Indonesia State Electricity Company (PLN) has transformed its business process into a more
modern practice with the intention of boosting efficiency and enhancing service quality. In the
digital economy space, customer expectation is more complex, influenced by the network effect
and public opinion. Facing these facts, it is imperative for PLN to be able to listen to its customers’
voices using publicly available social media channels. Understanding and acting upon large-scale
opinions and conversations is a nontrivial task. To filter large-scale data, we employ Natural
Language Processing (NLP) methodology called Latent Dirichlet Allocation (LDA) to find the
dominant topics in public discourse. It is followed by UTAUT2 model implementation to further
clarify the most critical issues found. Finally, this study modified the e-ServQual and eRecServQual methodologies to propose the best and most personalized solution for each issue. By
using the proposed approach, PLN could accelerate the implementation of data-driven decisionmaking and induce sustainability.