Assessment of service quality using text mining – evidence from private sector banks in India

As the retail banking institutions are becoming more customer centric, their focus on service quality is increasing. Established service quality frameworks such as SERVQUAL and SERVPERF have been applied in the banking sector. While these models are widely accepted, they are expensive because of the need for replication across bank branches. The purpose of this paper is to propose a novel, user friendly and cost effective approach by amalgamating the traditional concept of service quality in banks (marketing base) and sentiment analysis literature (information systems base).

In this study, the main objective is to analyze user reviews to better understand the correlation between RATER dimension sentiment scores as independent variables and user overall rating (customer satisfaction) grouping in “good” and “bad” as dependent variable through development of authors’ own logistic regression model using lexicon-based sentiment analysis. The model has been developed for three largest private banks in India pertaining to three banking product categories of loans, savings and current accounts and credit cards.

The results show that the responsiveness and tangibles dimensions significantly impact the user evaluation rating. Even though the three largest private banks in India are concentrating on the tangibles dimension, not all of them are sufficiently focused on the responsiveness dimension. Additionally, customers looking for loan products are more susceptible to negative perceptions on service quality.

This study has highlighted two types of scores whereby user provided overall evaluation scores help provide validation to the sentiment scores. The developed model can be used to assess performance of a bank in comparison to its peers and to generate in depth insights on point of parity (POP) and point of difference (POD) fronts.