Indonesian Marketplace Trust Analysis Using Text Mining: a Case of Tokopedia

Abstract

The existence of technological sophistication and the use of the internet can be done with various things such as using social media, e-commerce, and others. Tokopedia is a marketplace where e-commerce activities can be carried out. In its development, Tokopedia experienced several adverse events such as data leaks, a decrease in monthly website visits in 2020, and the performance of the Tokopedia brand, which was still below other marketplaces. Based on this incident, Tokopedia experienced a decline which could make the view of Tokopedia less favorable, such as trust. This study aimed to see how Twitter users trust Tokopedia based on trust factors in e-commerce: perceived privacy, perceived security, perception of website quality, perceived risk, and internet experience. Text mining methods were used in this study, i.e., multiclass classification and sentiment analysis. The study results show that trust in Tokopedia is still lacking. There are results from the ecommerce factor where perceived privacy and risk have the most negative percentages among other factors. These results can be of particular concern in knowing the recommendations for marketing tactics that Tokopedia should carry out. In addition, this study is fruitful in seeing how the application of big data can be helpful in management regarding trust using available technology, such as the application of text mining to analyze reviews on a topic.