On e-commerce platforms, to cope with the problem of information overload, the recommendation system is widely used by online sellers. Product characteristics are often used as basic features for algorithms to generate recommendations. However, how the interaction between product characteristics and recommendation sources, both product signal cues transmitting information of product quality, affect consumers decisions remains to be explored. By conducting an online experiment, we tried to compare the recommendation effects between different combinations of product characteristics and recommendation sources empirically. Our results make theoretical contributions to the research on signaling theory in e-commerce context and the research of content-based recommendation. Moreover, both e-commerce designers and sellers could benefit from the practical implications of our study.