Pillar 4: Using Data to Drive Priority-Setting, Decision Making, and Performance

This is the fourth blog in a series highlighting key insights from the IBM Center's Special Report, Five Pillars of Effective Government.

Government agencies face unprecedented pressure to demonstrate both efficiency and effectiveness. The ability to harness data for priority-setting, decision making, and performance management has evolved from an aspirational concept to an operational imperative. As federal, state, and local governments navigate challenges ranging from public health emergencies to infrastructure modernization, the strategic use of data has become the cornerstone of responsive, accountable governance.

Landmark legislation such as the Foundations for Evidence-Based Policymaking Act of 2018 reflects this reality, requiring federal agencies to develop learning agendas and demonstrate that programs deliver intended results. As the fourth pillar, data-driven decision-making builds upon and reinforces the other foundational elements of effective government. It provides the evidence base for strategic planning, the metrics for performance management, and the transparency mechanisms that foster accountability. Without robust data capabilities, government agencies operate with limited visibility into program effectiveness, struggle to identify improvement opportunities, and face challenges in demonstrating value to stakeholders and the public.

Data-driven governance demonstrates how quality data, measurement systems, advanced analytics, and the Evidence Act enable better decision-making while addressing fraud prevention, organizational capacity, transparency, and cross-agency collaboration.

The foundation of effective data-driven decision-making begins with high-quality data. Data provides the evidence needed to understand current conditions, identify problems, and evaluate the effectiveness of policies and programs. It enables decision-makers to set priorities based on objective criteria rather than subjective judgments or political pressures alone.

Recent research shows that 81 percent of U.S. federal government analytics professionals’ express confidence in their agency data quality. However, quality must extend beyond accuracy to encompass timeliness, relevance, and accessibility.[1] Automated data collection enhances both quality and efficiency, reducing human error while freeing analysts for interpretation. The Health Resources and Services Administration's UDS Modernization Initiative[2] demonstrates this approach, as do Internet of Things (IoT) sensors deployed by cities for air quality monitoring and smart traffic management.[3]

Agencies must also think strategically about data collection. Not all data is equally valuable, and collection efforts should be guided by clear questions about program effectiveness and mission achievement. The three critical questions for any data collection initiative are: What is the value from the proposed dataset? What is the feasibility of scaling the dataset? What is the sustainability of the data collection project? These questions help ensure that resources invested in data collection yield proportional returns in decision-making capacity.[4]

Public-facing dashboards and metrics transform abstract data into accessible information that drives accountability. However, selecting appropriate measures at program inception—and evolving them over time—is critical. Measures must reflect true outcomes rather than merely what is easy to collect.

The Pandemic Response Accountability Committee (PRAC), established to oversee pandemic relief funds, demonstrates this approach. Using the Pandemic Analytics Center of Excellence and through sophisticated data analysis, the PRAC identified fraudulent applications tied to stolen social security numbers, showing how measurement systems can simultaneously ensure program integrity and operational effectiveness.[5]

Another instructive example is the Federal Information Technology Acquisition Reform Act (FITARA) scorecard, which measures agency progress on IT modernization and governance. Originally focused on data center consolidation, the metrics evolved to allow agencies to choose "fit for purpose" computing environments[6] including cloud services, demonstrating how measurement systems must adapt while maintaining focus on core objectives, a principle that aligns with agile approaches to government operations.[7]

Comprehensive data strategy requires defining data elements across networks and including disaggregated data for population-specific insights. Organizational boundaries, incompatible systems, and cultural resistance to sharing information often impede the collaboration necessary for addressing complex public problems.[8] Since few holistic questions can be answered by single data sources, breaking down organizational silos becomes critical. This requires not just technical solutions but sustained leadership commitment and cultural change.

Effective data strategy begins by defining programmatic questions. The U.S. Department of Veterans Affairs' response to the opioid crisis[9] illustrates this: understanding and addressing the issue required data on educational and job opportunities, housing, food security, and transportation access from multiple agencies—Labor, Education, HUD, and Agriculture—highlighting the importance of reducing friction that prohibits information sharing.

