CONSUMER SENTIMENT ANALYSIS TO THE SERVICE PERFORMANCE OF THE INDONESIAN STATE ELECTRICITY COMPANY DURING THE COVID-19 PANDEMIC
This study aims to identify problems in the service performance of the State Electricity Company/PLN in Indonesia during the Covid-19 pandemic and to formulate priority solutions for service improvement. The methods used in this research are Naïve Bayes Classifier, Text Associations, and Analytic Network Process (ANP). The Naïve Bayes Classifier and Text Associations methods identify the most dominant problems based on social media's consumer perceptions. The recognized text associations refer to the marketing mix elements, including product, price, promotion, place, physical evidence, people, and process. The data sources in this study are consumer comments submitted on Twitter, Facebook, and PLN Mobile for the period March-June 2020. Furthermore, the ANP method is used to determine service improvement priorities, where the decision-makers in the weighting come from the company management. Based on Text Associations results, several dominant problems were identified: rising bills, slow response to consumer complaints, frequent power outages, unstable power supply voltages, and unreliable PLN-mobile applications. When referring to the marketing mix elements, these problems fall into product, price, and process. Furthermore, based on the ANP method's application, the main priority for improvement is to change the recording system for consumer kWh meters that are still using the conventional system (postpaid) to a token system (prepaid). This recommendation is considered appropriate to solve problems during the Covid-19 pandemic. There is no need to assign PLN employees to manually record bills through direct visits to consumer locations through its implementation. The next priority for improvement is the readjustment of distribution transformers and the upgrading of electricity grid technology. The combination of Naïve Bayes Classifier and Text Associations to Analytic Network Process (ANP) can broadly capture today's consumers' needs and wants to be used as a frame of reference in developing decision-making models to formulate priority solutions for service improvement.
Analytic Network Process, Marketing Mix, Naïve Bayes Classifier, Sentiment Analysis, Text Associations.