As the Chief Data Officer (CDO) role gets established at each federal agency as required by the Evidence-Based Policymaking Act of 2018 (Public Law 115-435) (Evidence Act), I can’t help but feel conflicted. On one hand, I am excited that the federal government has taken such an intentional step towards a more data-driven government: There are enormous potential benefits when CDOs partner with program owners to innovate and deliver real values to the federal government and our citizens. On the other hand, I am concerned this role, if not properly supported and empowered, could become yet another silo yielding little real value.
As someone who has successfully implemented the first federal open data law, the Digital Accountability & Transparency Act of 2014 (DATA Act), I performed a role similar to the “CDO” (based on the qualifications and functions defined by the Evidence Act) for the entire federal government. As the Treasury executive who was leading the government-wide implementation, I was responsible for the lifecycle management of spending data across more than 100 federal agencies. I coordinated with Senior Accountable Officials from all 24 CFO agencies as well as functional communities including acquisition, financial assistance, and finance to build the first integrated administrative function data standards (over 400 interconnected data elements). I oversaw the publication and standardization of the high-value data assets provided by all federal agencies. To the extent practicable, I engaged with internal and external stakeholders to maximize the use of the spending data. I intend to share what I learned from this monumental effort with the hope it could benefit those embarking on a data journey at the federal level.
As a newly minted C-suite role, the federal agency CDOs will feel the need to quickly build credibility within their agencies. Although the Evidence Act has some specific mandates, the CDOs will need to think beyond those mandates to have a sustainable impact on the mission of their respective agencies. Based on conversations with multiple federal and state CDOs, I have developed a model (as shown in Figure 1) for an approach that will help agency CDOs maximize the return from their data.
Figure 1. Data Value Optimization Model
First, all innovation should be anchored to the organization’s goals or driven by the need to solve a business problem. In the private sector, organizational roles are organically created based on business needs. In the federal government, due to its great size and complexity as well as the need for specific authority to drive significant changes, legislation is often required to mandate a government-wide change, including to establish organization roles. For example, the Chief Financial Officer (CFO) Act of 1990 officially created the CFO role. Most recently, the Evidence Act established the authority of the CDO and the Chief Evaluation Office roles. Unfortunately, more often than not, these statutory mandates result in a compliance-based approach as opposed to a value-based approach.
As a result, CDOs should start by articulating the value proposition of their work in terms of their potential impact on the organization’s goals and how they solve their business problems. The model consists of five key components represented in a circular mode:
- Data Strategy: Creating a successful data strategy requires the leadership to assess an agency’s unique business challenges, match those challenges with relevant data and resources, and establish processes that grow their capabilities. At the macro level, the Federal Data Strategy Framework provides a good foundation for each federal agency. Leverage the framework to accelerate the planning process.
- Organization Policy: In the history of the federal government, many policies have been implemented to facilitate compliance with laws and regulations. Some of these policies (e.g. security, privacy, ethics, etc.) can constrain the accessibility and use of the data by agencies. It is important to understand and consider how they impact existing programs, processes, and stakeholder interests so that CDOs and other executives can assess the practicality of the scope and change management required.
- Data: This component includes establishing data governance and the infrastructure for data acquisition and normalization. Its ultimate purpose is to make quality data accessible for use to achieve business value. In data science terms, this means creating robust data pipelines for analytics.
- Use: Unless data is used to create value, the effort spent on data governance and infrastructure serves no purpose. To use data for value creation, agencies need to build their analytics capacity. That includes human capital (i.e., program subject matter experts, data analysts, and data scientists), tools, and models (i.e., machine learning and artificial intelligence). This component is the closest to the point of achieving value and therefore is most likely to garner support from business owners.
- Value: The ultimate business value from using data well is the insights gained that lead to actionable solutions for complex problems. For example, one type of advanced analytics that could help agencies improve efficiencies is Natural Language Processing (NLP) due to its ability to process a large amount of unstructured data. At Elder Research, we have provided sentiment analysis of large volumes of unstructured data using NLP techniques to assist federal agencies on risk assessment and prioritization. The efficiency gained benefited the agencies directly by enabling them to achieve greater outcomes with fewer resources. In 2015, I first spoke about a simple formula at a data conference and coined the data + use = value principle. The principle remains relevant today and I am glad that Treasury continues to apply this formula in driving its data vision.
The circular direction and two-way arrows in Figure 1 highlight the iterative nature of the process as well as the way each component informs the others. It also provides the flexibility for an organization to determine the appropriate starting point. Depending on the level of maturity in adoption and momentum, an agency may need to start at the Use stage so the non-technical business owners can envision the applications and benefits of data analytics as well as understand the value of data governance.
Also, consistent with an agile software development approach, each iteration should be tested, and debriefed to provide a basis for learning, or pivoting as needed. Last and most importantly, innovating with data is a paradigm shift so requires sustained support from executive leadership to create the culture and environment needed for success and adoption.
Of course, no model can replace human wisdom gained from experience. Based on my own experience, the biggest challenges I faced in driving innovation and implementing the open data law across the entire federal government were not technical—they were all related to navigating people, policy, and politics. Agency CDOs who are most skilled in influencing the government ecosystem are focused and disciplined, have the technical competencies, and also understand human motivation and incentive alignment. Those leaders are most likely to succeed at 1) maximizing the value their agency obtains from its data and 2) leaving a legacy of positive impact for public good.
