12 minute read 18 May 2023

So your agency has a data strategy, now what?

Many organizations have a data strategy in place but may stumble when it comes to execution. Some early successes point to actions that could help public organizations use data to create mission value.

Bruce Chew

Bruce Chew

United States

Sarah Milsom

Sarah Milsom

United States

Chris Foss

Chris Foss

United States

 Sean Fitzgerald

Sean Fitzgerald

United States

In organizations large and small, both in government and in the private sector, many executives seem to have reached the same conclusion: Data must be at the forefront of their organizational vision. The growing consensus on the value of a data-centric and data-savvy organization, and the exponential growth of data and analytical tools, led to the creation of the 2021 Federal Data Action Plan, the CDO Council, and an increasing number of chief data officers (CDOs) at all levels of government.1 As a result, there has been a wave of recently created government data strategies. So, data strategies exist across much of government—at least on paper.

Challenges in implementing data strategy

Implementing a new strategy can be hard. A review of studies of strategy execution effort success rates found that, on average, about half fail.2 There is a well-worn list of the factors that can cause transformative efforts, including executing new strategies, to fail. The usual suspects can include insufficient funding, the lack of a clear and communicated vision and sense of urgency, inadequate support from top leadership, and so on.3

The usual suspects will be potential barriers to success for executing data and analytics strategies. But the focus of this discussion is on the distinctive challenges involved in implementing data strategies, and data strategies in government in particular. Even leaders with experience in leading change should consider these elements if they hope to translate their espoused data strategies into value and impact for their organization and mission.

1. Data strategies often cut across current business processes and organizational boundaries

To effectively advance data maturity and center analytics in the day-to-day work of a complex organization, the realities and nuances of the current state processes and practices should be thoroughly understood to effectively design for the future state.

Data strategies can entail massive changes across an enterprise that can strain organizational, technical, and cultural norms. Advancing data across all the processes and practices of an organization can be disorienting for an organization and the working-level personnel tasked with the change.

2. The exponential growth in data within organizations can make data strategy implementation particularly daunting for many, creating substantial cultural barriers

Government agencies are not one, monolithic enterprise, but a confederacy of component organizations with their own quirks, attitudes, and institutional priorities. Each bureau or office often will have a different baseline level of data maturity, meaning that they will require an implementation approach that’s tailored to their challenges and goals. Data maturity differences can be compounded by the variety in data types and data storage, which can create a layer of technical challenges that require a sophisticated approach. Leaders should understand that a one-size-fits-all approach to implementing a data strategy is unlikely to succeed. It is imperative that leadership meets people where they are to understand their priorities, their concerns, and the practical challenges they face.

It’s important to consider that various sub-organizations may have their own data management processes, policies, and practices. Consequently, implementation efforts may encounter resistance to change and hesitancy to share “sensitive” data, particularly if that organization is the only owner of the data product. Organizational components may see their ownership of certain data as a source of “power,” so it is important to facilitate a cultural shift toward a world where the real “power” comes from sharing data to enhance decisions and operations for the enterprise.

3. The case for change may not be well understood or accepted across the enterprise

Data has been an input and output of processes for decades–so why is there a change in mindset now? The value of a data strategy may not be inherently evident to many important components of the enterprise. Stakeholders may not understand how improving data management, standardization, access, and interoperability will impact their priorities and goals. It is imperative that a data strategy implementation approach connects to the priorities of organizational stakeholders to generate urgency for change and adoption.

4. The skills and language of data and analytics may be unfamiliar

Data literacy poses a significant challenge for many organizations to bring data to the forefront of their day-to-day work. Personnel often lack academic or professional experience with data analytics and can struggle to independently derive insights from data. This unfamiliarity can hinder the engagement and enthusiasm that is often essential to organizational transformation. To overcome these obstacles, data strategies and implementation efforts need to place a high importance on data literacy programs that are inclusive of all skillsets, while connecting data to their day-to-day work.

