In recent years, the scale of the electronic commerce (EC) market has continued to expand along with the spread of the Internet, both for business and personal transactions. Many consumers use EC sites, and it can be said to have penetrated their daily lives. Many consumers make purchasing decisions based on reviews written by other users when purchasing on EC sites. Therefore, reviews are considered to contain a lot of important information indicating the characteristics of products from the viewpoint of users who have purchased and used the products. Until now, many experiments have been conducted on the effects of reviews on consumer behavior and methods for extracting important information from reviews. However, there is room for debate regarding the utilization and validation of the usefulness of information representing product characteristics obtained from reviews. In this study, we propose a method to create product descriptions using information extracted from product review data and to verify the usefulness of such descriptions. Specifically, we identify important words that are considered to represent the characteristics of products by analyzing the sentiment of the reviews and selecting feature words based on the TF-IDF method. We also verify the effectiveness of the feature words by using a screen-based eye tracking device to create product descriptions. The results showed that there was no statistically significant difference in the median gaze duration between the product descriptions created using the feature words obtained from the reviews and the conventional product descriptions.