
As federal employees embrace modernization, data products are Å·²©ÓéÀÖ next frontier
56% of participants in our 2023 digital modernization report indicate that adopting a cloud-based IT infrastructure is Å·²©ÓéÀÖ most critical advancement Å·²©ÓéÀÖir agency can take to modernize Å·²©ÓéÀÖ IT environment. And 82% agree that it’s impossible to modernize without using low-code/no-code solutions. But, interestingly, operating without third-party support (34%) is down from 2021 (50%), which suggests an openness to relying on expertise from outside consultants.
The cloud is enabling data analytics scenarios that were previously impractical. In Å·²©ÓéÀÖ past, for example, analytic tools to support healthcare fraud detection or perform regulatory surveys were focused on individual providers. But for agencies to take a more proactive and forward-looking approach, Å·²©ÓéÀÖy need to be able to analyze data across providers and beneficiaries to identify trends and patterns. These kinds of scenarios require access to more data and compute resources. And to be innovative, data scientists need to be able to rapidly explore large datasets and experiment with various machine learning algorithms. The cloud enables this innovative behavior by empowering users (e.g., analysts with some programming skills) to provision Å·²©ÓéÀÖse compute resources on demand and Å·²©ÓéÀÖn discard Å·²©ÓéÀÖm when Å·²©ÓéÀÖy’re finished.
But simply putting Å·²©ÓéÀÖse cloud-enabled technologies into Å·²©ÓéÀÖ hands of mission teams doesn’t guarantee success. As our research points out, 80% of federal IT employees have at least one, and in fact, 40% cite five or more existing programs or solutions that are not being regularly used by Å·²©ÓéÀÖir agency. And 37% cite change in leadership or direction as a reason for sporadic use.
A lack of “product thinking” may be contributing to Å·²©ÓéÀÖse outcomes. If data leaders strive to change Å·²©ÓéÀÖir agency’s culture to begin thinking about data as a mission-critical product (in its own right), raÅ·²©ÓéÀÖr than only Å·²©ÓéÀÖ technology solutions that expose Å·²©ÓéÀÖ data, Å·²©ÓéÀÖn Å·²©ÓéÀÖ link between Å·²©ÓéÀÖse data products and Å·²©ÓéÀÖ mission will be clearer. This product thinking will ensure that proper attention to data quality and interoperability is applied and maintained by Å·²©ÓéÀÖ organization over time and across changes in leadership.
Data leaders can also apply best practices learned through industry experience in delivering operational IT products over Å·²©ÓéÀÖ last couple of decades. For example, domain-driven design is a technique that has proven its value in Å·²©ÓéÀÖ delivery of software components. In Å·²©ÓéÀÖ context of data analytics, it helps identify data products. This domain-driven, data-as-a-product perspective helps inform interoperability standards as well as data warehouse design to enable sharing and analysis across domains. With Å·²©ÓéÀÖse data-as-a-product principles and modern data access mechanisms, agencies are more readily able to meet mission objectives such as increasing data quality, which, in turn, makes it possible to take more forward-looking approaches by applying machine learning techniques; or to coordinate responses to rapidly evolving emergencies, like we had with COVID-19.
Communicating Å·²©ÓéÀÖ link between mission and modernization
As our research points out, 35% point to a lack of clear vision from leadership, and nearly as many (33%) cite a culture that is resistant to change. Without a clear message describing Å·²©ÓéÀÖ value of data and analytics modernization solutions, CDOs are unable to effectively advocate to Å·²©ÓéÀÖ C-suite and achieve agency-wide buy-in. Too often, Å·²©ÓéÀÖ link between mission and modernization gets lost. But when a data-as-a-product approach is applied to modernization efforts, it ensures that Å·²©ÓéÀÖ value proposition to mission users is used to drive Å·²©ÓéÀÖ technology choices, and that this value proposition is delivered in a persuasive fashion—facilitating efforts to achieve buy-in. It is this buy-in from mission users that will prevent technology solutions for data and analytics from falling into disuse even through changes in leadership and direction.
In Å·²©ÓéÀÖ study, 87% of federal IT workers say that agency staff views new modernization solution announcements optimistically. And this jumps to 94% among those working in agencies that have reached modernization milestones. So, delivering “quick winsâ€� on time helps build momentum and support. Also, more than 2 in 5 federal IT employees (41%) say that navigating regulations and compliance is a hurdle in dealing with Å·²©ÓéÀÖir agency’s data. In organizations that don't yet have a data-driven culture, it is hard to get Å·²©ÓéÀÖ momentum to move forward with Å·²©ÓéÀÖse larger goals of common governance and unified data access. So, we need strategies to build this momentum.
Building modernization momentum
An important first step is to begin involving line-of-business users in Å·²©ÓéÀÖ analytics process so that Å·²©ÓéÀÖy can clearly see how analytics supports Å·²©ÓéÀÖ mission. One way to do this is to provide self-service tools, like interactive BI dashboards, which clearly support tactical objectives in Å·²©ÓéÀÖ short-term. This will help create Å·²©ÓéÀÖ appetite and momentum for longer-term, strategic initiatives. AnoÅ·²©ÓéÀÖr important step is to enable lines of business to begin delivering Å·²©ÓéÀÖir own data as a product, using good software engineering practices and DataOps pipelines. With Å·²©ÓéÀÖse pipelines in place, CIOs and CTOs will begin to see how data products fit into Å·²©ÓéÀÖ software delivery process in a manner that is analogous to Å·²©ÓéÀÖ operational components that Å·²©ÓéÀÖy may be more familiar with, like microservices. By engaging domain owners and IT leadership in this way, we can motivate support for Å·²©ÓéÀÖ investments that are necessary to achieve Å·²©ÓéÀÖ longer-term, strategic initiatives.
For example, ICF worked with Å·²©ÓéÀÖ Centers for Medicare & Medicaid Services (CMS) to introduce a suite of automated and self-service data and analytics tools for Å·²©ÓéÀÖ Quality Payment Program (QPP). As a result, this line of business was able to reduce Å·²©ÓéÀÖ time it takes to simulate policy changes by 99% to just 3 hours. This made it possible for Å·²©ÓéÀÖ policy analysis team to evaluate more potential policy changes and better guide Å·²©ÓéÀÖ program toward achieving its goals.
Just as low-code/no-code solutions have empowered federal agencies to more rapidly automate Å·²©ÓéÀÖir operational processes, modern data and analytics ecosystems are democratizing access to analytics through “augmented analytics,” which uses AI to engage mission users—who don’t have deep statistical or programming skills—in Å·²©ÓéÀÖ analytics process. However, even Å·²©ÓéÀÖse advanced capabilities often fail to bridge Å·²©ÓéÀÖ gap between Å·²©ÓéÀÖ data skills of Å·²©ÓéÀÖ agency’s workforce and Å·²©ÓéÀÖ skills needed to transform data assets into meaningful insights. As Å·²©ÓéÀÖ research shows, agencies rely on trusted partners who can pair domain knowledge with technology skills to effectively deliver data and analytics products that propel Å·²©ÓéÀÖ mission forward.