Customer Loyalty through Improved Customer Service: A Case Study of Biman Bangladesh Airlines Limited

Abstract

This dissertation presents a comprehensive, data-driven investigation into the determinants of customer satisfaction and loyalty at Biman Bangladesh Airlines, addressing a critical gap in the literature on aviation in emerging markets. The research employs a rigorous mixed-methods approach, integrating quantitative analysis of survey data from 425 passengers with qualitative insights from 43 semi-structured interviews with stakeholders. The study examines two primary dependent variables (DVs) as Customer Satisfaction and Customer Loyalty, against eight independent variables (IVs) operationalized through 40 distinct attributes, while controlling for three key demographic variables (CVs): Age, Education Level, and Travel Frequency. A novel Multivariate-Machine Learning Synergistic Model (MML-SM) was developed, synergizing traditional statistical methods with advanced machine learning to uncover both linear and non-linear relationships. The Ordered Logistic Regression (OLogit) analysis identified Improved Customer Service (β=1.365, p < 0.001) as the most powerful predictor of satisfaction, with Brand Reputation (β=0.586, p< 0.01) and Price (β=0.541, p<0.001) also showing strong significant effects. The Multivariate Linear Regression (MLR) analysis specifically quantified the drivers of loyalty, revealing Improved Customer Service (β=0.341, p<0.001) as the strongest predictor, followed by Brand Reputation (β = 0.138, p < 0.01) and Price (β = 0.133, p < 0.001). The OLS regression further confirmed Improved Customer Service as loyalty's primary driver (β = 0.7413, p < 0.001; R² = 0.659). The machine learning models revealed critical, non-intuitive realities. Random Forest not only validated the hierarchy of drivers, attributing 56.0% importance to x Improved Customer Service, but also uncovered the phenomenon of diminishing returns on service quality (R² = 0.718–0.784), indicating that blanket improvements beyond a certain point are inefficient. Most strikingly, the XGBoost model assigned 100% feature importance to Improved Customer Service, suggesting that in the complex reality of passenger behavior, all other factors are subsumed by this central construct. The study challenges conventional wisdom by uncovering a paradoxical positive Price-Loyalty relationship across all models (MLR β = 0.133, OLogit β = 0.541), where Bangladeshi travelers perceive premium pricing as an indicator of quality. Conversely, it exposes a critical "service visibility gap" where Staff Services, despite a strong correlation with satisfaction, demonstrated statistically insignificant impacts on loyalty in both MLR and OLogit models, revealing that frontline efforts fail to translate into enduring allegiance. The implications provide a strategic roadmap for Biman to transform operational challenges into a sustainable competitive advantage. Recommendations include targeted interventions such as AI-driven personalization to overcome diminishing returns, value-tiered bundling to leverage the price-perception paradox, and a comprehensive service excellence framework to bridge the identified gaps. The research establishes a new benchmark for customer-centric aviation in emerging markets, demonstrating the power of a triangulated analytical framework to reveal the complex, non-linear drivers of passenger behavior. Keywords: Customer Loyalty, Customer Satisfaction, Service Quality, Biman Bangladesh Airlines, MLR, OLS Regression, Ordered Logit, Random Forest, XGBoost, Emerging Markets, Aviation Service Optimization.