Provider feedback information and customer choice decisions on crowdsourcing marketplaces: Evidence from two discrete choice experiments

•Develops a model of the direct and interaction effects of provider profile information on customer decisions on crowdsourcing marketplaces.•Shows that highly visible and costly-to-generate profile information (e.g., high number of reviews and high average weighted rating) have the largest impacts.•Reports the results from two separate, identical experiments, one in laboratory and one in natural settings.•Based on a state-of-the-art discrete choice experimental design, with Pareto optimal choice sets and random utility models.

Crowdsourcing marketplaces are increasingly becoming popular for the online transactions of services. On these marketplaces, profile information of providers, especially feedback left by previous customers, is the main information source for choice decisions of prospective customers. In the study reported in this paper, we examined the impacts of various feedback information components on provider profiles on the decisions of customers. We conducted two fractional factorial discrete choice experiments, one in a controlled laboratory setting and one online on a crowdsourcing marketplace. We found that the feedback information components “number of reviews” and “average weighted rating” have the largest impacts on the decisions of customers. We also found that “positive ratings” and “positive comments” have significant impacts on customers' decision-making, especially when they appear on the first feedback page. We also found in the lack of highly visible feedback components on the subsequent feedback pages, “negative comments” become a significant determinant of customers' decisions. We also showed the significant impact of information consistency on customers' decision-making, through the synergistic interaction effects between different feedback components. Finally, we found evidence that the cost of evaluating a feedback information component has a negative impact on the likelihood of customers evaluating that information component. The article concludes with implications of the findings of the study for theory and practice.