Guest blog by Jane Wiseman, Institute for Excellence in Government.

As a fellow at the Harvard Kennedy School helping to support a national network of state chief data officers, I’ve had a front row seat to leading edge analytics for the past three years.  As a more recent observer of chief data officers in the federal government, I’ve been impressed by the excellence and diversity of innovative data work being done by the pioneering data officers in federal service.  

While reading the Federal Data Strategy Action Plan, I was inspired by how detailed and thoughtful it is.  Actions address a range of important topics, and it’s wonderful to see data governance and data literacy getting attention, as these topics are not glamorous and sometimes get too little focus at the state and local level.  There’s an aggressive timeline in the action plan and a remarkable amount has already been accomplished.

I was honored to share my thoughts at the Federal Data Forum hosted by the Data Coalition and the White House in early July, and was energized by the range of experts who shared their ideas. The simple fact that OMB is creating a data strategy for the federal government is one of the most exciting developments in data-driven government in my career and offers a tremendous opportunity to deliver more value for taxpayers.  

Listening to other speakers during the forum, I was surprised twice – first at the common concern about the gap between the needed data literacy for government managers and current skill levels, and secondly by how few voices were calling for agencies to create and publish their data strategy.  

Observation #1: The federal government needs to invest in data literacy.  

Most leaders and decision-makers in government are not digital natives and many lack confidence in their data and digital skills.  And this gap will slow the adoption of data-driven government.  

• With 2.5 quintillion bytes of data created every day, we are swimming in data.  Unfortunately, few senior leaders are fully comfortable with their own data knowledge and skills, or their organization’s capacity for data use – 4 out of 5 government leaders say decision-making is either somewhat or rarely data-driven according to a PwC study.

• There are 90,000 units of local government in the US. Only a small percentage have a designated data leader such as a chief data officer (CDO), and only 20 have chief data officers who are part of the Civic Analytics Network (CAN), a peer network of leading state government data leaders.  

• Data-driven government can bring tremendous value. For example, the Health and Human Services Office of the Inspector General chief data office organization returns $5 in results for every $1 in cost at, including $1.3 B takedown of medical fraud.  

• Data-driven government is about decision-making.  But if decision-makers are not data-literate, you can have all the data scientists in the world in an agency but decisions won’t use their results – it will be as if the results are in Greek and no one translated.

Executives need to know how to ask for data analysis (what’s possible and what’s not), and how to critique results. They don’t need to be able to code or build an algorithm or map themselves, but they should know the power and capability of modern data tools. Basic data literacy means knowing how to ask good questions that inspire analysis, and then having the confidence to interpret and use the results. Let’s take a comparative study that was done between the US and 23 countries. In a comparison of the skills of the adult workforce in the US to those in 23 countries on "literacy, numeracy, and problem-solving skills," Japan and Finland led the pack on numeracy skills, while the United States ranked a disappointing 21 out of 23 participating countries. To close this achievement gap and to move toward large-scale organizational change to data-driven government will take a variety of training and coaching offerings. And it will take time.  

Recommendation: The federal government should provide a full suite of data literacy training, support and coaching for senior leaders and management executives in government so that they have the confidence to lead data efforts.

Observation #2: Each agency should have their own data strategy and plan. 

As the saying goes, “If we don’t know where we’re going we might end up somewhere else.”  

In the Federal Data Strategy Action Plan, actions 12-16 call for parts of what would be agency-specific data strategies. But this call to action falls short of asking each agency head to publish a multi-year data strategy and to report annually on progress toward achieving that strategy. We need to know where each agency is going in order to achieve data-driven decision making at every level of the organization and across all bureaus and agencies. Without a long-term strategy, we can’t hope for permanent culture change.  

The elements that exist in the action items — from data governance to open data to important data questions —  need to be knit together into a multi-year plan for optimizing the use of data. Targets also must be set for each that can be reported on an annual basis in a cohesive integrated, department-wide plan.   

With a clear charter from the chief executive, and armed with a strategy, a chief data officer can define a roadmap describing the difference the chief data officer team can make in government over a three- to five-year time horizon. The best chief data officers at federal, state and local levels operate from a strategy, a guiding document that sets mission and vision. They share their strategies and that helps them stay focused, and helps communicate to others what is and is not expected (e.g. chief data officers are not the one you call when you need someone to fix the printer and have an email server issue).  

Agency heads must be the individuals ultimately responsible for their data strategy and must invest their chief data officers with the authority and resources to carry it out. It needs to be clear to all the chief executive has a strong relationship with and relies on data, and views the chief data officer as a trusted and important resource to the organization.  

Strategy is about mission and outlining key questions. Asking good questions matters and clearly defining how the analytics research question will generate public value is important. The importance of understanding the connection to priority policy questions is summed up well by one of the CDOs I interviewed last year who said, “You might be Lord Algorithm but if you don’t stop to understand the problem, you will never succeed in making government better.”   

When the UK created the Government Digital Service in 2011, their focus was on long-term transformation of government to digital services, and they’ve made consistent progress every year.  But it didn’t all happen overnight. One of the keys was to make sure every agency had a point person and that he or she had a roadmap. We need that same level of focus on individual federal agency level strategies.  

Recommendation: OMB should hold agency heads accountable on an annual basis for progress on achieving data-driven government, should require them to be the leader of their department’s multi-year strategic plan for data, and require the publication of their data strategy.  

Successful Implementation Holds Promise for Lasting Impact 

With this exciting momentum at the federal level toward becoming more data-driven in decision-making, there is a tremendous opportunity for the federal government to support the development of capacity in state and local government as well.  For example:  

• The tools created for federal agencies will benefit state and local governments, where there is far less capacity to build knowledge capital and to document tools and methods.  A great example is the data ethics framework and data protection toolkit called for in actions 3 and 4 in the federal data strategy.

• The federal government should create Data Analytics Hubs or Centers of Excellence in a service model, ensuring state and local governments do not have to replicate analytics capabilities Rather, major economies of scale can be achieved if the federal government establishes analytics services hubs to provide data analysis for local governments.

• Procurement expertise and advice could also be part of these Centers of Excellence. With the rapid growth of the analytics and data mining vendor field, it would be impossible for any local official to have time to stay abreast of trends and to know the right questions to ask to assure that procurements are able to properly protect data privacy and security and to properly take into account data ethics.  A central resource at the federal level could support better investments at the state and local level and reduce the risk of data breaches and improper or wasteful spending of state and local funds for data analytics projects. 

The federal government has a unique opportunity at this moment to help incubate and advance the field with actions in every agency. Leadership and investment in capacity now will pay dividends long into the future.  

The Federal Data Strategy, with an iterative and collaborative approach, should support agencies moving forward. If it’s done right, the strategy will lead to a major transformation of government.