5. New players and new roles are being created

For many organizations, the development of a data strategy often includes the establishment of new roles: chief data officers (CDOs), chief analytics officers, or chief AI officers are all emerging across public sector executive teams. The creation of any new organizational verticals entails the complicated task of establishing responsibilities, oversight mechanisms, and bureaucratic turf. Because data can and should be incorporated throughout the entire organization, the boundaries of a CDO’s reach can be difficult to establish, both on paper and in practice. Ironing out the details of a new CDO’s purview from a policy and culture perspective is imperative to successfully propelling a data strategy forward. The CDO should play a critical role in the implementation of any data strategy, as they will have both the foundational data understanding and organizational authority to speak on behalf of the strategy’s mission. The CDO should encourage leaders to buy-in to the strategy’s implementation across their own parts of an organization, while also continuously communicating the strategy’s value to the internal and external partners that can help drive it forward.

6. The tools and technologies are rapidly evolving

The tools and technology available to an organization can often be disconnected from the aspirations of a data strategy. Technology is a critical component to modernizing the data infrastructure of an organization, but necessary upgrades can be costly, complex, and cumbersome. Operating within the constraints of the current technological environment, prioritizing high-impact investments, and understanding the distance between present and future state can be critical for a data strategy to be effective.

The challenges posed by a data strategy implementation should be overcome with creativity, collaboration, and a whole of organization vision.

Ways to effectively implement a data strategy

Data strategies are typically conceived with the sincere aspiration to transform an organization but may come up short as they encounter low levels of engagement, entrenched processes, technological constraints, and cultural barriers. One key to overcoming these obstacles is a thoughtful implementation approach that integrates data with the priorities, leadership, and day-to-day tools of the enterprise. Through Deloitte’s experience implementing data strategies across more than 28 public sector organizations, we have identified several actions that organizations should consider to effectively implement their data strategies and enhance their chances for success.

1. Create a campaign

One option to secure executive buy-in and scale a data strategy across an organization is to use a campaign approach to implementation. This novel implementation method anchors the data strategy in key organizational and leadership priorities, advancing data and analytics maturity while tackling issues that executives care about.

The idea of a campaign is simple–for time-bound sprints (e.g., 3-6 months), the organization would surge data and analytics resources on one or two key priorities for leadership that are critical to mission effectiveness, rather than trying to “move the needle” across every data strategy objective at once. This surge of resources can provide data engineering support, executive engagement, project management, and policy expertise to advance data capabilities and create impactful analytical products for the organization (figure 1).

The campaign approach is designed to address the root causes of data strategy implementation on multiple fronts. With campaigns closely linked to advancing key organizational priorities, leaders can better understand the incentives and benefits that data activities provide and may be more inclined to support implementation and remove impediments to progress. Similarly, alignment to mission and management challenges puts data silos in the spotlight and reframes the “why we can’t share data” conversation to “why we need to share data, and how do we do this safely and securely.” The campaign model also narrows the focus of implementation to address the practical challenges that specific organizational components face, rather than “boiling the ocean” for a one-size-fits-all approach. Finally, anchoring the implementation in critical mission challenges often illuminates a menu of small, quick win activities that can provide value to the organization on an accelerated timeline, enabling more consistent demonstration of value throughout execution. In all, the campaign model can help to rapidly demonstrate the value of a data strategy and ensure it is connected to the day-to-day challenges and priorities of an organization.

2. Go all in on talent

Individuals with strong data and analytics capabilities are in extremely high demand, with no sign of that changing in the near future. In fact, data scientist has been called “the sexiest job of the 21st century.”4 Government must compete with the private sector for this talent, without being able to offer the same salaries or stock options.5 So, should CDOs compete for talent, train their internal people, hire skilled contractors to work internally, outsource data activities, seek out open-source packages and ready-made tools, or what? The answer is “yes” to all of the above. This multipronged approach is proving effective for multiple agencies.

