Bent Flyvgjberg and Atif Ansar have a new article providing evidence that SpaceX projects outperform historical and contemporary competitors by using a platform approach. The authors believe the same approach and success is available to government agencies. In some
… [P]rices for payload on a Falcon Heavy launch start as low as US$90 million — about 5 times cheaper than the Delta IV Heavy made by United Launch Alliance (ULA) — jointly owned by Boeing and Lockheed Martin. And less than a third of the price of its closest competitor, the Russian Proton family of rockets, which has been in service since the 1960s.
SpaceX also outperforms in terms of development cost. NASA (2010) verified SpaceX’s total development cost for the Falcon 9 rocket at approximately US$390 million (including US$90 million towards the Falcon 1 rocket) in year-of-expenditure prices 2002–2009. NASA’s 2010 report goes on to estimate “the predicted cost to develop the Falcon 9 if done by NASA would have been between $1.7 billion and $4.0 billion.” SpaceX’s Falcon 9 system was thus 4–10 times cheaper in development cost than NASA’s own estimate.
… SpaceX’s platform strategy also outperforms NASA in terms of cost adherence, measured as cost overrun on the final approved baseline budget. Figures 2a & 2b and Table 2 show cost overruns of NASA versus SpaceX’s missions.
Across these metrics, SpaceX’s platform strategy has vastly outperformed NASA’s bespoke strategy in some cases by multiple orders of magnitude.
SpaceX has a 10% average cost overrun, and 50% median. So their “extreme” values are actually cost savings, pulling the mean down. That’s really quite incredible. Not only are costs much lower than what government could achieve, but cost growth is lower than his public megaprojects dataset. So what’s going on?
Here are the authors making a distinction between “platform” development and “bespoke,” or “quantum leap” development paradigms:
Platforms differ from quantum leaps in two primary assumptions about the world. Quantum leapers believe, or pretend to believe:1) The world is a linear, controlled, environment (a la Hirschman’s “pseudo-imitation” technique);
2) Experts possess all the necessary information and skills to marshal the risks that may arise in the future (a la Hirschman’s “pseudo-comprehensive-program” technique). Platformers, in contrast, believe: 1) The world is a dynamic, complex adaptive system (Baldwin and Clark, 2000); 2). Future adjustments to risk in the external environment cannot be pre-ordained. Instead, carefully thought through options must be designed to flexibly respond to multiple probable worlds in the future (see McGrath, 1997; 2004 on “real options”).
This use of “platform” is is perhaps slightly at odds with how many folks in defense communities use it — which is one vertically integrated system that goes through block upgrades. And platform economics I think of as multi-sided players transacting on something like Facebook, Google, iPhone app store, etc.
Even though SpaceX is vertically integrated, they built and tested incrementally to get from their earliest rockets. Moreover, they share code base such that nine separate vehicles shared a relatively small staff of 50.
Leave a Reply