Influencing factors of Indonesian coffee product customer experience in international market: an aspect-based sentiment analysis with GPT-3 Davinci model

Abstract

International trade is increasingly facilitated by e-commerce platforms, fostering user-generated content through customer reviews, which can be analysed using sentiment analysis to uncover the customer experience. This study applies aspect-based sentiment analysis with the GPT-3 Davinci version to identify key factors influencing the international customer experience of Indonesian coffee products, where Indonesia ranks as the fourth-largest coffee exporter. By analysing 3470 Amazon customer reviews of 14 Indonesian coffee products, key factors emerged across sensory, price, and packaging aspects. Sensory aspects, especially flavour and aroma, were highly rated (84.54% positive), though issues like bitterness and acidity were noted (15.46% negative). Most customers (80.59%) found the pricing affordable, although 19.41% viewed the products as expensive. Packaging received 71.96% positive feedback for sustainability, but 28.04% of customers expressed concerns over functionality. These findings enrich the literature on customer experience and sentiment analysis, offering valuable insights for exporters into marketable coffee attributes in the global trade context.

Keywords

• Customer experience
• aspect-based sentiment analysis
• GPT-3
• Indonesian coffee
• E-commerce
• international trade