Back in 2014the United States stood its ground small internal space race. NASA chose two companies to take us back to the International Space Station and provide them with funds to build a manned spacecraft: One Established, proven company for aerospace technology over many yearsand a Upstart run by a man who names all his children like they were Warhammer characters. Ten years later, the race has a clear loser: Boeing.
Boeing received $4.2 billion in federal funding to build its Starliner system, far more than the $2.6 billion SpaceX received for its Dragon. With over a century of aviation experience, it is by far the more established competitor, but Starliner is only now preparing for its first crewed flight – four years after Dragon’s and $1.5 billion over budget. Ars Technica I checked what went wrong:
But at least SpaceX was in its natural environment. Boeing’s space division had never received a large fixed-price contract. Its executives were used to operating in a cost-plus environment where Boeing could charge the government for all expenses and earn a fee. Cost overruns and delays were not the company’s problem, but NASA’s. Now Boeing had to deliver a flyable spacecraft at a fixed price.
Boeing struggled to adapt to this environment. When it came to complicated space projects, Boeing was used to spending other people’s money. Now every penny spent on Starliner meant one less profit (or ultimately greater losses). This meant that Boeing gave Starliner fewer resources than it needed to succeed.
…
With a fixed-price contract, a company receives compensation when it reaches certain milestones. Complete a software review? Earn payment. Are you proving to NASA that you built a spacecraft component as you promised? Earn payment. This type of contract structure naturally incentivized managers to achieve milestones.
The problem is that while a company may do something to unlock a payment, the underlying work may not actually be completed. It’s a bit like students copying homework throughout the semester. You get good grades, but you haven’t acquired all the knowledge you need to understand the material. This will only be determined during a final exam in class. So Boeing essentially carried forward the technical debt so that the additional work was focused on the final milestones.
Boeing, as we’ve all seen recently, is no longer an aerospace company – it’s a profit company, the kind of business you get when MBAs driven by “efficiency” replace engineers. This thinking did not fit well with a government mandate, and the so-called “efficiencies” never proved helpful. Ars Technica continues:
Boeing didn’t have a single flight software team. The responsibilities were distributed. A team at Kennedy Space Center in Florida handled the ground system software that kept the Starliner intact during ground testing and the countdown until the final minutes before launch. Separately, a team at Boeing’s facilities in Houston, near the Johnson Space Center, managed the flight software for the vehicle’s launch.
However, neither team trusted the other. When the ground software team visited their counterparts in Texas and vice versa, interactions were limited. In the end, the two teams largely worked in isolation and didn’t really share their work with each other. The software team in Florida concluded that the Texas team working on flight software had fallen behind, but would not admit it. (A Boeing spokesman denied that there was any such friction).
The full Ars Technica piece is worth a read as it delves into the countless ways in which Boeing failed on the Starliner project. Cost constraints, siled development, a lack of proper testing, and even the company’s lack of vertical integration all contributed to derailing the project. Flying into space is hard enough without making it as cheap as possible.