to a central data team that has no expertise. For example, someone in government might be looking at education data and wonder how educational outcomes correlate and overlap with public health initiatives. Analysts in the education domain can request access to health-care domain data products and they will look to the owners of that data to spur their ideas about meaningful collaborative insights driven by combining these data sets.
EXPLORE: What is data governance and how can it enhance data management?
Taking Small Steps for Data Governance
When tackling improving overall governance and enhancing overall value of analytics systems, there’s a lot to think about and many possible frameworks to adopt. In my experience, even the most forward-thinking organizations start small. It’s not only easier to start small, it’s usually strategically superior.
If a state wants to improve its data governance position, one path forward could involve taking inventory of all data assets and identifying a small subset of mission-critical data assets for a pilot program to monitor data quality. As ROI on a small initiative is achieved, this gives guidance for the overall data quality strategy, and it helps drive adoption as people see the wins. This same method can be used to start implementing the principles of a data mesh. Starting with one domain that’s building up data product capabilities, or even starting with giving ownership over a few specific data assets to the domain experts, can create some initial wins on the way to designing and implementing broader change.
I find it instructive to remember the Pareto principle, also known as the 80/20 rule. Almost universally, organizations find they can get most of the value needed out of data management initiatives — perhaps 80 percent — by strategically focusing on the most important 20 percent of the work. States don’t need to adopt every fancy data tool out there and don’t need to perfectly match all the latest frameworks that big tech firms adopt. They can start small and keep things simple by beginning with specific use cases that have clear ROI, and then expanding over time.
LEARN MORE: How is data literacy in government is enhancing data-driven decisions?
Supporting Data Literacy Across Government Agencies
If we take a step back and think about what governments are intended to do, we recognize inherently that governments need to understand their constituents and the people they are trying to serve. They need to understand how the programs they support affect people. This requires investment in data systems. So, it makes sense that data management is perennially one of the top areas to focus on.
Sometimes those outside of data organizations don’t see the value that data teams create. This problem can be exacerbated by territorial behavior. If data teams try to position their knowledge as unique, it ultimately hampers their ability to grow more valuable inside their organization. In fact, the more people understand data and the value it drives, the easier it is for a state or city to adopt data initiatives.
In addition to helping with adoption, employees outside of data management often are the best business owners for data assets because they understand the domain better than data professionals. For that to happen, there must be a groundswell of understanding within an organization that data is going to help move everything forward. Otherwise, it is just seen as extra work that the data team should do.
It is important to support data literacy among all government employees, not just data teams. Data literacy brings with it an understanding of the data management goals for government agencies and awareness of the potential impact. But how does that happen? Data literacy often starts with a skilled data storyteller who shares how impactful data can be in the decisions and programs of an organization. Presentations showcasing the role of data in outcomes make a big difference in promoting data literacy.
The good news is, in almost every large agency, there’s usually at least one analyst that can do this. There’s almost always someone who has studied Edward Tufte’s work, or has been inspired by Florence Nightingale or John Snow’s revolutionary data visualizations, or otherwise knows how to create an impactful data presentation. If not, there are always consultants happy to help.
Ultimately, increasing overall data literacy will grease the wheels of progress and make data initiatives more efficient and easier to adopt. While data engineers will still be the ones doing most of the data work, everyone must own the work from a business perspective.
This article is part of StateTech’s CITizen blog series. Please join the discussion on X (formerly Twitter).
