Differences in Online Review Content between Old and New Products: An Abstract

Abstract

Online product reviews are one of the most important sources of information that influences the consumer decision making process. Studies have shown a positive relationship between online reviews and product sales. Not surprisingly, extant research has studied the factors that influence reviews such as include reviewer credibility and review content, review volume, valence, and dispersion, the impact of negative versus positive reviews, and the influence of product category.

Despite the important influence of online reviews on consumer decision making, there are several key empirical and theoretical questions in the domain of online reviews that remain unanswered. Extensive research on online reviews has been taken place for movies, books, and television shows. There has been relatively limited research into reviews for new products in other, less well-known product categories. In addition, most research has focused on the volume, valence, and variance of reviews. Research investigating review content is emerging, although much of this research is based on standard attributes of the text using packages like LIWC (Linguistic Inquiry and Word Count). Research investigating more specific aspects of review content is less prevalent. Finally, there have been several studies focused on the impact of online reviews on new products sales and review-based mining to get the pricing power of different attributes. However, these inquiries have not taken into consideration how reviews for new products differ from reviews for similar products that already exist and are available online.

The aim of this paper is to compare content of online reviews for new and established products. We seek to understand how consumers evaluate old and new products in the same category based on online reviews and whether there are systematic differences in the type of content discussed for each type of product. We use Amazon review data from two categories to compare review content for a new product and the leading established product in each category. The present study serves as a pilot to examine the efficacy of machine-based coding of product reviews along a number of relevant dimensions and probe for potential differences to inform future research.

We find that there is a difference in the way consumers review old and new products in a category. The reviews of new products are highly descriptive and detailed with experiences shared by the users. Additionally, when we look at the attribute-wise impact on the ratings, it is found that attributes that are common to both seem to have a higher impact on the ratings than attributes which stand out in the new products.