Crowdsourcing in the form of human-based electronic services provides a powerful way of outsourcing so called human intelligence tasks (HITs) to a large workforce of people over the Internet. Because of the limited control over that workforce, it is challenging to ensure the quality of the work results. Several approaches have been proposed that can be applied to specific types of HITs. However, it is difficult to identify a suitable quality management approach for any given type of HIT. This paper aims to provide a first sketch of a decision matrix.