While industry experts say the strategy lays a much-needed foundation, the devil’s in the details when it comes to implementation.
The Defense Department’s first enterprisewide data strategy, released in October, is succinct, coming in at less than 20 pages. But many defense observers say the strategy plants a clear and necessary stake in the ground declaring the importance of good data management at the Pentagon.
The road to realizing the vision isn’t free of challenges, though, and details regarding how the strategy will be implemented are just starting to emerge. Still, experts suggest DOD stay the course.
Tara Murphy Dougherty, chief executive officer of Govini, a decision science firm supporting the Pentagon and the defense industry, told Nextgov in a recent conversation the foundation of the strategy is “quite sound.”
“I really would call upon the incoming Biden administration to adopt the data strategy and run with it,” Dougherty said.
“So much of what is in it is right and spot on from a needs perspective, they're better served by defining the ‘how’ than going back and re-litigating any of the first principles, which all data professionals would tell you are quite sound,” she added.
How the Strategy Came to Be
The Pentagon’s first enterprisewide data strategy breaks from previous departmental guidance on data by changing the focus to center warfighters and senior leaders who need to make better decisions, Chief Data Officer David Spirk said at a National Defense Industrial Association event in October.
The DOD Net-Centric Data Strategy, signed in 2003, prescribed an “outside in” view of the data landscape, according to Spirk, but this new strategy means data requirements will now flow from decision-makers.
Development of the new data strategy preceded Spirk’s appointment to the CDO position, a role he assumed in June. Spirk told the audience during ThoughtSpot’s Beyond.2020 Digital event in December that the department struggled to complete the strategy for a couple of reasons, including the sheer size of the agency as well as disagreements on what the focus of an enterprise-level data strategy should be.
“There were some that thought the business analytics component was the most important, there were some that thought the warfighting business was the most important component,” Spirk said.
Spirk said DOD fostered a “community of interest” around the development of the strategy in order to work out some of these foundational questions. The document grew, incorporating more priorities and interests until at one point it encompassed 75 pages, he said.
Eventually, the community of interest settled around an emphasis on operational advantage, Spirk said. This consensus is what allowed the department to finally finish the strategy after Spirk assumed the CDO position, he added.
“We will pay for operational advantage,” Spirk said. “The good news about doing data is, if you do it right, the efficiency naturally follows.”
But what does doing data right look like? According to the strategy, it means making data visible, accessible, understandable, linked, trustworthy, interoperable and secure.
Defense Digital Services Director Brett Goldstein, who is the former CDO of the City of Chicago, said during a Center for New American Security webinar he collaborated closely with Spirk on developing some of the strategy’s core concepts. For Goldstein, doing data right means getting back to the basics.
“I like to introduce this concept of what I call ‘no new bad,’” Goldstein told Nextgov in November. “Historical systems are what they are. But starting tomorrow, let’s make sure we’re architecting things for a way that allows for integration. Historically, we spent a lot of time thinking about things in the service or the component level. Instead, we need to think about that holistically.”
In an email to Nextgov, Spirk shared some of the steps the department has taken since October to strengthen data management. This month, Deputy Secretary of Defense David Norquist, the official who signed out the data strategy, issued department-wide guidance around several data commandments, Spirk said.
“First, he directed that data be published in data catalogues using common interfaces or APIs,” Spirk said. “Second, that the Department employ open industry-standards to the maximum extent practical. Third, that data be managed and stored in a portable manner not dependent on particular infrastructures. Fourth, that data be protected with robust cybersecurity and access control. Additionally, this is the first Department-wide guidance directing that we will move to an open data standard architecture.”
Spirk himself also issued guidance on federated data cataloging for all levels of classification and has drafted guidance on reviewing public data releases, protecting sensitive datasets, and reporting data purchases. He also recently hosted the first meeting with CDOs from the Five Eyes intelligence alliance—which includes Australia, Canada, New Zealand, the United Kingdom and the U.S.—as well as the inaugural meeting of data leaders from the combatant commands.
“There is clearly a shared urgency across the Department, with our partners in the federal government and allied nations, to address data issues in a collaborative and transparent manner,” Spirk said. “It’s been great to see their interest grow in applying the lessons learned from DoD’s data journey and I look forward to working with all our partners as we drive forward on this effort.”
Where to Go from Here
One of the foremost reasons some industry experts are encouraged by the data strategy is because it lays a clear groundwork without getting muddled by competing priorities.
“These things can go lots of different ways,” Dougherty said. “They can often get diluted through the coordination process internal to the Pentagon … but [the strategy] really didn't dilute the core message, and that is so badly what this space needs.”
But while the strategy is a good first step, it’s just that: a first step.
Sonny Hashmi, managing director for global government at cloud company Box, told Nextgov ensuring decision-making and design work are as distributed as possible is key for successfully implementing the strategy. Hashmi is the former chief information officer for the General Services Administration, where he helped implement a data strategy for the organization.
“What that means is that to the extent that you can be closer to the mission, to the edge where commanders, people in the field, different mission owners can actually define the taxonomies the data standards for themselves while working within an overarching framework that's been defined at the headquarters level, at the Pentagon level, that's going to be successful,” Hashmi said.
Hashmi added it’s impossible to define every single taxonomy layer, or data definition, or data access rule in a central way for an organization as large as DOD, so distributing decision-making will be critical.
An immediate next step for the department would be to identify some low-hanging fruit where having access to better data in real time would make a clear and measurable difference, Hashmi said. The data strategy in isolation is not the goal, Hashmi said; instead, everything must drive toward specific mission outcome benefits. And those benefits should be showcased, he added.
Chad Cisco, general manager for federal business at DataRobot, an artificial intelligence company that works with DOD, also told Nextgov that many of the challenges described in the data strategy have already been solved by commercial industry—sometimes “many times over.”
“There's just so many companies out there that can solve a lot of these problems for the DOD that I think it is quite important, that efforts like the [Defense Innovation Unit] get expanded and they really find ways to bring in the nontraditional kind of defense contractors to provide technology that's already been built,” Cisco said.
Leaning on industry may be important particularly because DOD lags behind industry on advancements with regard to data. Adapting commercial products to fit DOD requirements can help the department catch up quickly as long as security and compliance measures are taken care of, according to Hashmi.
Ultimately, Hashmi sees the data strategy as an ongoing project without a clear finish line. The journey matters more than any specific destination, he said.
“In fact, the more effort you put into it, the more improvements you see, the more opportunities start identifying themselves and where you can take it even further,” Hashmi said.
NEXT STORY: Wastewater Testing Tracks COVID Infection Trends