How diversity contributes to innovation in crowdsourcing: An evolutionary diversity model

Organizations are increasingly engaging crowds to generate innovative solutions to complex organizational problems using collaborative-based crowdsourcing events such as innovation contests. The assumption underlying collaborative-based crowdsourcing is that innovation is more likely when the crowd is diverse because, as the online interactions surface diverse perspectives, new ideas are sparked, knowledge is recombined in new ways, and innovative solutions emerge. However, management diversity literature states that there are three forms of diversity: one (variety) that helps innovation and two (separation and dominance) that hurt innovation. Moreover, because the management diversity literature focuses on intact groups when studying the effect of diversity, it ignores the possibility that, in a crowdsourcing event, diversity may evolve or accumulate over time as individuals engage in different discussion neighborhoods with different degrees of diversity in each. This paper uses data from a crowdsourcing event to explore the relationship of diversity to the innovativeness of ideas generated during the event. The three traditional forms of diversity (variety, separation, and dominance) measures were applied to both the neighborhoods in which innovative ideas emerged as well as to neighborhoods that idea initiators engaged in prior to posting their innovative ideas. Results of the analysis reveal that, in contrast to the management diversity literature, separation actually fosters innovation when it occurs in earlier neighborhoods. This suggests a notion of evolving diversity in crowdsourcing in which the neighborhoods people participate in affect them in ways not currently suggested by existing management theories. The implications for organizing crowds for innovation are drawn.