This paper investigates actual Venture Capital (VC) decision making as it occurs over time in its natural decision environment. Our qualitative analysis is based on a comprehensive, longitudinal data set comprising 11 years of archival data from a European-based VC firm. During this time, the VC managed two funds, reviewed a total of 3631 deals, and made 35 investments. By adopting this research methodology, we can overcome several limitations of post-hoc methods and experiments commonly used in this research stream, and also have the opportunity to tackle some fundamental, yet hitherto elusive research topics. For example, our findings reveal how the importance of decision-making criteria varies between different stages of the evaluation process. They also show how VC portfolio composition and VC management time serve as important, yet to-date largely neglected decision-making criteria. Implications for research and practice are outlined.