Justin Berg, Stanford University

Betting on the most promising new ideas is key to creativity and innovation in organizations, but predicting the success of novel ideas can be difficult. To select the best ideas, creators and managers must excel at “creative forecasting”—i.e., predicting the outcomes of new ideas. In this research, I build hypotheses on the conditions for accurate creative forecasting, focusing on the impact of creator and manager roles. To test these hypotheses, I use a field study of 339 professionals in the circus arts industry, as well as a lab experiment. In the field study, creators and managers forecasted the success of new circus acts with audiences, and I assessed the accuracy of these forecasts using data from 13,248 audience members. Results suggest that creators were more accurate than managers when forecasting about others’ novel ideas (but not their own ideas). This advantage over managers was undermined when creators had poor ideas that were successful in the marketplace anyway. Results from the lab experiment provide causal evidence for the proposed mechanism driving creators’ advantage over managers, suggesting an intervention that may improve managers’ accuracy.