The Effects of Recommendation Sources and Content on Source Credibility and Product Informativeness

Although online recommendation agents (RAs) have become increasingly critical to the success of online commerce, RAs require users to have their preferences and needs elicited. This user interrogation has recently been criticized, as users may not have the motivation or ability to indicate a full set of product attributes. In this study, we compare the RA and its associated interrogation feature against three alternative online recommendation sources (i.e. store, consumers, and experts) that do not require the initial interrogation process necessary for RAs. We draw upon the Elaboration-Likelihood Model (ELM) to compare the differences among the sources as well as the content (i.e. high vs. low quality products). Using an online experimental study with 124 subjects, we found that 1) the RA was not perceived to be more credible and informative than experts, thus an expert’s recommendation is suggested as an alternative to an RA; 2) recommendations from RAs and experts, but not consumers, were found to be superior to the store’s recommendation in terms of perceived source credibility and perceived product informativeness; 3) regardless of the source, recommendations with high-quality content were perceived to be significantly more credible and informative than those with low-quality content. The results of our study not only increase our understanding that persuasion can take place through a central route and a peripheral route simultaneously, they also inform practitioners as to what recommendations can be incorporated in their e-commerce websites to form a credible and informative impression.