To coordinate these efforts, ensure quality, and make sure scarce resource are deployed against important problems, leading CDOs are also creating centers of excellence for data and analytic efforts. Whether that’s within the IRS office of Research, Applied Analytics and Statistics, or the North Carolina Government Data Analytics Center, the consolidation of analytical talent can offer increased effectiveness and efficiency while potentially creating a better work environment for data-skilled agency employees and/or on-site contractors.6

3. Institute deputized chief data officers

CDOs are critical to the advancement of data and analytics across an organization, but formal responsibilities, organizational structure, and cultural barriers can be constraints on their reach. One approach to compound data and analytics efforts is to embed personnel with responsibilities, such as a CDO, within discrete offices or components of the organization. This approach can connect the organization through a web of data and analytics experts with a shared mandate to mature capabilities in their respective offices. These deputized CDOs enable enhanced coordination on foundational elements of the data strategy, such as facilitating the adoption of data standards, propelling data-sharing agreements, and cooperating on metadata management through data inventory and cataloging. On a more granular level, these embedded data officers can have greater visibility and familiarity with their business unit’s processes, challenges, and opportunities. This can promote greater cooperation, heighten data-driven decision-making, and allow for the creation of more tailored analytic products at the suborganizational level.

The deputized CDO model can focus dedicated data resources and expertise in a way that is highly responsive to the varying data needs across offices, while establishing a mechanism that enables the enterprise to learn and grow together at scale. Leadership may then become more inclined to support data strategy efforts, since they have dedicated resources focused on the nuanced data needs in their respective office. The deputized CDO can also help translate the value of the data strategy efforts to leadership, and is well-positioned to champion new data tools, products, and standards that are deployed enterprisewide.

Additionally, embedding personnel responsible for maturing the data and analytics capabilities within a component of the organization–while at the same time being responsible for privacy and security of that data–can help mitigate concerns related to data-sharing. In some cases, this may mean that data remains siloed, but value is still gained when the use cases can be shared and adapted elsewhere in the organization. While this may limit some economies of scale since solutions are deployed multiple times, it can also increase the potential for innovation.

Large, highly federated government agencies, such as the Department of Defense (DoD), Department of Transportation, and Department of Homeland Security, have employed the deputized CDO model across their subcomponents (e.g., there are CDOs for each branch of the military in addition to a central DoD CDO).

Deputized CDOs can be powerful agents for advancing data and analytics within an organization, but the onboarding and deployment process should be well-crafted for them to succeed. Design a comprehensive selection, training, and onboarding process that holds technical knowledge and organization-specific processes in equal regard to make the most of a deputized CDO’s efforts.

4. Prioritize a platform

Developing a centralized platform for data assets, analytic products, and resources can enable broad, enterprise adoption of the data strategy. It provides an organization with a one-stop-shop to share and interact with enterprisewide data, tools, and guidelines that could otherwise be difficult to utilize. The DOD’s Advana platform for instance provides users with quick, easy access to common business data, decision support analytics, and data tools, that may otherwise exist in different locations and require extensive time to discover.7

A centralized platform can showcase the enterprise’s data strategy successes and provide important mechanisms for gathering evidence of user adoption. The platform can simplify and provide transparency to user analytic metrics that demonstrate demand, while at the same time establishing a method to measure certain key performance indicators (KPIs) tethered to data strategy implementation. This can provide greater transparency around how products are being used and their relative demand across the organization. As a result, leadership has a better understanding of the value of successful data strategy activities and may be more likely to invest in further resources and lend political support for efforts.

Additionally, a centralized platform can challenge data silos across an organization. It facilitates data-sharing, cross-organizational collaboration, and encourages data literacy. It also provides a shared space to publicize the work being done in different parts of the organization, thereby encouraging innovation and adaptation of existing work to new uses cases.

Action here will, of course, depend on the particular situation of a given agency, recognizing organization’s existing technological landscape. Technological solutions to help advance data capabilities can be expensive and often take place over a long timeline, while leaders need to deliver quick wins that can be accomplished with the existing tech stack. The technology component of a data strategy should match the skillset of the personnel–leaders should consider workforce skills, customizability, and the end goal prior to making large investments.

5. Invest in communications, outreach, and data literacy

Caryl Bryzmialkiewicz, an early government CDO and the first for Health and Human Services Office of the Inspector General, highlights proactive communication when asked for advice for government CDOs: “The ability to motivate and to pull people together depends on good communication skills and a bit of marketing.”8

Often an afterthought, early investment in communications, branding, marketing, and outreach can be crucial to the success of data strategy implementation. Transforming the way people use and manage data may require a great deal of change, and change is never easy. They should be brought along the journey from data strategy development throughout implementation to feel like they are included and have ownership of the outcomes. They can be fearful that new data and analytics products may replace their jobs, when in many cases it will simply reduce routine tasks while freeing them up for more value-added analysis and evidence-based decision-making.

