Air Force needs a data-driven acquisition supply chain

The recent confirmation of Frank Kendall as the Secretary of the Air Force is a hopeful sign that the Air Force’s much needed acquisition reforms will have the expertise and political weight to come to fruition. Secretary Kendall’s previous government experiences as the Undersecretary of Defense for Acquisition, Technology and Logistics from 2012-2017 paved the way of modernizing acquisition pathways and starting the trends of valuing the importance of big data and dual-use technologies. Yet, the Air Force has its work carved out. In order to achieve sufficient readiness level, especially in the backdrop of great power competition, structural changes are needed. Specifically, the Air Force should implement a data-driven approach to its acquisition and supply chain.

The current readiness level of the US Air Force (USAF) is dire. There are three major challenges. First, much of the unexpected maintenance is ad hoc rather than being tracked in a central database or repository. The most recent 2020 GAO report showed that out of the 46 types of aircraft from FY 2011-2019, only three met their annual mission capable goals. About 59 percent of those that did not meet mission goals was due to unexpected replacement of parts or repairs, stemming from a lack of technical data, delays in depot maintenance, unscheduled maintenance, and lack of trained personnel.

Second, the just-in-time manufacturing model for aerospace spare parts used by USAF is fundamentally misaligned with commercial contractors’ business models. The USAF buys old parts years apart, in small batches, on a needs-basis. This is not viable for commercial parts contractors to keep a production line open with labor and capital costs for potentially years without definitive orders.

Third, supplier availability and diversity are far between, especially those that make expensive parts that have infrequent demands but are key readiness drivers. One example is the lack of rotor blades for helicopters produced by Bell, Sikorsky, and Boeing. There is a limited number of suppliers — three to be exact — for a dizzying array of types of rotors for various aircrafts, but helicopters cannot function without the blades. These key issues are hindrances to USAF’s surge capability, especially in the broader context of great power competition and increasing instability in counter-terrorism areas of interest, such as Afghanistan.

The USAF has had limited success in using a data-driven approach to address these challenges. One project is the creation of automated inventory algorithms to make decisions for individual parts based on optimizing against overall metrics of materiel availability targets at minimal cost. While this approach works well for often-used parts, it vastly discriminates against expensive, infrequently demanded parts that are key readiness drivers, such as rotor blades. The algorithms themselves are fundamentally flawed and need to be updated based on advancements in machine learning.

Another way USAF is aiming to modernize its supply chain and acquisition processes is shifting from siloed Logistics Management Solutions towards sensor-enabled Supply Chain Enterprise Resource Planning (ERP) technology. According to a Govini report, “these ERP solutions create connected systems that bring greater asset visibility and tracking capabilities” with a 22.3 percent increase in annual growth from FY14-FY17.

Despite USAF’s efforts to address these critical issues, a more holistic data-driven approach is needed for the USAF to reach its full mission capabilities.

While the automated inventory algorithms deployed by DLA is a start to the integration of data, the algorithms need to be updated and adapted to including the efficacy of critical parts, even if they are infrequent. The current emphasis on efficiency rather than efficacy has led to the overlooking of critical parts. Machine learning algorithms, such as unsupervised learning, can be helpful in correcting these mistakes. Furthermore, predictive analytics can be used in ERP technologies to foresee maintenance issues as well as monitoring the status of critical parts. It can help achieve cost efficiencies by understanding exactly what DOD will need, where it is needed, and when.

Both these recommendations are underlaid by the necessity of having a central repository of information on maintenance schedules for every major aircraft/weapon system, details on what needs to be replaced, and having personnel available to do the maintenance. The repository not only could get rid of the current ad hoc approach, but it could also serve as training data sets for DLA algorithms and ERP technologies as a broader push for a data-driven approach to supply chain and acquisition.

Finally, having a database of suppliers of subcontractors and vendors in addition to the primes is critical to increasing diversity. Working with partners, such as NTIB and NATO, is another way to increase the number of trusted vendors.

USAF is starting its modernization process, and FY21 NDAA requirements are proof of that effort with an increase in budget for boosting its supply chain and acquisition risk assessment. But it needs a data-driven framework with the political weight behind its implementation. 

Evanna Hu is a nonresident senior fellow at the Atlantic Council’s Scowcroft Center for Strategy and Security. Additionally, she is the CEO and partner of Omelas, an artificial intelligence (AI) and machine learning company working on mapping the online information environment. She is a subject matter expert in messaging and propaganda of countering violent extremism (CVE) and counterterrorism (CT) in both Salafi-jihadism and neo-Nazism and has worked at the intersection of governance, security, and technology in Europe, the Middle East, and Africa, including extensive time in Kenya, Iraq, the Gaza Strip, Syria, Tunisia, and Afghanistan. Prior to Omelas, she successfully founded two technology ventures, one based in Nairobi, Kenya, and another in Amman, Jordan. To date, she has briefed six national heads of intelligence and has advised 12 cabinet or ministerial members on technology and security. At the Atlantic Council, she specializes in emerging technologies for NATO and member countries with a focus on AI and 5G.