This paper develops and tests a model of the effectiveness of selection processes in eliminating less fit organizations from a population when organizations are undergoing adaptive change. Stable organizational traits, such as a search strategy or routine, do not imply that an organization's performance will remain stable over time or that cross-sectional differences in performance will persist. These properties create the possibility that population-level selection processes will be inefficient in that organizations with potentially superior long-run performance will be selected out. We theorize that organizational-level adaptation often results in fluctuations in current performance across time. These fluctuations may attenuate the degree to which current performance differences among organizations are indicative of future performance. As a consequence, search strategies that generate systematically different performance trajectories, even if they share a common long-run outcome, will generate differing survival rates. These ideas are explored using a formal simulation model employing the framework of NK performance landscapes. Our central finding is that selection may be systematically prone to errors and that these selection errors are endogenous to, and differ markedly across, firms' search strategies.