Running Experiments

The logic concerning uncertainty described in the “Disciplined Entrepreneur” incorporates a robust understanding of statistics, and the pitfalls of human intuition.  The authors describe that a normal distribution of outcomes is often assumed, it is what we learn in school after all.  However, they caution that the distribution of outcomes is more likely highly irregular, and painfully opaque.  Business leaders sitting around a board room can go around in circles, describing what they believe is the probability of the risk they face.  They may have good reasons for their assumptions, logic and experience on their side, and a solid consensus of their peers.  However, these may mean nothing if they are encountering an unfamiliar risk.  Human simply are not good at anticipating levels of risk and uncertainty intuitively, we are full of biases and heuristics that cloud our judgment.  We must rely on other data driven models or natural experiments that can inform our decisions with more accuracy.

I was intrigued reading about the experience of Noodles and Co.  It is unlikely that keen and experienced business leaders could draw up reasons why their restaurant in Madison failed and their restaurant in Denver did not.  Even if there were such a priori opinions, there would likely be equally compelling arguments why the restaurant in Madison would succeed.  There are simply far too many variables at play, some weigh heavily on an outcome, while some have no effect.  Only experiments have the power to uncover how the real world works.  The restaurant in Denver succeeded.  Noodles and Co. now know that fact, and they should be encouraged to conduct further experiments in other cities.  Once they have more data, they might be able to reconstruct the factors that make their restaurants successful, and adjust their business model accordingly.  Business leaders and entrepreneurs should have their ear tuned to the results they see on the ground, and be constantly flexible and adjusting to their circumstances.  It is simply not enough to have a well thought out business model.  That business model will definitely not be accounting for important variables, or may otherwise be over-emphasizing unimportant variables.