This article suggests a model to predict product sales with the herd behavior effect; the tendency to gravitate toward, and buy products that others are buying, ignoring comparable unordered products available at the same time. Some products attract many buyers and become coveted, the center of buying attention, while other products are overlooked, receiving few orders. The overall model has two components: (1) a logit to predict the product’s probability of choice set inclusion (2) a poisson regression to predict the sales quantity of each cases (inclusion/no inclusion), which is, in turn, integrated with the logit. This article illustrates its modeling approach with online group buying data for CD. The results show that sales quantity is significantly related to the herd behavior via products’ cumulative sales and price range, and the overall effect is large when compared to product- or seller-specific factors that predict product sales. The practical implications of the findings, and promising research opportunities in this area are also discussed. This provides a decision aid for seller seeking to understand and forecast products’ demands.