THE DEVELOPMENT OF CLOTHING MSME’S LIVE COMMERCE CONTENT MODEL BY CONSIDERING VIEWER EMOTIONS AND THEIR IMPACT ON PURCHASE INTENTION

• Main Author: M.J. Desanto, Tirza
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/77868
• Institution: Institut Teknologi Bandung

Live streaming commerce is one of the features most widely used by clothing MSMEs and commerce platform users during the Covid-19 pandemic. The use of live streaming really helps MSMEs because they can still run their business in a pandemic situation. However, the problem that MSME sellers always experience when doing live streaming is that there are still many viewers who do not continue purchasing after they watch the live streaming. This phenomenon indicates that there is something wrong or lacking in the live streaming strategy carried out by MSMEs so that the intention to buy live streaming viewers is still lacking. One method that can be used to evaluate live streaming content is Facial Expression Recognition (FER). However, no previous research has examined FER in live streaming. To fill this gap, this research will try to find out what kind of live streaming content is liked by viewers so that it can increase purchasing intentions in live streaming commerce for clothing MSMEs in Indonesia. The model development of this research consists of research related to live streaming content, emotions, and purchase intentions. The live streaming content used in this research is content that is usually provided by sellers, namely visualization, discounts, professionalism, interactivity and entertainment. The emotions measured in this study were positive emotions, negative emotions and neutral emotions which were measured from the recognition of facial expressions shown by respondents. Purchase intention in this study was measured using a questionnaire. This research uses a quantitative research approach using purposive sampling. The preliminary study in this research was carried out by interviewing 4 respondents to find out the forms of live streaming content for each variable. Furthermore, the quantitative method used in this research is divided into two, namely an experimental design for 32 respondents to determine respondents' emotions when watching live streaming content which has been designed based on interview results and followed by logistic regression analysis. The experimental design of this research uses facial expression recognition analysis with the help of FaceReader 9.0 software developed by Noldus. The data processing analysis in this research consists of several analyses. First, descriptive analysis to find out the profile of the respondents in this study as well as descriptive analysis related to the emotions that dominate in live streaming. Second, non- parametric difference tests consisting of the Kruskall Wallis test and the Friedman test. Kruskall Wallis test to determine the differences in each subcategory regarding the emotions expressed. Friedman test to determine the effect of providing a stimulus video on purchase intentions. The final test is logistic regression analysis to determine the influence of emotions on purchase intentions. The finding of this research revealded that neutral emotion is the only emotion that dominates in every video in this study. This indicates that providing a combination of video stimulus with different live streaming content was not able to influence the emotions expressed by the audience. Furthermore, neutral emotions that dominate each video were unable to influence the purchase intentions of live streaming viewers. Based on the results of data processing on differences in the combination of content for each video on purchase intentions, it was found that there were significant differences in purchase intentions for each video. This means that providing different combinations of live streaming content for each video influences purchase intentions. Furthermore, this research found that the combination of live streaming video content contained in video 3 was the combination of content with the highest increase in purchase intention. The content contained in video 3 is a visualization of trying on clothes but does not provide clothing zoom, with a minimum discount of 25%.