With the proliferation of recommendation functions (RF) on e-Commerce websites, there is growing confusion about how various RF types affect consumers’ beliefs and behavior. Despite the importance of understanding the differential effects of RF types, research focusing on the comparison between provider-generated recommendations (PGRs) and user-generated recommendations (UGRs) has received little attention. This paper reports on two empirical studies that examine the differential effects of PGRs and UGRs on cognitive, affective and relational aspects of consumer beliefs and show how these perceptions influence RF usage intentions. The findings from a field survey (N=366) and a laboratory experiment (N=161) indicate that UGRs (such as consumer reviews) have stronger impact on users’ trusting beliefs and perceived affective quality (i.e. on relational and affective perceptions respectively) than PGRs. Conversely, PGRs (such as collaborative filtering-based RFs) are superior to UGRs in affecting perceived usefulness (i.e. cognitive perceptions). Further, trusting beliefs and perceived affective quality were found to be stronger predictors of usage intentions than perceived usefulness in UGR rather than PGR contexts. By showing which RF types influence different consumer perceptions, the study provides practitioners with clear guidelines on how to design sales efficient e-Commerce websites while enhancing online-consumers’ overall shopping experience.