Entrepreneurial activity is an important source of innovation in information technology products and services. Prior literature suggests that IT innovators should be agile, adaptive, willing to change direction frequently, and acquiring the necessary resources to facilitate the change. Social networks have been suggested as essential for acquiring information and resources and therefore in facilitating the venture development process. Yet, we know little about whether these adaptive processes could ultimately affect outcomes of IT ventures at the early stages and whether externally introduced changes in social networks can modulate the effect. With the ability to control presented content in a massively open online course (MOOC), we conduct a randomized field experiment that encourages students to either pursue an adaptive or a planning-based approach in building their startups and to seek mentors with different types of social networks. Measuring their performance at the end of the class and in a longer-term follow-up (two years later), we find that instructing entrepreneurs to have a strong, persistent vision for their startups often results in better performance at early stages. However, the disadvantages of adaptive strategies at early stages are mitigated by engaging a mentor with diverse social ties, and this strategy generates the best long-run outcomes (measured two years later). The results suggest that mentor selection -- and especially the social networks of the mentors -- are critical for achieving the benefits of agile product development.
Class learning outcomes, longer-term outcomes related to ventures created/ worked on in the course
Planning-based approaches are more effective than adaptive approaches in generating short term success, but this gap decreases in our long-term measures. However, combining the adaptive approach with a mentor with a diverse network of contacts generates the highest short-term and long-term outcome.