Why failures are important

In recent weeks, a video from SpaceX has been circulating on the internet documenting the many failures of this space company. The video is a good opportunity to discuss the importance of failures in systems engineering. While we are ultimately interested in success, real innovation is only possible if failures are dealt with properly.

SpaceX: Really a failure?

Let's be honest: SpaceX would never publish a video with failures if the company wasn't incredibly successful in reality. Here are a few statistics (Source):

• SpaceX receives more than 50% of the global commercial launch orders put out to tender
• In 2017, SpaceX has so far more launches than any other launch provider in the world, with a 1001TP3 success rate
• SpaceX has just completed its 11th successful ISS resupply mission, making it by far the world's the most important supply company of the western ISS partners
• And finally: For the first ten years of its existence, SpaceX has spent less than a billion US dollars, or about as much as the Elbphilharmonie has cost.

However, the video is a clear signal that innovation cannot be realized without failures. Especially when speed is of the essence. The timeline in the video's subtitles is also significant: the first crash took place in September 2013, the first successful landing on land in December 2015, a good two years later.

Failures in systems engineering

Systems engineering is usually about development: either a further development or a new development, but in both cases something new is created. And many project managers know the saying derived from the military: no plan survives contact with reality. Contact with reality must therefore be established. This is done - as with SpaceX - via prototypes

No plan survives contact with reality [tweetthis]No plan survives contact with reality[/tweetthis]

Of course, using these prototypes productively reduces costs enormously. Because if there is no failure, then the project is already co-financed. In practice, this approach has already gone so far that people feel honored when they are allowed to work with prototypes: Just think of the pilot programs for Google Glass or Oculus Riftwhere developers were allowed to apply to buy prototypes for expensive money.

Agility

These trends also explain why agile systems engineering The pressure to innovate is enormous, and in order to really learn from prototypes, a correspondingly large number of prototypes have to be developed and produced.

Agile systems engineering is not new, by the way, as I wrote about its use back in 1973. Nevertheless, a lot has accelerated since then, and with the right tools, efficiency can be drastically increased, as I wrote in the article Agile requirements management with Jama. Incidentally, SpaceX also uses Jama.

Incidentally, there are now also voices saying that too few failures indicate that there is not enough innovation in the game. However, this view should not be taken too far, as the following video shows.

Fail Harder

Conclusion

Ultimately, failures have to happen in the right context. A large-scale recall certainly doesn't fall into that category. But if the failures are seen as the results of experiments, then it starts to make sense. This includes planning for failure from the outset. Surely SpaceX will have appropriate insurance policies in place to deal with the failures. At the same time, they would (hopefully) not consider a rocket launch with humans on board an experiment: By then, the prototype stage should be complete.

None of this is new. Henry Ford already knew this when an employee came into his office meekly because he had destroyed an expensive test engine. "Am I fired now?" the employee is said to have asked. Henry Ford replied: "Why should I fire you now, after I've just invested a small fortune in your training?" That's what I call a confident way of dealing with failure.

Image source: SpaceX/YouTube