Synopsis: Maintaining statewide enterprise data is key to states being able to effectively make investments and decisions, leverage federal program funding, and support important governmental and social objectives, such as health, safety, development, and equity. To be successful states need to be able to bring together diverse stakeholders from their own agencies, from nonprofits, from local government, private sector partners, and sometimes federal agencies, in an organized way.
AppGeo has completed more than 35 strategic and business plans in 24 different states across the country. The webinar (recorded on October 6th, 2021) brings together Bill Johnson, Carpe Geo Evangelist at AppGeo, and Kate Hickey, Chief Operating Officer at AppGeo to discuss the common challenges of maintaining statewide Enterprise data, and to introduce three data governance models for consideration. By thinking about statewide data programs through the lens of these models, states can better plan for the resources, engagement, and authority needed to manage their statewide data assets.
The Three Models
Johnson: “It hasn’t gotten easier to do these enterprise projects. While the technology and tools we have have gotten better, the process of working with a wide range of stakeholders to bring these data together is hard! The real key here we have learned is that the non-technical aspects are what make these challenging. That’s why we have taken what we’ve learned in our years of experience and broken it down into three models. These models encompass a variety of factors: the funding, the governance, the authority of who has the data, data standards and so on. Each of these has to be adapted to the individual state’s needs.”
Different states, depending on their internal structures, on the stakeholders involved, and on the data they are attempting to consolidate have chosen different models for organizing the effort to maintain state data layers. The three models that we see – decentralized, federated, and centralized – each represents a different level of centralization of authority and control over the process. Each is a viable approach and they all present different challenges and trade-offs.
Johnson: “Many times states will have to address data being held in a decentralized way. This is what happens in the absence of state leadership, and it’s not that anyone is doing anything wrong, but they’re just doing things themselves and making it work for their own needs. For state agencies they might just be focusing on their own particular mission. The issue is that this can lead to duplication of effort and data. Sometimes agencies might receive funding to do things in a certain way which can make them reluctant to collaborate or change their methods. Decentralization is also the baseline for local data, things like parcels and addresses which fall under a local responsibility, and therefore there is a tendency for standards, data handling, and data quality to differ too.”
Decentralization occurs when there is a lack of business drivers to centralize data that is managed locally. This can be a useful model because it keeps data accessible and accountable to local decision makers who will make the most use of it. However, it can also be a hindrance to larger (multi-jurisdiction or statewide) projects because of differing standards and management practices between various stakeholders, and the cost of collecting, aggregating and harmonizing data. Without an economic incentive to centralize, the effort and cost required to gather and standardize these data mean that it is unlikely to happen.
Hickey: “The heavy need for coordination and cooperation in federated data models points to the non-technical elements. Time and time again what we are seeing in strategic planning is that those are the most challenging. You have to create buy-in, make the framework, identify the roles and responsibilities, and have that communication in place to keep everyone involved. What makes it hard is that this is not a one time thing, it’s a process that keeps going. There’s not really a shortcut to getting all these human factors in place and then supporting it.”
The advantage of a federated model is that it allows stakeholders to play a more active role in the areas that they have the most expertise in, and reduces the need for constant oversight and micromanaging. On the other hand, creating a framework that balances the actions and needs of individual stakeholders with the goals of the larger organization requires clear communication and dedication to maintain.
Hickey: “Centralization offers standardization of information. For example, in many states, imagery is produced through a single statewide contract, as it is expensive and therefore inefficient to have each county or other small municipality creating or buying their own data. For many types of data it makes the most sense for there to be one centralized data set. In the case of imagery, it would be extremely difficult for data to be useful if not centralized and standardized in some way. Factors like different flight times, height, tiling schemes, or even resolutions can make data non comparable from one jurisdiction to the next.”
Centralization of data ensures that data is collected and stored in a more uniform way than if such processes are handled by different actors. Additionally, it allows for uniform distribution of the data among various stakeholders. But this centralization is often harder to achieve than other models of data maintenance because of the coordination and buy-in required from all the actors involved. An example of a successful centralized program is the Texas Imagery Service provided by the Texas National Resources Information System (TNRIS), who AppGeo has worked with to help create a centralized database of imagery across Texas.