Our Perspectives

Challenges Faced by Federal CDOs

“Data is the new oil.” When identified, extracted, processed, and refined properly, data can be yielded into flexible resources that can be used in powerful ways. And yet, this way of thinking has only become common over the past few years, and in many sectors, data is still a relatively untapped resource. This is especially true in the federal government, which has vast stores of some of the most exquisite and most useful data in the world.  

In 2018, the Foundations for Evidence-Based Policymaking Act (The Evidence Act) was passed to correct this deficiency, creating a pathway for effective data use within the federal government, starting with the requirement for government agencies to appoint leaders to oversee their use of data.  

As the first Chief Data Officer (CDO) at the Department of Homeland Security (DHS), I experienced first-hand the challenges of realizing the true value of data within the federal government. I had the opportunity to survey stakeholders across DHS and understand the range of use cases for the data. The vast majority of users had relatively unsophisticated use cases for the data. Their immediate requests usually involved simply being able to collect reliable data and display it to a senior decision maker. This usually involved collating data from forms and displaying it in a dashboard that was kept updated on a regular basis. More sophisticated users looked to synthesize and analyze operational data to better understand principal drivers and optimize performance. A prime example was work done to understand the drivers in processing asylum seekers. This involved synthesizing data from multiple agencies across DHS. The most sophisticated users such as the ICE HSI lab were using AI/ML to examine large amounts of data and fully integrate it into their process. 

Federal data sources are both an opportunity and a challenge – they are huge and unique. Some of the challenges are technical – most of the data is siloed off in legacy programs with differing data standards and with limited options to share or interact with the data. The organizational and budgetary challenges are more severe. The Evidence Act laid out a vision for data-powered policymaking, and the Office of Management and Budget (OMB) has fleshed out many of the requirements, but it is still a largely unfunded mandate. Many agencies struggle with whose responsibility it is to fund enterprise data efforts. 

Because of this lack of funding, many CDOs focus on relatively inexpensive measures to “prep the battlefield”– to include establishing data policies, governance, best practices, standards, and data dictionaries. These measures make data more available and reduce the friction to integrating data from multiple programs but are only the first steps to improving mission performance. Agencies are still wrestling with where to focus their funds to best leverage the data they have.  

How should the government prioritize funding? The mission should serve as the primary filter to prioritize the funding, and the CDO must define the mission needs to see how data analytics and AI/ML will help. These opportunities need to be weighed against the availability of data and maturity of AI/ML algorithms to be applied. By considering all aspects of the opportunities via the CDO’s expertise, government agencies can avoid the scattershot approach prevalent in government activities at this point. 

To identify the best opportunities and take advantage of the “new oil,’ the government needs to bring together both mission experts who understand the problems that need to be solved and data scientists who understand the emerging technology and the realm of the possible. This combination is vital to select the true game changers that will improve the way we meet the mission.   

Redhorse Corporation combines sophisticated data science tools with artificial intelligence and machine learning to find new insights to accelerate the decision-making process across the challenges that government faces. Our team includes 400 plus engineers, data scientists, and subject matter experts that are experienced at developing and tailoring focused solutions and surfacing actionable intelligence and value from data. We partner with best of breed AI/ML product companies to deliver best of breed capabilities.