November 23, 2011 by startupengineering .
Yesterday was the anniversary of the Kennedy assassination, and Errol Morris marked the occasion by releasing a six-minute documentary about the umbrella man.
The umbrella man was a guy who was standing along the route that John F. Kennedy’s motorcade took in Dallas, holding a black umbrella over his head. It was a perfectly sunny day, and of the thousands of people lining the route, he was apparently the only one holding an umbrella. Kennedy was shot just as he passed the umbrella man, giving rise to multiple conspiracy theories. The indelible, incontrovertible, pregnant fact of one strange figure, standing with his umbrella open on a sunny day, just as the fateful shots were fired, proved that something suspicious was going on.
But in Congressional hearings five years later, the umbrella man came forward. It turned out that he had been protesting Kennedy’s father, Joseph Kennedy’s involvement in Neville Chamberlain’s appeasement policies toward Hitler before World War Two. The umbrella was a reference to Chamberlain’s famous umbrella. Wild theories that linked the umbrella man to the assassination were wrong, and the chain of events that had brought him out there with his umbrella was beyond the reach of any plausible theory.
Founders go wrong by depending on inductive reasoning to sound out intent. It’s natural to assume that the umbrella man was intending something sinister, because he was doing something odd, and something sinister happened. But inferring intent from behavior is perilous, because there’s an infinite supply of possible intentions and it’s natural to choose one based on how you already see the situation.
If Kennedy had NOT been assassinated, then it’s plausible to think that passers-by might have seen the umbrella man and recognized the reference to Chamberlain and appeasement. But the glaring light of the assassination made that connection invisible.
In the same way, the glaring light of a founder’s idea makes it extremely difficult to identify (or easy to mis-identify) the intent behind customer behaviors.
Instead of inductive reasoning, try effectual reasoning. Imagine the change you want your product to create in customer behaviors. Start with a thought experiment: Is it possible that the existence of my product would lead people to act in this way? How would they go about it? What might prevent them? You can eliminate lots of ideas that just won’t work simply by approaching this rigorously and thinking it through.
Next, start constructing “out of the building” experiments. One way to think of these experiments is through the lens of minimum viable product. “Viable” means “functionally able to test the hypotheses you’re using it to test.” “Minimum” means “don’t spend money designing in stuff that you don’t need specifically to test the hypothesis.” The classic example is using a landing page instead of a whole built-out website to gauge interest.
Errol Morris’s film about the umbrella man talks about one guy who painstakingly designed an elaborate flechette-shooting umbrella to show how the man shot Kennedy. But unlike conspiracy theorists, founders don’t have all the time (or resources) in the world.