Attention: You are using an outdated browser, device or you do not have the latest version of JavaScript downloaded and so this website may not work as expected. Please download the latest software or switch device to avoid further issues.
| 29 Jul 2026 | |
| Blogs |
By Hyon Kim and Chris Marcum, Senior Fellows, Data Foundation
In the pre-digital era, researching a question often started at a library. If you needed to know the level of trade between the United States and China in the previous year, you would go to a research library that held the statistics published annually by the federal government. To find the answer, you would have used the library's card catalog, which organized the library's holdings. Each card contained consistent information about a publication: author, title, date, description, and physical location. That information would lead you to the shelf holding the annual trade statistics published by the Department of Commerce.
The card catalog is now mostly ornamental at any library that has kept one. We're used to getting instant answers to questions like the size of the trade deficit. But the source is still the same. The federal government collects, processes, and publishes the data that informs everyday decisions and supports business, research, and policymaking. For a search engine or an AI tool to find the right answer, the government's data holdings need the right metadata, just as every card in a catalog once carried consistent information about each book in a library's collection.
Congress recognized the importance of finding and accessing federal data in the Foundations for Evidence-Based Policymaking Act of 2018 (Evidence Act), which included the OPEN Government Data Act (OGDA). The OGDA requires all federal agencies to make their data open and accessible, and to maintain comprehensive inventories of metadata for their datasets so they can be included in a central data catalog, Data.gov. The General Services Administration has operated Data.gov since May 2009. The OGDA also directed the Office of Management and Budget (OMB) to issue detailed implementation guidance for how agencies should comply with the Act.
OMB's guidance, M-25-05, was released in January 2025 and gives agencies specific instructions for their comprehensive data inventories (CDIs), the metadata records that include title, description, date, keywords, and an access or download link (among other critical information). Complete, high-quality metadata in those CDIs is what lets the public get an answer that's fully informed by federal data. And because the federal government holds such vast and varied data on so many topics, that metadata also needs to work with international standards, so a query made from anywhere in the world can draw on it.
That's why OMB's guidance directs federal agencies to adopt a metadata standard called DCAT-US v3.0, a U.S. profile of the widely used international W3C DCAT specification. It was developed through a collaborative effort of the Chief Data Officers Council, the Federal Committee on Statistical Methodology, the Interagency Council on Statistical Policy, and OMB, known as the FAIRness Project, a name borrowed from the Findable, Accessible, Interoperable, and Reusable data governance principles.
The new standard was released in May 2026 and is documented on both the DCAT-US Schema v3.0 page at resources.data.gov and the GSA GitHub repository, reflecting strong work by the teams at GSA, OMB, and the Chief Data Officers Council. The documentation shows agencies how to describe the essential metadata for every dataset so it can be found and used.
M-25-05 requires federal agencies and Data.gov to be updated to DCAT-US v3.0 by September 30, 2026. Given reductions in agency staff and the short time since the technical standards were published, it's unlikely every agency will meet that deadline. Fortunately, some agencies, including the Departments of State, Agriculture, and Health and Human Services, have already begun implementation. And because most existing metadata elements carry over into the new standard, agencies won't need a wholesale overhaul of their CDIs.
Agencies should consult the detailed technical guidance at resources.data.gov and follow the Migration Guide, which explains the changes from the previous schema and prioritizes the required elements. Agencies that haven't yet built a CDI can start fresh with the updated schema. With complete, updated CDIs, agencies can make all of the federal government's data assets more discoverable and accessible to everyone.
There's broad, bipartisan support for open data, but OGDA implementation has never had a specific appropriation. Even in flusher budget years, this kind of work tends to fall low on the priority list. Congress should direct agencies to make the modest investment needed to keep federal data discoverable and accessible with up-to-date metadata, and they should fund it too.
Such an investment will pay returns in efficiencies. For example, M-25-05 clarifies that Congress intended that the metadata in the Federal Data Catalog would be the official source of information for other government systems, such as the Evidence Act’s Standard Application Process portal for discovering and applying for access to confidential statistical data.
The decisions consumers, businesses, researchers, and policymakers make every day should be informed by data the federal government has spent years, and real money, creating and maintaining. Skimping on metadata means an answer to a consumer's question, or a policymaker's, could miss important evidence. The same is true for AI tools, which are only as good as the data they can confidently discover and retrieve. Data is also central to the Trump Administration's priorities in the AI Action Plan and the Genesis Mission. Federal agencies need the capacity to make government data open and available, as the OGDA intended.
