With the limited screen size of mobile phones, consumers cannot read product reviews as freely as they do at computer terminals. To rectify the insufficient number of helpful review voters, consideration should be paid not only to the value reviews provided by consumers but also to the relationship between reviews so as to recommend the optimal reading order for consumers and enable them to obtain more product information. In this paper, we measure product uncertainty to show the relationship between reviews with consumers continually reading reviews and to explore the influence of their perception change on their online behaviors. The product uncertainty is computed by the improved Shannon entropy, which measures the product review text on websites. With our data collected from Amazon.com by python coding, three significant findings are detected as follows. First, regardless of the order the reviews are sorted in, whether by the most recent or the most helpful, the results demonstrate that the varied product uncertainty of a review has a significant relationship with its helpful voters. Moreover, the more the product uncertainty is varied by a review, the more possible it will be for the product to obtain voters. Second, when analyzing the influence of product uncertainty on consumers’ purchasing behavior, we find that the lower the product uncertainty computed by many reviews, the higher the product rank, and the more likely consumers are to purchase the product. Third, by exploring the experimental results with knowledge of relevant behavior psychology, we offer different meanings for display reviews on different e-commerce sites.