Perceived usefulness of online customer reviews: A review mining approach using machine learning & exploratory data analysis

Online customer reviews, considered as electronic word of mouth, have become very useful in the era of e-commerce as they facilitate future purchase decisions. The present study discusses the central and peripheral sources of influence, such as the content of the review, star rating, review length, and the total number of votes on the perceived usefulness of the review. It analyses reviews from Amazon.com on three products, namely, a videogame, digital music, and a grocery item. Using text mining, the study uncovers sentiment polarity, identifies sentiment patterns, and finally, analyses the perceived usefulness of reviews under the moderation effect. The study establishes that the impact of the central route is not significant for search goods. The study concludes that peripheral sources have a significant impact on the search products. Our study provides insights on how mar-keting strategies can be formulated by online retailers based on the product type.