Learning from the unexpected: Towards a behavioral theory of serendipity

We study how entrepreneurs learn from the unexpected and make valuable discoveries through serendipitous “matching pairs.” Through semi-structured narrative interviews with 51 novice and seasoned entrepreneurs, we develop a behavioral theory of serendipity that explains why some individuals can see what others fail to see. Our findings suggest that, to learn from the unexpected, entrepreneurs often behave in ways that contradict canonical predictions from learning theories. For example, instead of continuing alternatives after experiencing successes and changing them after failures, as performance feedback theory would predict, serendipitous entrepreneurs update their beliefs slowly and spend more time exploring the implications of surprises, sometimes pivoting even when the outcomes are coded as successes. Similarly to Normal Accident Theory advising operators to pay attention to weak signals to avoid cascading errors, our propositions illuminate how entrepreneurs can learn from the unexpected in a contrarian way to capture the upside.