Designing a system often begins with matching existing solutions to current problems. Specifically, integration mechanisms are mapped onto situations. Novices are not good at this task, and experts are rare. Could crowdsourcing, that is, aggregating the suggestions of individuals working independently, be effective? Two experiments, one with design students in a classroom, and another with participants on the web, demonstrated that the crowd possesses wisdom about how to match mechanisms to situations. Participants also categorized situations, and those who name their categories were better at matching than those who didn’t. The results have pragmatic implications, suggesting it is possible to crowdsource design, and providing new ways of eliciting, testing, and training expertise. More generally, the paper suggests a new model for information system design based on analogical mapping.