Research concerning diversification and specialization of venture capital funds typically does not consider how a VC's effort might influence performance of different portfolios. We develop a model that analyzes VC effort when there is the potential for cross-sectional and/or serial knowledge spillover among projects. The model generates two implications concerning VC effort and performance. First, VC post-investment effort is a nonmonotonic function of performance shocks, especially for diversified VCs. Second, greater cross-sectional and serial knowledge spillovers improve the performance of specialization relative to diversification, and shape how the number of decision stages in a project affects portfolio choice. Copyright © 2011 Strategic Management Society.