Successful communications, branding, and outreach serve to build the case for change early and continue throughout implementation. When deployed at the enterprise level, a thoughtful engagement effort can accelerate the dissolution of organizational silos. It could increase the desire for all parts of the organization to be active participants in implementation when there is an opportunity for their work to be prominently showcased and can open feedback mechanisms to create solutions that work for a broader set of use cases. Likewise, a robust suite of communications, branding, and outreach capabilities can be important for demonstrating value beyond the immediate stakeholders impacted by an implementation activity. It also serves to highlight success and to keep an organization energized around progress, quick wins, and achievement of milestones along the road to implementation.

Finally, increasing the investment in training and data literacy efforts could increase adoption rates and ease many of the challenges presented by the development of new analytics products. Integrating the language of data into existing training and onboarding requirements can help to get ahead of the adoption curve, while enabling users to consistently derive valuable insights from the platforms available to them. Data literacy programs should be inclusive of all skillsets, providing helpful resources for both the data savvy and data inexperienced. More importantly, these programs must connect data concepts to the day-to-day work of the organization, so personnel learn how to use data tools and analytics platforms in practice. Data capabilities are only as useful as they are used–providing an organization with sophisticated tools without the skills and support necessary to use them is a recipe for disappointment.

Strategy implementation is rarely easy, and data strategies provide additional challenges due to a legacy of ingrained practices and the complexity of changing the way data is used and managed across the enterprise. But there are novel approaches to implementation that can overcome these challenges, and help organizations achieve their vision to harness data as a strategic asset.

  1. Federal Data Strategy, “2021 Action Plan,” accessed May 5, 2023; Federal CDO Council, “About us,” accessed May 5, 2023; CalGovOps, “The future – Cal data,” accessed on May 5, 2023; US Department of State, “Appointment of Dr. Matthew Graviss as Chief data officer,” press release, January 4, 2021;, “Mayor Breed and city administrator Chu appoint Michelle Littlefield as Chief data officer,” press release, February 7, 2022.

    View in Article
  2. Jeroen Kraaijenbrink, “20 reasons why strategy execution fails,” Forbes, September 10, 2019.

    View in Article
  3. John P. Kotter, “Leading change: Why transformation efforts fail,” Harvard Business Review, May-June 1995.

    View in Article
  4. Davenport, Thomas H., and D. J. Patil, "Data Scientist: The Sexiest Job of the 21st Century," Harvard Business Review 90, no. 10 (October 2012), pp. 70–76.

    View in Article
  5. Thomas H. Davenport and DJ Patil, “Is data scientist still the sexist job of the 21st century?Harvard Business Review, July 15, 2022.

    View in Article
  6. The Evidence Team, OMB, “In the spotlight – IRS Office of Research, Applied Analytics, and Statistics,”, November 17, 2021; Janice DeGarmo, “The Department of Analytics launches the Center for Analytics,” State Magazine, March 3, 2020; North Carolina Department of Information Technology, “N.C. Government Data Analytics Center,” accessed May 5, 2023.

    View in Article
  7. Brian B. Joseph, “Data analytics: Using data to enhance acquisition outcomes,” Acquisition: Office of the Assistant Secretary of Defense, accessed May 5, 2023.

    View in Article
  8. Jane M. Wiseman, Data-Driven Government: The role of Chief data officers, IBM Center for The Business of Government, 2018.

    View in Article

Cover Image by: Jim Slatton

Cultivate an Insight Driven, Connected Government

AI and automation have tremendous potential to improve business operations and drive mission impact. Join Deloitte’s Government & Public Services (GPS) Consulting and Advisory Leaders as they discuss key AI and data trends transforming the government and enabling users to cultivate connectivity through data.

Bruce Chew

Bruce Chew

Managing director, Monitor Deloitte | Deloitte Consulting LLP


to receive more business insights, analysis, and perspectives from Deloitte Insights