Improving new venture success: the role of market experimentation in supporting entrepreneurial learning

Entrepreneurial learning researchers have argued that experimentation is a useful learning technique for entrepreneurs to better understand customer needs (Nicholls-Nixon et al., 2000; Politis, 2005). However, a detailed analysis of extant entrepreneurial learning literature suggests that the experimentation concept is underdeveloped in explaining various aspects of experimentation. In this respect, little is known about: what an experiment is; how experiments are undertaken; how they are evaluated for success or failure; how they are developed and evaluated within the entrepreneur’s social network; and how entrepreneurs learn from these experiments. To advance understanding, a theoretical framework of the entrepreneurial experimentation learning process is constructed and using this, the learning processes of three novice digital technology entrepreneurs, each developing a new venture opportunity, are investigated and analysed. The research takes an interpretive, qualitative case study approach comprising multiple in-depth longitudinal interviews with the entrepreneurs and key members of their social network. The results offer five key contributions to theory. First, these entrepreneurs started the experimentation process with attempts to exploit the opportunity through their experiments, rather than seeing their first experiments as part of the exploration process. Second, experimentation involves multiple different types of explorative and exploitative experiments and levels of investments. Third, rather than learning from small failures, these entrepreneurs, with their predisposition towards success, tend to over-optimistically evaluate small successes in their experiment outcomes as evidence to support continued exploitation. In doing so, these entrepreneurs repeatedly ignore the build-up of evidence pointing towards failure. Fourth, in all cases, the entrepreneurs exhibited low-level learning from exploitation activities, and in most cases, this only changed with a critical incident and full recognition of a small failure. Fifth, and finally, once involved, social actors’ advice supported high-level learning, with critical reflection, new market ideas, and guided pre-experiment design and measurement. This was a key area missing when entrepreneurs experimented in isolation. That said, the evidence suggests that social learning was most effective when the entrepreneurs received regular, consistent support and advice from a single actor before, during, and after experiments. Contributions to practice include: first, entrepreneurs should be encouraged to objectively verify and validate experiment results; second, individuals and organisations supporting entrepreneurs need to avoid the descriptive term small failures and refer to setbacks; and third, entrepreneurs should be made aware of support available prior to these setbacks and encouraged seek out social actor engagement before encountering setbacks. This study is subject to inherent limitations of the qualitative approach taken, which restrict the generalisability of these findings. The concluding section, therefore, outlines these limitations and proposes ways in which future research can address them.