Unfolding the characteristics of incentivized online reviews

The rapid growth of social media in the last decades led e-commerce into a new era of value co-creation between the seller and the consumer. Since there is no contact with the product, people have to rely on the description of the seller, knowing that sometimes it may be biased and not entirely true. Therefore, review systems emerged to provide more trustworthy sources of information, since customer opinions may be less biased. However, the need to control the consumers' opinion increased once sellers realized the importance of reviews and their direct impact on sales. One of the methods often used was to offer customers a specific product in exchange for an honest review. Yet, these incentivized reviews bias results and skew the overall rating of the products. The current study uses a data mining approach to predict whether or not a new review published was incentivized based on several review features such as the overall rating, the helpfulness rate, and the review length, among others. Additionally, the model was enriched with sentiment score features of the reviews computed through the VADER algorithm. The results provide an in-depth understanding of the phenomenon by identifying the most relevant features which enable to differentiate an incentivized from a non-incentivized review, thus providing users and companies with a simple set of rules to identify reviews that are biased without any disclaimer. Such rules include the length of a review, its helpfulness rate, and the overall sentiment polarity score.