Successful intergovernmental data sharing requires persistent leadership, cross-functional teams, and robust governance structures. Examples from state and local governments demonstrate that while data sharing agreements can take months to negotiate—Allegheny County’s agreement with Pittsburgh Public Schools required eighteen months—the resulting integrated data systems enable dramatically improved service delivery and resource allocation. The Commonwealth of Virginia Data Trust, developed initially to address the opioid crisis, enabled the state to deploy COVID-19 dashboards in days rather than months, demonstrating how investments in data sharing infrastructure pay dividends during emergencies.[10]

Disaggregated data is particularly crucial for understanding how different groups are affected by policies and programs, allowing for more targeted and effective interventions. For example, disaggregated data on education outcomes can reveal disparities between demographic groups, guiding efforts to address educational inequities. This granular understanding enables agencies to move beyond one-size-fits-all approaches toward interventions tailored to specific community needs.

Modern data-driven governance extends beyond descriptive statistics to predictive analytics using AI and machine learning. However, these tools should be viewed as evolutionary extensions of existing efforts rather than revolutionary replacements. Research shows that communicating and interpreting results has become the top focus of analytic effort, surpassing data gathering[11]—insights only create value when effectively communicated to decision-makers in accessible formats.

The AI-Driven Internet of Things (AIoT) offers transformative possibilities. Smart sensors provide real-time data while AI algorithms analyze patterns to enable predictive maintenance, optimize traffic flow, and improve public safety. Pittsburgh's smart traffic control systems, which optimize signal timing to reduce travel time and emissions, exemplify how AIoT creates adaptive systems.[12] These AIoT applications combine sensing, analytics, and control capabilities to create adaptive systems that respond dynamically to changing conditions.

Evidence-Based Decision-Making and the Evidence-Based Policymaking Act

The Evidence Act institutionalizes data-driven decision-making by requiring agencies to develop evidence-building plans, conduct annual evaluations, and assess their capacity to use evidence.

Implementation of the Evidence Act provides valuable lessons in integrating evidence-based practices into strategic planning. Agencies must move beyond viewing evidence requirements as mere compliance exercises and instead embrace them as opportunities to fundamentally improve how they understand and advance their missions. This requires building organizational capacity—not just in technical analytics skills, but in the ability to frame answerable questions, interpret findings in context, and translate insights into action.

The creation of learning agendas—systematic plans for building evidence to inform decision-making—exemplifies the Evidence Act approach. Rather than conducting isolated studies, learning agendas establish coherent research programs aligned with agency priorities. This ensures that evaluation activities build upon each other and contribute to cumulative understanding of what works, for whom, and under what circumstances. The iterative nature of learning agendas reflects agile principles of continuous learning and adaptation based on evidence.[13]

Technology and data infrastructure are necessary but insufficient for effective data-driven governance. The fourth pillar requires organizational capacity spanning technical skills, analytic methods, and—perhaps most critically—cultural commitment to evidence-based decision-making.

Leadership buy-in emerges consistently as critical. Survey research on government analytics found that persistent, visible support from senior leaders ranks alongside clear communication about results as the most significant factors for success.[14]  Leaders must actively use evidence in decision-making, ask data-informed questions, and create space for findings that challenge assumptions. This commitment proves particularly important when adopting approaches like agile methodologies that require sustained cultural change.[15]

Building this culture requires bridging the knowledge gap between technical analysts and program leaders requires reverse mentoring and collaborative workshops. These human capital investments may matter more than technology investments, as sophisticated tools deliver little value if insights fail to reach decision-makers.

Recruiting qualified analytics staff requires competing with the private sector by emphasizing government's unique value proposition: meaningful work affecting millions combined with access to comprehensive datasets. Agencies with analytics-friendly cultures that value data-driven decision-making find themselves better positioned to attract top talent.

Data-driven governance extends beyond internal decision-making to transparency and stakeholder engagement, enabling citizen oversight, informed participation, and institutional trust.

Public-facing dashboards transform raw data into citizen insights, but accessibility demands thoughtful design for diverse audiences. Data disaggregation—breaking down information by demographic categories, geographic areas, or program components—enables understanding of community-specific impacts.

The trend toward self-service analytics reflects both technological advancement and philosophical commitment to empowering stakeholders. Self-service models allow users to explore, visualize, and analyze data using dashboards and cloud-based tools rather than depending entirely on centralized analytics teams.[16] This democratization of data access can increase efficiency, broaden perspective, and enhance accountability. However, it also requires careful attention to data literacy, ensuring that users understand not just how to access data but how to interpret it responsibly.

Government's most pressing challenges transcend agency boundaries, requiring integrated data and coordinated action. Yet organizational silos and legal barriers impede collaboration.

