Article Preview
Top1. Introduction
Online user reviews have gradually become an important display content of many online shopping platforms with the full application of Web 2.0 and the rise of social networking sites and new media. It refers to consumer's judgment (evaluation, response, recommendation, or complaint) on a product or service after purchase on the Internet. These reviews contain a large amount of useful information, and because the data comes from different individuals, it can avoid the one-sidedness of information. An online review is one of the most important influencing factors for customers' decision-making when buying unfamiliar products. Meanwhile, from the perception of business firms, it contains various user feedback on products or services, which can provide manufacturers with Research & Development and improvement strategies. User reviews have increasingly become the focus of theoretical and social practice. Recently, the amount of consumer reviews has been increasing rapidly, creating significant big data challenges for consumers and businesses (Singh et al., 2017).
It is essential for online businesses to continually improve and even innovate their products from user experience analysis. Online review, an important component of user-generated content, is an excellent source for research and commercial use and it shows the feasibility of getting quality improvement tips from it (Yang et al., 2019). For example, through gamers’ comments, game developers could discover the advantages and disadvantages of games and bugs they did not notice before. After that, they can release a new version to fix and update the game according to information obtained from reviews. It also can give them directions in developing new games because of the understanding of users’ needs. Nowadays, the development of emotional analysis technology and text mining technology has greatly improved the ability to perceive customer needs and provide timely feedback.
For consumers, online user reviews give them an opportunity to know more about products before making decisions. These reviews were generated by customers who have purchased and used the products. This is a good way to reduce information asymmetry in market transactions. However, due to too many reviews, the amount of information is much higher than what the customers can consume, withstand, or need. It is impossible for a potential buyer to read through all the reviews of the product. In most cases, customers refer to the first few comments displayed on the product review page, so only comments that consumers have read are likely to help them make decisions. If helpful reviews are more visible to potential buyers, it will be easier for buyers to get the useful information they need (Singh et al., 2017). Therefore, it encourages online retailers to evaluate the helpfulness of reviews. Many websites now set a question next to consumer reviews, like “Is this comment helpful to you?” with “thumbs up” and “thumbs down” buttons.
The helpfulness of reviews is usually just assessed by the number of helpfulness voting. Most websites rank reviews in the order according to the usefulness of reviews voted by consumers and put the review that gets the most votes in the first. It helps consumers to reduce the cost of information search and increase the efficiency of purchasing decisions. By improving the user-friendliness of the website, more users will be attracted. However, there exist some problems in identifying the value of reviews in this way. Firstly, the latest user reviews may be placed after the previously less useful ones since too many comments are posted at the same time, so that cannot be voted in time. Secondly, for some products that are not so popular, almost no one proposes for the usefulness, so these reviews cannot be ranked by usefulness. These two reasons will both cause the failure of the voting mechanism. Consequently, it is necessary to research factors affecting the helpfulness of reviews and build prediction models for online businesses to forecast helpfulness of any new reviews and sort them effectively.
Many previous studies did research on online user reviews of various product categories, such as movies, phones, restaurant services, hotels, clothes and others. As determinant factors of review helpfulness might differ for different products, this article will focus on the game reviews on a digital game social platform, Steam platform, with little research in this area.