Exploring mobile banking service quality dimensions in Pakistan: a text mining approach

Customer reviews of mobile banking (m-banking) apps contain the most direct and first-hand accounts of customer experiences with m-banking. However, surprisingly little effort has been made to understand m-banking service quality using these reviews. Therefore, this study aims to discover m-banking service quality dimensions from customers' reviews of the m-banking apps through a text mining approach.

Reviews of m-banking apps of 24 banks operating in Pakistan were scraped from Google Play Store. Latent Dirichlet allocation (LDA) method was applied to discover the dimensions of m-banking service quality from 24,529 positive and 29,569 negative useable reviews.

Different dimensions of m-banking service quality are discussed in positive and negative reviews. Positive reviews focus on security, convenience, ease of use, continuous improvement, usefulness and app attributes, whereas negative reviews discuss system availability, responsiveness, faulty updates, login problems and reliability.

The results are based only on customer reviews in one country and generalization may not be possible. Moreover, due to the unavailability of demographic information about reviewers, the effect of demographic characteristics on users' perceptions of m-banking quality could not be determined.

The study provides managers with useful insights to improve the service experience of m-banking customers. The study also demonstrates how managers can employ text analytical techniques to assess and improve the quality of m-banking services.

In addition to enriching the understanding of m-banking quality based on direct and first-hand user experiences, the current study also provides initial evidence for the two-factor structure of m-banking service quality.