A longitudinal consumer feedback analytics approach for theory building and testing: Examining consumer satisfaction and loyalty in buy now, pay later (BNPL) platforms

This study investigates the post-adoption factors influencing consumer satisfaction and loyalty in Buy Now, Pay Later (BNPL) platforms, with a focus on the culturally specific and religiously regulated context of Saudi Arabia. A longitudinal dataset of 61,631 cleaned and validated bilingual user-generated evaluations from prominent BNPL applications, collected between 2018 and 2025, was analyzed to capture temporal shifts in consumer perceptions. A mixed-methods approach was employed, combining topic modelling, sentiment analysis, and large language models to identify underlying experiential themes and map them to theoretical constructs. Confirmatory analysis using Elastic Net regression and Random Forest, with 80/20 train–test splits, 10-fold cross-validation, and bootstrap resampling, demonstrates that Information Quality and Perceived Security are the strongest predictors of Consumer Satisfaction, while Sharia Compliance functions as a cultural boundary condition and significant driver of Consumer Loyalty. Perceived Financial Risk consistently exerts a detrimental effect on both outcomes. The findings extend retailing and consumer behaviour as well as IS/FinTech literature by incorporating culturally specific constructs and by demonstrating how user-generated content (UGC) analyzed with machine learning, providing unsolicited, ecologically valid, and large-scale insights into post-adoption behaviour. Practical implications highlight the importance of transparency, ethical adherence, and financial clarity in enhancing the BNPL user experience, while also emphasising cost–benefit considerations, regulatory compliance, and consumer segment differences in implementation strategies.