Recently, I was speaking to a congressional staffer about the need for data governance if agencies, companies, and the evidence community want to do artificial intelligence (AI) well. Their response was, “Great—but what does that mean, and how do we do it?”
Let’s start with a definition for data governance and stewardship. I like this one from Data Management Association International:
The exercise of authority and control (planning, monitoring, and enforcement) over the management of data assets.
So what does that look like in practice? When organizations think about managing financial assets, they look to align financial expenditures with business priorities. Many of the ideas behind managing data are the same, but the difference is that data can be used more than once—and the more copies we have, the less control we have over how it gets used. So organizations looking to manage data as an asset have to place a premium on data governance and stewardship to ensure that they are making decisions based on the right data.
Based on what I’ve seen organizations get right (and wrong) along their data governance journeys, here are some key themes to get started:
Start with your business drivers
When it comes to data governance, it’s important to understand that what an organization considers to be data assets are distinct to that organization. Determine what your organization’s business priorities are. Are there enterprise measures or key performance indicators (KPIs) that must be aligned with staff work? Do you have a mission or vision for the organization that guides how you operate? Do your KPIs align with that mission or vision? This is where you start. If you don’t have these things documented, then you need to talk to your different business units to figure out how they are running their business.
Identify key stakeholders
Your organization almost certainly employs staff who help mitigate risk to its financial or physical assets. An organization also requires staff who can help you use data to understand how your business is being run and manage data risks. Small organizations might be able to use more informal methods of data governance, but with growth it becomes harder to ensure that the key data you need to run and measure your organization is of high quality, is available on a timely basis, and answers the questions that enable you to run your organization.
It’s not difficult to determine who’s collecting, curating, and using data within an organization. Often, you find that the same person is engaged in all three functions. Once you identify who is doing this informally, it’s important to formalize the role and define other roles explicitly, which also helps with efficiency.
Document where data lives and who has access
Establishing data governance requires an understanding of where data live and who is accountable for the data. Who has access and what do they have access to? What can they do with the data, how long will they have access to the data, and how is the data accessed, stored, and used? Conducting an inventory of where the data lives is the first step in understanding who should have access.
Create an enterprise glossary
As organizations grapple with data and technology challenges, it is critical to ensure staff are speaking a common language. Data Management Association International’s Data Management Body of Knowledge is an extensive and time-tested resource, offering common definitions, practices, and templates that can help organizations at any point along their data governance and data management journey. It’s also important to identify an internal champion that understands the importance of data to the organization and is a trusted voice.
Engage a partner that can help guide your journey
Whether your organization is ahead of the curve on governance, or just beginning to understand the importance of getting AI right, it’s important to have someone by your side on your data journey. As a mission-driven organization trusted for our ability to turn high-quality data into reliable insights, we know AI is only as good as the data behind it. This foundational commitment to being a trusted data steward grounds our approach to AI. We merge cutting-edge digital technology with deep subject matter expertise, a keen understanding of the policy landscape, and rigorous data governance to identify and address gaps in data, ensure accuracy, and uncover action-ready insights. We’re eager to help you harness the power of AI by ensuring the data that inform it are comprehensive, high-quality, and free of bias. Let’s progress together on data governance—and get AI right.
