This morning as I was driving into work, I listened to an interview on NPR with Peter Goelz, a former managing director of the National Transportation Safety Board (NTSB). The interviewer asked Mr. Goelz why the Federal Aviation Administration waited to ground the Boeing 737 MAX 8, when major European nations had done so three days prior. He answered that the FAA is a data-driven organization and there was not enough data to make that decision. He stated that those European regulators and other commentators calling for action were acting solely on conjecture. Yesterday, there was enough data – in the form of a recreated flight profile that was similar to the previous crash – to issue the grounding order.
This statement regarding the use of data, where hundreds of thousands of lives are at potential (fatal) risk, made me think: is data always the answer? When would it not be? Is this the best use of data? As a member of a group advocating for data-driven logistics (D2L), this is not an easy position to consider as we are already sledding uphill. However, the military profession, and specifically logistics in MCDP-4, is both art and science. Our D2L team argues (correctly, I might add!) that we’ve over-weighted the art aspect – the proverbial gut feeling of experience – at the expense of the scientific component.
So back to the question: when has ignoring the data been appropriate? The most probable Marine Corps answer would be Inchon. Against the advice of his Navy and Marine Corps advisors, General MacArthur ordered Operation CHROMITE and the amphibious operation was an overwhelming success in changing the course of the Korean War. I would speculate that the operation would not have modelled and simulated success, if those tools were available. So what made General MacArthur eminently qualified to overrule the advice of his subject matter experts and theoretical modeling/simulation?
Unfortunately, in his case, it was simply hubris as later proved by his ignoring of significant intelligence data about Chinese movements into Korea and their upcoming offensive against UN forces. Lately, the Marine Corps has been operating under the same hubris even though circumstances – and our unchallenged military advantage – have changed. So what would have been ideal? Ultimately, the ideal MacArthur would have been a well-educated officer in both the military arts and the data sciences who could have understood the underlying assumptions within the data analysis and the cognitive biases affecting both his and his advisors’ judgment, but with a deep reservoir of military judgment from personal experience and study of military history.
This archetype would be driven by data, but not encumbered by it. An organization staffed with this type of personnel could balance the data – the Boeing 737 MAX 8 has thousands of successful flights with only two crashes – against the anecdotal similarity of the incidents to determine that something isn’t right and issue the grounding order. Data, artificial intelligence, algorithms, machine learning, predictive analytics…we must leverage these tools as part of the future of warfare or we will be left behind. While we should always strive to make decisions supported by data, we can’t fail to make decisions in the absence of it. As logisticians and Marines, we must educate ourselves on how we can use these tools, but recognize that while it is certainly an aspect of the solution set, it’s not the only answer.
Great analogy to bring home the point that data is a tool. Surveys from 1,500 global corporations show 14 percent of business data is critical, while 32 percent is obsolete, and 54 percent is not yet understood. The human element is still needed to decipher the outputs and create insight. Great blog post.