ANALYSİNG DİGİTAL GAME FEEDBACK WİTH SENTİMENT ANALYSİS: IS IT POSSİBLE TO TRANSİTİON FROM E-WOM TO R-WOM?

Abstract

The vast number of comments on online platforms makes it nearly impossible to evaluate this feedback holistically using only the human mind. In this context, there is a need for new tools that offer advantages in terms of cost, time, labour, and informational value for market and marketing research conducted on online platforms. The main purpose of this study is to test whether market segmentation can be achieved based on game genres by evaluating e-WoM data using sentiment analysis methods. For this research, user reviews of the top 100 highest-rated video games for the PlayStation 5 console on the Metacritic website were analysed. The automated extraction and processing of user reviews were conducted using web scraping techniques and Python programming. A total of 44,065 comments across all games were evaluated using sentiment analysis. Within this framework, difference tests were conducted to compare polarity scores across different game genres. Sentiment analysis of user comments showed consistent results according to game genres. Conceptually, however, a clear distinction between action and adventure games was not feasible. As a result, no significant difference was found between these two genres based on the study’s test results. On the other hand, there were significant differences in polarity scores between racing, role-playing, shooting, unique, and adventure/action game types. The findings of this study provide an alternative perspective on how e-WoM and r-WoM applications can contribute to marketing strategies within the gaming industry.