In 1937, Robert McNamara graduated from university with a BA in economics, maths, & philosophy: a dangerous combination.
He then got an MBA from Harvard and taught accounting; he became their highest-paid and youngest assistant professor.
In 1943, he joined the Army, teaching analytical business approaches, to soldiers.
Then he became a colonel in the Air Force at the Office of Statistical Control.
In 1946, he joined Ford, running planning, and management control systems.
He led Ford in adopting computers; what he called ‘Scientific Management’.
He was so good at it he became President of Ford.
He was a brilliant man who knew the answer to everything lay solely in data analysis.
In 1960 he was recruited by President Kennedy to be Secretary of Defence.
Via computer modelling, he did the job by spreadsheets, graphs, and trends.
His data convinced him that the war in Viet Nam was going well.
In 1962, he said “Every quantitative measurement shows me we are winning the war.”
According to McNamara’s projections the war would be over by 1964.
But by 1965, the war wasn’t over, in fact America was losing.
McNamara’s data analysis consistently showed the only way to win was for America to escalate the war.
So much so, that in government it became known as ‘McNamara’s War’.
He knew the metrics wouldn’t lie, so he decided the way to prove they were winning was by a superior body count: kill more of the enemy.
So ‘body count’ became the most important metric on the spreadsheet, that would prove that the US was winning, numbers don’t lie.
And so the main goal of every military unit was to achieve the highest body count.
But McNamara’s quantitative style, based on number-crunching by computers, missed the human dimension.
General William Peers wrote: “With improper leadership, ‘body count’ could create competition between units, particularly if these statistics were compared like baseball standings and there were no stringent requirements as to how and by whom the counts were to be made”.
The obsession with body counts meant promotions for officers with larger numbers, which led to exaggeration of enemy losses.
Soldiers would claim kills that hadn’t been made, officers would inflate those numbers, and so on up the chain of command.
If the top brass just wanted numbers, they’d get numbers.
So the data was useless because it was based on lies.
But whatever the evidence of common sense, McNamara wouldn’t be swayed from his belief in statistics and data.
Senator Richard Russell said “McNamara is the smartest fellow any of us know, but he’s opinionated as hell and he’s made up his mind.”
Eventually, even McNamara admitted the models and stats he attached such importance to were “grossly in error” and he was “now pretty well convinced that our present policy can lead only to disastrous defeat”.
But that’s what happens when we have the arrogance to depend purely on data and computers to tell us how the world will behave.
Because people will always game the system, any system.
So when we create a system, we are creating something to be gamed.
And if it can be gamed, it will be gamed.
We need to learn, we mustn’t confuse algorithms with brains.