Abstract: Recently, some e-commerce platforms launched the function of follow-up product review, so consumers can give a follow-up review within a certain period of time after their initial review on the purchased product. Follow-up reviews have become an important factor affecting product sales. Based on the reviews and sales data of 51 kinds (102 pieces) of products, this study examines the impacts of valence consistencies between initial and follow-up reviews given by the same consumers on sales, where the valence of the reviews is evaluated by applying the text mining technology on review contents. Results show that the valence consistencies between homologous initial and follow-up reviews have different effects on sales, and the effects are moderated by product brand. Specifically, for the valance-positively-changed reviews sets, they have a positive effect on sales of low-end products but a negative effect on sales of high-end products; for the valance-negatively-changed reviews sets, they have a negative effect on product sales regardless of brand. It is concluded that the sets of initial and follow-up reviews exhibit recency effect on low-end products and negative deviation effect on high visibility products. For low-end products, managers should pay special attention to after-sale service to improve the follow-up reviews; as for high-end products, after-sale service is only a basis and managers should make efforts to strengthen product quality.

Key words: electronic-word-of-mouth(eWOM), product sales, follow-up review, valence consistency of initial and follow-up reviews