Customers’ reviews in e-commerce sites play a significant role in influencing potential customers’ purchasing decisions which ultimately affects products sales. Chinese e-commerce sites like Tmall, Taobao and JD.com contain a collection of aspect tags that group reviews with similar comments tags to help customers browse reviews and evaluate products more conveniently. To validate whether these tags are useful and actually playing a role in promoting future sales, we collected data including product information and review tags on a regular basis for consecutive 8 weeks from Bestore, a snack seller on Tmall. We classified the collected review tags into 9 types based on their semantic meanings. Finally, we analyzed and performed generalized estimating equations (GEE) modeling on the data set consisting of 234 products with a total of 734 tags. The results show that most of the aspect tags are related to immediate period sales volume and certain tags are more capable of nowcasting next immediate sales.