Purpose -Although crowdsourcing has gained significant attention and is being used by numerous companies to develop new products and solve practical issues, the performance of crowdsourcing is not optimistic. The purpose of this paper is to develop a validated risk profile of crowdsourcing and investigate the relationships among different types of risks and those between risks and performance in crowdsourcing. Design/methodology/approach -Based on the quantitative data collected from 136 crowdsourcing participants in China, two dimensions (i.e. social system and technical system risks) and five sub-dimensions (i.e. crowdsourcer, relationship, crowdsourcee, complexity, and requirement) of crowdsourcing risks are developed and validated. A theoretical model that integrates crowdsourcing risks and performance is developed. The technique of partial least squares is employed to assess the measurement model and test the hypotheses. Findings -The empirical evidence determines the positive association of social system risks with technical system risks, which in turn negatively affect crowdsourcing performance. Specifically, relationship risk is positively affected by crowdsourcer and crowdsourcee risks, and these risks positively affect requirement and complexity risks. However, requirement and complexity risks negatively affect crowdsourcing performance. Originality/value -This study explores the interrelationship between various risks and the relationship between risk and performance in the context of crowdsourcing by integrating risk-based view with socio-technical theory. Systematic but different risk mitigation strategies should be designed in crowdsourcing to manage risks and enhance performance.