Assessing the Commercial Viability of Seed- and Early-Stage Ventures

Abstract

The question of how venture capitalists assess the unique characteristics of potential ventures lies at the basis of all venture investing. Yet the data on human versus purely statistical prediction models is troubling, often indicating that humans are poor decision makers in the face of uncertainty. This article looks at over 500 potential new ventures, each reviewed by skilled engineers with some business background. From a list of 37 criteria, four were found to be statistically significant: expected profitability; development risk; functional performance; and IP protection. The rate of accuracy of the skilled predictors was surprisingly high.