When former U.S. Defense Secretary Robert McNamara died last week, much was made of his quantitative abilities — the success he had a Ford and later the miscalculations made in the escalation of the war in Vietnam that stemmed, in part, from his reliance on quantitative analysis. Here is a brief quote from the obituary in the Boston Globe:

“’Every quantitative measurement we have shows that we’re winning this war,’ he said. It was a telling response from someone who would be presiding over a struggle in which, as the United States came to learn, hearts and minds did more to determine the outcome than body counts or bomb tonnage.

It was also a characteristic response from a man who saw number-crunching as equal parts contact sport and higher calling. Give him enough statistics to analyze, Mr. McNamara seemed to believe, and almost any problem might be solved. ‘You can’t substitute emotion for reason,’ he liked to say.

This quantitative bent was the foundation of his extraordinary certitude, a certitude critics called inflexibility. No one was more can-do, no one more gung-ho. Mr. McNamara brought an almost-missionary zeal to problem-solving. ‘I would rather have a wrong decision made than no decision at all,’ he once said. It was an emblematic statement.”

The argument can be made that he was simply looking at the wrong data, of course, and that it wasn’t analysis per se that caused his miscalculations.

Another possible root of the limits of McNamara’s abilities is advanced by Lane Wallace in The Atlantic: He ignored rhetoric — the second path to the truth, along with logic, advocated by Aristotle.

“When it comes to planning for the future, or making decisions in domains where things can be “other than they are,” Aristotle believed rhetoric was far more useful than analysis. ‘Humans have never predicted the future by analyzing it,’ Golsby-Smith says. Designing effective strategies for the future–especially in areas involving potentially irrational human actions and reactions–requires imagining various scenarios and perspectives on the truth, and then making judgments based on the persuasiveness of each one. “

The lessons for analytics professionals are many: Numbers alone aren’t always the answer Emotion can never be fully removed from any process in which humans are involved and thus should temper our certitude in our analysis. One needs to push for the right data, not just the data that is easy to collect (hearts and minds vs. body counts and bombs).  One should use equal fervor in exploring what we don’t know as we exert on what we do know (McNamara’s “every quantitative measure we have” obviously wasn’t sufficient). Finally, one can never go wrong with a little humility. New York Times columnist James Reston, quoted in the Globe obit summed it up well in 1966:

“He is tidy, he is confident, he has the sincerity of an Old Testament prophet, but something is missing: some element of personal doubt, some respect for human weakness, some knowledge of history.”

Even the most astute analyst will get some of the calls wrong. Grasping one’s own fallibility can both help prevent some of those errors and help you survive the ones you do make.