Text mining approach to explore dimensions of airline customer satisfaction using online customer reviews

Abstract

The airline industry operates in a highly competitive market, in which achieving and maintaining a high level of passenger satisfaction is seen as a key competitive advantage. This study presents a novel framework for measuring customer satisfaction in the airline industry. Using text mining methods we explore Online Customer Reviews (OCRs) to provide guidelines for airlines companies to improve in competitiveness. We analyze a database of more than 55,000 OCRs, covering over 400 airlines and passengers from 170 countries. Using a Latent Dirichlet Allocation model we identified 27 dimensions of satisfaction described by 882 adjectives. Dimensions and adjectives were used to predict airline recommendation by customers, resulting in an accuracy of 79.95%. The most relevant dimensions for airlines' recommendation prediction were calculated. OCRs were stratified according to several variables. Of those, type of passenger impacted the least on the number of dimensions of customer satisfaction, while type of cabin flown impacted the most. Observing results in different publication years we showed airline customer trends through time. Our method showed sensitiveness to identify variations in dimensions distribution according to different passenger characteristics and preferences. Practical implications are that airline service providers aiming at maximizing customer satisfaction should focus their efforts on (i) customer service to first class passengers, (ii) comfort to premium economy passengers, and (iii) checking luggage and waiting time to economy class travelers. Regression analysis revealed cabin staff, onboard service and value for money as top three dimensions of satisfaction to predict the recommendation of airlines. Designing services that excel in those dimensions is likely to improve the company's performance with customers.