Implementing text mining-based IPCA on Islamic versus conventional mobile banking in Indonesia

Purpose This study aims to analyse the importance–performance competitor analysis (IPCA) of mobile banking service quality dimensions in Islamic and conventional banks, based on online reviews, using a text-mining approach. Design/methodology/approach To conduct the IPCA analysis, the authors used (1) latent Dirichlet allocation (LDA) to extract service dimensions from online reviews, (2) term frequency-inverse document frequency to calculate importance values and (3) lexicon-based sentiment analysis to calculate performance. Findings The LDA results revealed six service dimensions: stability, login, internet banking, UI & UX, enjoyment and security. Meanwhile, the IPCA matrix results show that Islamic banks exhibit significant differences in the dimensions of login, internet banking, UI & UX and security. Islamic banks must improve application stability and customer experience. Meanwhile, conventional banks can provide more enjoyment than Islamic banks. Research limitations/implications The research results provide insights for bank managers, especially Islamic banks, on the service dimensions that need improvement in the future. The study is useful for enhancing the performance of both banks in serving their customers. Originality/value The novelty lies in combining text mining with IPCA to analyse online reviews of mobile banking services. This approach provides new service dimensions and comparative insights between Islamic and conventional banks.