Choosing Team Members versus Ideas: Evidence from a Field Experiment in Entrepreneurial Team Performance

With the rise of novel institutions for new venture creation, such as accelerators, incubators, company builders, and corporate venturing, a plethora of patterns can be observed how early-stage entrepreneurship is organized. Oftentimes, these institutions make an attempt to professionalize team formation by deliberately assigning team members and topics with the aim of achieving higher complementarity between team members and their innovation topic than would have been achieved with purely endogenous choices by the founders themselves.

The existing entrepreneurship literature tells us that most entrepreneurs choose team members they know well when starting their companies. While this can carry benefits of being familiarized and building on prior shared experiences, it can also favor comfort, complacency and uniformity. Existing research does not provide convincing evidence whether such endogenous team formation leads to more successful entrepreneurial teams or not. On the other hand, there is also no empirical evidence to date on whether and how these novel institutions are effective in fostering early-stage entrepreneurial success by their deliberate means of forming entrepreneurial teams.

We systematically investigate the factors that make team formation effective in early-stage entrepreneurship: exogenous (i.e. pre-assigned) versus endogenous (i.e. self-selected) team formation based on team members versus based on business topics to be explored. More concretely, we are interested in the following research question: How is performance in early-stage entrepreneurship helped or hindered by team formation based on (i) endogenous selection of team members and/ or (ii) endogenous selection of topic, compared to an exogenous pre-assignment of team members and topics? What is the better model: freedom in team members and topic choice or pre-assignment?

We implement a large-scale field experiment in a German university, whose students are required to take an introductory entrepreneurship class. At the end of this class, equal-sized teams of students must submit a business plan in form of startup pitch deck. The first treatment dimension in our two-by-two experimental design pertains to the assignment of team members: students are randomly assigned to either endogenous choice of team members or exogenous pre-assignment of team members. The second treatment dimension applies to the assignment of topics: students will either be allowed to select a topic of their choice or be pre-assigned to a topic. The outcome of interest will be performance in this business plan exercise as well as changes in entrepreneurial self-efficacy and motivation. Randomization will permit us to causally identify if successful team performance rests on choosing ideas versus team members.