The pandemic response illustrated both the necessity and difficulty of data sharing.[17] When the U.S. Department of Labor initially indicated it lacked authority to demand unemployment insurance data from states, Inspectors General offices were forced to issue individual subpoenas to obtain information critical for fraud detection. This reactive approach cost precious time during a rapidly evolving crisis. Later legislation addressed some of these barriers, but the experience underscores the importance of establishing data sharing frameworks before emergencies strike.

Successful data sharing requires attention to technical standards, legal authorities, privacy protections, and governance structures. Memoranda of understanding between agencies can establish frameworks for routine data exchange, but these agreements often take months to negotiate. Standardizing data elements and formats across agencies—while respecting legitimate differences in mission and operational context—can dramatically reduce friction in data sharing when collaboration becomes necessary.[18]

As the fourth pillar of effective government, using data to drive priority-setting, decision-making, and performance represents more than a technical capability—it embodies a commitment to rational, evidence-based governance that serves citizens effectively and respects taxpayer resources. The journey toward truly data-driven government requires sustained attention to multiple dimensions: technical infrastructure and analytical capabilities, certainly, but also organizational culture, leadership commitment, staff capacity, and mechanisms for transparency and accountability.

Looking forward, several priorities merit particular attention and may include developing comprehensive data strategies, investing in fraud prevention, expanding self-service analytics, fostering data literacy, strengthening cross-agency collaboration, adopting agile approaches, and maintaining transparency for public accountability.

Using data to drive priority-setting, decision-making, and performance rests on a simple premise: government should make decisions based on the best available evidence, evaluate them against clear metrics, and maintain transparency. This requires balancing quick wins with long-term investments in infrastructure and capacity, recognizing that data-driven governance is a continuous journey of learning and improvement—one that promises more effective government, better outcomes for citizens, and stronger democratic accountability.

Footnotes 

[1] Jennifer Bachner, Optimizing Analytics for Policymaking and Governance (Washington, DC: IBM Center for The Business of Government, 2022), 6,

https://www.businessofgovernment.org/sites/default/files/Optimizing%20Analytics%20for%20Policymaking%20and%20Governance.pdf.

[2] Health Resources and Services Administration (HRSA), Bureau of Primary Health Care. “UDS Modernization Overview.” Updated 2025, https://bphc.hrsa.gov/data-reporting/uds-modernization.

[3] Gwanhoo Lee, Creating Public Value using the AI-Driven Internet of Things (Washington, DC: IBM Center for The Business of Government, 2021), 28. https://www.businessofgovernment.org/sites/default/files/Creating%20Public%20Value%20using%20the%20AI-Driven%20Internet%20of%20Things.pdf.

[4] Bachner, Optimizing Analytics for Policymaking and Governance, 15.

[5] Goodrich and Westbrooks, A Prepared Federal Government: Preventing Fraud and Improper Payments in Emergency Funding, 23.

[6] U.S. Government Accountability Office (GAO), "Information Technology and Cybersecurity: Evolving the Scorecard Remains Important for Monitoring Agency Progress," GAO-23-106414, December 2022, https://www.gao.gov/products/gao-23-106414.

[7] Ganapati, Adopting Agile in State and Local Governments, 47,

[8] Wiseman, Silo Busting: The Challenges and Success Factors for Sharing Intergovernmental Data, 21.

[9] U.S. Department of Veterans Affairs, “Social Determinants of Health” (updated 2025), https://www.va.gov/health-care/health-needs-conditions/social-determinants-of-health/.

[10] Wiseman, Silo Busting: The Challenges and Successes of Intergovernmental Data Sharing, 7, 18–19.

[11]Bachner, Optimizing Analytics for Policymaking and Governance, 29.

[12]Lee, Creating Public Value using the AI-Driven Internet of Things, 15.

[13] Ganapati, Adopting Agile in State and Local Governments, 19.

[14] Bachner, Optimizing Analytics for Policymaking and Governance, 17.

[15] Ganapati, Adopting Agile in State and Local Governments, 25.

[16] Bachner, Optimizing Analytics for Policymaking and Governance, 27.

[17] Goodrich and Westbrooks, A Prepared Federal Government: Preventing Fraud and Improper Payments in Emergency Funding, 15.

[18] Goodrich and Westbrooks, A Prepared Federal Government: Preventing Fraud and Improper Payments in Emergency Funding, 30.