Despite the increasing importance of social media marketing, little extant research has studied what factors that consumers would take into account in the decision-making of endorsing a product to their peers with established ties via social media. This paper examines if online reputation (restaurants’ displayed Yelp ratings), which helps update consumers’ expected utility, is a causal factor that affects consumers’ decisions of endorsing via Facebook and purchasing products (the restaurants’ vouchers). I build a stylized Bayesian learning model and derive the hypotheses: (1) a higher online reputation leads to increased social media endorsements and voucher sales, but only when the number of review ratings is sufficiently large; (2) these effects are greater for restaurants with more reviews; and (3) these effects are greater for restaurants with a larger variance in the review ratings. Interestingly, the third hypothesis contrasts to the predictions by established theories (e.g., cue diagnosticity theory). I test the hypotheses using data of Groupon and LivingSocial deals. To identify the causal effects of online reputation, I use a regression discontinuity design by exploiting the institutional feature that displayed Yelp ratings are rounded to the nearest half star. The empirical results largely support the hypotheses. In particular, I find the effect of displayed Yelp ratings on Facebook Likes is greater when the variance of ratings is larger, suggesting that perhaps consumers are risk averse in the decision-making of endorsing a product to their peers. Yet, the effect on voucher sales does not significantly differ with the variance.