Helpfulness of online reviews: Examining review informativeness and classification thresholds by search products and experience products

Information overload often makes it difficult for consumers to identify helpful online product reviews through the traditional “helpful votes” function; therefore, it has become particularly important to efficiently identify helpful reviews. By differentiating search products from experience products, this research examines the impact of different measurements of review informativeness on review helpfulness, and proposes different classification thresholds to individually identify the helpfulness of online reviews for search products and for experience products, respectively. Further, our study applies machine learning algorithms to predict the performance of the classification based on our proposed review informativeness measurements and classification thresholds. All experiments were conducted using a dataset from JD.com, one of the largest online electronic marketplaces in China. Our results offer guidelines to design different helpfulness classification standards for search products and for experience products.