Managing Complexity and Unforeseeable Uncertainty in Startup Companies: An Empirical Study

Novel startup companies often face not only risk, but also unforeseeable uncertainty (the inability to recognize and articulate all relevant variables affecting performance). The literature recognizes that established risk planning methods are very powerful when the nature of risks is well understood, but that they are insufficient for managing unforeseeable uncertainty. For this case, two fundamental approaches have been identified: trial–and–error learning, or actively searching for information and repeatedly changing the goals and course of action as new information emerges, and selectionism, or pursuing several approaches in parallel to see ex post what works best. Based on a sample of 58 startups in Shanghai, we test predictions from prior literature on the circumstances under which selectionism or trial–and–error learning leads to higher performance. We find that the best approach depends on a combination of uncertainty and complexity of the startup: risk planning is sufficient when both are low; trial–and–error learning promises the highest potential when unforeseeable uncertainty is high, and selectionism is preferred when both unforeseeable uncertainty and complexity are high, provided that the choice of the best trial can be delayed until its true market performance can be assessed.