What Do Lean Startups Have to Teach the Department of Defense?
Like many observers in the military, I have been cheering the Department of Defense’s growing interaction with Silicon Valley. I am hopeful that the exciting culture of innovation that drives much of our country’s technological progress will inspire meaningful improvement in Department processes.
Aside from astronomical home prices, hooded sweatshirts, and millennial executives, I am sure the personnel manning the newly christened Defense Innovation Unit Experimental (DIUX) will discover the Valley’s widespread devotion to the “lean startup” model. Explored by Eric Ries in his seminal 2011 book, The Lean Startup, the concept continues to be a dominant model of operations for tech startups seeking to integrate cutting edge technology quickly and effectively into American life.
The defense acquisition process, widely criticized for chronic inefficiency in fielding new weapons systems, can be improved by incorporating the basic tenets of this model. It offers a solution to a massive efficiency barrier for defense programs — effectively leveraging technological innovation into useful products for users in a timely manner.
All War on the Rocks readers are intimately familiar with major programs that take so long to develop and deliver that the technology is nearly obsolete by the time the finished products reach the hands of warfighters, often decades after initial design. Prioritizing speed to capture a first mover advantage, the “lean startup” model emphasizes the rapid development and deployment of a minimum viable product — a basic version of the product than can quickly be redesigned and improved — to targeted customers for initial user testing. Testing generates feedback in the form of detailed metrics to help the company learn how to improve the product in a quick iteration cycle. This loop, “build, measure, learn,” can help future defense programs avoid some of the endemic headaches caused by the current systems engineering-based process.
Admittedly, the Department’s preferred systems engineering process, which has been the basis of the acquisition system over the past many decades, has proven its value in organizing and tracking increasingly massive and complex defense programs. However, timeliness has become a serious problem, with most new systems being measured in decades from the definition of requirements to actual operational capability. One fundamental problem with the process is the very first step: defining capability requirements. Beginning a program with defined end-state requirements can be a sound strategy, but it is also the exact point where many programs have sealed their own fate to produce something obsolete.
A perfect modern-day example is the integration of unmanned systems. Over the past decade, there have been countless numbers of strategy meetings, think tank papers, and institutional studies to determine how the U.S. military should utilize unmanned systems in the future to maximize advantage. However, the inability to predict how the technology will ultimately be most useful in the coming decades has made defining exact requirements a huge challenge. The inability to reach consensus on requirements further delays getting the systems in the hands of warfighters. There are simply too many unknowns, and commanders have not had the ability to experiment with the technology to unlock its value and see how unmanned systems might best be used. We know that unmanned systems are a revolutionary technology — we just don’t know yet what ultimate utility the technology will provide.
With that in mind, instead of trying to define particular requirements for a major unmanned system that will be fielded in the 2020s, the Department of Defense should build, measure, and learn from multiple prototypes to determine what utility warfare commanders can derive from the technology. Producing different types of unmanned systems, the equivalent of minimum viable products, and getting them to warfighters during large training exercises will allow users to employ and experiment with the technology, determine what works and doesn’t work, and return feedback to designers to iterate quickly on future models. This process will lead to the rapid deployment of a workable product in the field and improve over time to become an optimal finished product.
To this end, defense reps can take advantage of the talent present in the Valley by launching a design competition, soliciting inputs from the pool of hardware and software engineers in the area to capitalize on their ingenuity and skill. One only needs to take a look at the technology used by amateur unmanned aircraft operators to realize the sophistication present among the ranks of this “maker community.” For instance, homemade unmanned aircraft regularly incorporate 10 Degree of Freedom Inertial Measurement Units (IMUs) — with 3-axis accelerometer, magnetometer, and gyroscope, and a barometric pressure sensor — on par with many manned aircraft.
Awarding a small contract to promising designs from a competition can produce prototypes with which commanders can experiment during exercises. Based on a real-time feedback loop with designers, commanders can initiate the build–measure–learn cycle. From a small dollar investment, with consistent iteration, this venture could harness the best of American ingenuity, maximize warfighting capability, and ultimately inform the picture of what the best use of unmanned aircraft will be.
This rapid development and iteration is not a new phenomenon. In the Jet Age, aircraft designers built, tested, and produced jets at an incredibly rapid rate, partly because there was an intense race with the Soviet Union to gain a first mover advantage.
One classic example is the revolutionary U-2 aircraft that, legend has it, was designed by famed Skunk Works pioneer Kelly Johnson on the back of a bar napkin (an epic War on the Rocks article waiting to be written, surely). Based on a Request for Proposals in 1953, the first U-2 flight occurred in 1955, and the jet was flying missions one year later in what can be considered a minimum viable product role. Over the 60 years since, iteration of the aircraft design has yielded 18 different variants, including a version capable of launching from aircraft carriers in the 1960s — and the aircraft is still operational.
When you walk through an aviation museum, there are hundreds of different types and models of aircraft from this era, each with very different aspects of design. You can literally trace the development of new technology by looking at new aircraft — supersonic fighters, swept wings, radar, air-to-air weapons. These aircraft had relatively low unit cost compared to fighters today, and designers had the capacity to quickly build hundreds of each type, iterate on the design, create a new variant incorporating new tech, and produce hundreds more. Another example is the F-4 Phantom II aircraft that was prototyped in the spring of 1956 and operational in 1960. Over the decades that followed, through the Persian Gulf War of 1991, over 20 variants of the F-4 were operational, each iteration incorporating new technology to adapt the baseline product to the modern needs of the warfighter. Compared to the incredibly high unit cost and relatively inflexible design of systems today, the lean startup concept proves to be an interesting alternative for how to approach fielding new systems.
Ultimately, I am optimistic about the foray of defense leadership into Silicon Valley. Like nowhere else on Earth, the culture of innovation, progress, and creating solutions is tangible there, and I am excited to see how the DUIX capitalizes on the insights to be gained. Above all, I am hopeful that some of the lessons learned will inform how we can hasten the transfer of new technology from its completion in the design process to the hands of warfighters for practical use.
Jeff McLean is a naval aviator and a member of the Editorial Board at the U.S. Naval Institute. He also currently serves as a Millennium Leadership Fellow with the Atlantic Council. The opinions and views expressed are those of the author alone and do not represent the views of the U.S. Department of Defense, the U.S. Navy, or any other agency.
Photo credit: Secretary of Defense
