Tech alone isn’t enough: how data-driven methods lead to transformational success.
With agency partners and private industry, components across the Department of Homeland Security (DHS), including the Transportation Security Administration (TSA), U.S. Customs and Border Protection, and Cybersecurity and Infrastructure Security Agency, are implementing the next generation of technologies in security and efficiency. While these upgrades are critical, technology alone isn’t enough; DHS components have an urgent need to adopt data-driven methods for a seamless and effective rollout. Our LMI experts share key takeaways for technological transformation success.
Create a Baseline for Data
When organizations focus too much on individual metrics, they often fail to discern the broader vision of technology implementation, hindering success. Although a technological transformation can save time, improve resource management, and enhance service delivery, these results come from and are sustained by achieving the goals of the implementation system-wide. Zain Malik, a principal systems engineer at LMI, says that, as organizations transform their technology and become data-driven, they must set a comprehensive baseline of operations to avoid narrow metrics that don't tell the whole story.
"Organizations must start with the goal in mind, or at the least, work against metrics that they already collect and don't always lead to an accurate understanding of the problem," he explains. "They must examine the complete production environments to understand where the silos are, what data is already coming in, and where the gaps are. By evaluating across an organization, often organizations find that they already have some of the information that they thought they were missing and can craft a better problem or ideal solution set.”
Creating a baseline of operations, capabilities, and data saves time up front. Project teams can map technological enhancements from that baseline and select the underlying capabilities and operations to improve. Attempting to improve one or two specific metrics can lead to unintended consequences or failed projects because changing the metrics may not solve the root problem: the solution needs to be holistic.
Steve Pak, vice president of data operations at LMI, adds that creating a baseline helps organizations understand early on where they lack data maturity. Pak explains: "Organizations take note of analytics and emerging technologies and say 'Yes, this is how we will fix our problems,' but can only leapfrog so far. If an organization doesn't have good data to put into these systems, it can't do the work expected. Getting to data maturity requires understanding the data you have, whether it needs to be standardized, and what data still needs to be collected to use data analytics technology to improve operational outcomes."
“Organizations must start with the goal in mind, or at the least, work against metrics that they already collect and don't always lead to an accurate understanding of the problem.”
Zain MalikPrincipal Systems Engineer
Use Synthetic Data Modeling to Test Technology
Once organizations have a baseline and the data to understand the problem, synthetic data modeling can uncover solutions. For the government, testing new technologies or data solutions in the field is helpful but high risk. Testing with synthetic data modeling prior to piloting in the field saves time and resources.
"Homeland security components like TSA must respond to many threats. Often, they seek emerging technologies to manage those threats, but those technologies must be adequately tested before implementation," says Brant Horio, a data and computational sciences fellow at LMI. "With synthetic data, you model how well technology responds to changes in threats and understand the resources required to respond ahead of time."
Synthetic data modeling also assists with adversarial insights—learnings that hackers might already be trying to develop about technology systems. "If you have artificial intelligence trained to seek certain things, you understand what is important to the system. Then, you can start thinking of ways to beat that system, which adversaries are already doing," he says. "Now, you've got a sense for potential exploits before they happen and create security problems. That's very valuable data."
“With synthetic data, you model how well technology responds to changes in threats and understand the resources required to respond ahead of time.”
Brant HorioData and Computational Sciences Fellow
Build on Incremental Success
Synthetic data modeling supports project teams as they create pilots or incremental technology prototypes for the technological transformation. Keith Rodgers, vice president of advanced analytics & AI at LMI, says these data help organizations understand a project’s potential and can ultimately lead to the long-term internal support necessary for success.
"There is a tendency with organizations that, if a pilot or small project fails, the broader project gets abandoned. That's not ultimately helpful to achieving big goals like technology transformation," Rodgers says. When organizations discover potential through insights from data analytics, then project teams have wider latitude to build pilots and prototypes (which may need refinements over time) without a setback that tanks the whole project.
According to Rodgers, building an integrated project team that includes people from across an organization makes technology transformation more likely to succeed. If this team creates smaller project prototypes that can fail, the lessons learned from those small projects lead to a solution that works better. "Building on incremental successes and failures is how big projects move forward, and you can do that in a controlled way so you aren't negatively impacting operations in the field," he says.
Paul Dufresne, a senior analyst at LMI, agrees. "Often, the technology tool works, but not in the way an organization needs. If you have representatives on a project team from all parts of an organization who can stand up and say, 'This adds too much time to our process' or 'We can't use this because it doesn't solve for x,' then you can make changes before your frontline people have to implement it," he explains.
Manuel Galvan, a program manager at LMI, adds that organizations using an integrated project team construct help ensure that resources for leveraging existing enterprise solutions and a project’s effort don’t get wasted. “Large organizations tend to silo,” he explains. “Sometimes one group purchases a technology that was previously enabled by another group but isn’t being leveraged as an information technology (IT) service or capability by others. When a project team works across silos, they uncover where enterprise or services already exist and can exploit those capabilities to create efficiencies. These efforts multiply the return on IT investments throughout entire organizations.”
“Building on incremental successes and failures is how big projects move forward, and you can do that in a controlled way so you aren't negatively impacting operations in the field.”
Keith RodgersVice President, Advanced Analytics & AI
For homeland security organizations that respond to constantly changing threats, emerging technologies improve overall response capability and help manage resources more effectively. However, simply putting new technology in the field isn't enough. Stepping back and creating a baseline for the agency helps project teams understand which technologies to use. Synthetic data models the effectiveness of those technologies before fielding. Creating an internal project team with the latitude to test these technologies and make changes if they fail ensures implementation of only the best solutions. Together, these steps greatly increase success for technology transformation projects.
For nearly two decades, LMI has provided services to agencies throughout the Department of Homeland Security. From acquisition and financial management support for U.S. Customs and Border Protection and laboratory services for the department's Science and Technology directorate to program management for the Federal Emergency Management Agency and beyond, LMI is committed to ensuring homeland security mission success for our nation.
Brant HorioSr. Fellow, Applied Research & Partnerships Meet Brant
Brant HorioSr. Fellow, Applied Research & Partnerships
Brant Horio leads and develops a large team of data scientists, operations researchers, and computational modelers. He heads the development of capabilities and data science strategy for the evolving technology landscape of artificial intelligence, machine learning, cloud, DevOps, ModelOps, as-a-service technologies, and the democratization of data science.