As the first chief data officer of the Department of Defense (DoD), Michael Conlin described the challenges of approaching a vast mountain of data from numerous systems at the Pentagon, and making it actionable information for the department.
Speaking Wednesday at FCW’s Citizen Engagement Summit, Conlin addressed the mandates he saw as most important for his work. Conlin noted that the National Defense Strategy (NDS) sets the goals of reforming the defense business enterprise. As a complement, the National Defense Authorization Act (NDAA) pushes an enterprise view of DoD, pushing the department to take an enterprise-wide data approach and get a common view of the service branches.
While his focus on the National Defense Strategy and the NDAA were expected, Conlin pointed to the President’s Management Agenda (PMA) as the final policy that he finds most important.
“Congress said to enhance department-wide operations and management, now a specific target from the NDS–improved affordability and performance. How do you know when you’ve improved it enough? Well the answer is in the President’s Management Agenda,” he said. “It says you will be as efficient and productive as an equivalent commercial-sector organization.”
With his mandate in place, Conlin described the next challenge–what is common enterprise data?
“The NDAA said common enterprise data is data having to do with the operations and management of the defense business enterprise,” he noted. But with two systems of record and over 1800 registered business systems discovered across the department, getting permission from each system owner and gathering all of it together would be a Herculean task.
So Conlin took a different approach.
“You want to come at this one data story at a time. Know the question you’re trying to answer, and define just enough data architecture to answer that question,” said Conlin.
In the face of calls to build out a big data analytics platform to begin, or compile a massive data lake, Conlin noted that you have to carefully pick source systems for data, keep track of the data’s provenance, and work off of the hypotheses you want to test.
Detailing some success stories, he noted that they had built a repository of common enterprise data being used by decisionmakers. He said that they have baselined about 75 percent of DoD’s budget to understand costs and how they compare to the commercial sector. With the recent DoD audit, the repository tracks all the findings and remediation efforts on financial info, and the department is working to mature from affordability data to performance data.
Conlin described the challenge of creating a data-driven culture, and admitted that the Pentagon is not there yet.
“We do not have a culture of data-centric decision making in the department. We have a culture of experience-centric decision making,” he said. Conlin noted that his goal is to persuade colleagues to embrace data-centric decisions, and that the department was making “good progress” on the issue.
Finally, he addressed the issue of workforce, calling it his biggest challenge. He described how the combination of computer science, math, and domain knowledge is very rare, and something the department needs more of.