Ideas from employees are a major source of value creation in firms, yet the merits of rewards for incentivizing the generation of ideas are highly contested. Using a computational model, we show that firms can improve performance by offering low-powered rewards for the selection and implementation of employee ideas. Low-powered incentives provide a sufficient stream of good ideas, but few exceptional ones. Higher-powered incentives, in contrast, do not systematically translate into exceptional ideas either, but generate an excessive number of good ideas. Performance-based rewards thus appear to be a blunt tool to harness the long tail of innovation. We develop propositions to guide empirical research and discuss their implications for strategy and organizational design. Copyright © 2013 John Wiley & Sons, Ltd.