The model shows that managers and entrepreneurs make better decisions under uncertainty if they adopt a scientific approach in which they formulate and test theories. The model predicts that
they are more likely to terminate projects with negative returns, commit to projects with positive returns, or pivot to projects with higher returns. These implications are tested by combining the results of four Randomized Control Trials (RCTs) involving 754 start-ups and small-medium enterprises and 10,730 data points over time. The empirical analysis corroborates the predictions of the
Innovation decisions: project termination, radical change to the project (pivot), and performance.
Firms in the sample pivot at most six times during the observation window. Fifty nine percent of the sample never pivots, and six percent of the firms pivot more than two times. Thirty four percent of the firms terminate their projects within the observation window. The average amount of revenue is EUR 15,538, with large variation in the sample since a substantial number of firms has zero revenue within the observation window. The intervention raises the probability of pivoting once or twice and lowers the probability of not pivoting or pivoting more than twice. Results on the full sample show that on average treated firms earn EUR 6,500.691 more than control firms.