Professor Manuel Sosa, INSEAD
The empirical literature has consistently found that lone inventors significantly underperform teams in creating better quality output. Yet theory and experimental evidence suggests that single inventors need not always underperform. Using utility and design patent data from 1983 to 2009, we compared the effect of working alone (vs. working in teams) on the probability of creating breakthroughs. Over the entire population of patents, we find results consistent with existing literature that working alone lowers the probability of achieving a breakthrough by a significant margin. However, we also find an important boundary condition – the disadvantage of the lone inventors disappears for design innovation. A fundamental difference between design and technological inventions is that the design task is largely integral (i.e., design is evaluated holistically and structurally as a single idea, as opposed to consisting of distinct chunks). We empirically test the effect of the number of chunks of an invention on its probability of becoming a breakthrough, and find that (i) innovations with a lesser number of chunks relatively advantages an inventor working alone over that inventor working on a team; and that (ii) this factor almost entirely explains why lone designers do not suffer a disadvantage with respect to teams in creating breakthrough designs. Finally, we test whether the benefits of collaboration lie beyond the current task at hand. Indeed, our results show an important moderation effect of past collaborations – a lone inventor with a large number of past collaborations significantly improves their propensity to create breakthrough innovations. In carving out areas where lone inventors do relatively better than teams, our results deepen our understanding of the role of collaboration in innovation.