Work

Leveraging AI and Deep Learning for ECommerce Customer Segmentation

Year: 2023
Type: article
Abstract: As e-commerce platforms compete to retain customers, predicting and preventing customer churn remains a critical challenge, particularly among new shoppers. In this paper, leveraging anonymized large-scale transaction records from an e-commerce platform, we develop ClusChurn-EC, a systematic two-step framework for interpretable customer segmentation and churn prediction. The framework first clusters new customers into distinct behavioral segments based on purchasing patterns and engagement metrics. We then design a deep learning pipeline using LSTM networks and an attention mechanism to predict customer churn with minimal initial interaction data, effectively capturing sequential shopping behaviors. By identifying at-risk customer segments, ClusChurn-EC enables businesses to implement targeted retention strategies. Extensive experiments on real-world anonymized transaction data demonstrate that ClusChurn-EC provides actionable insights into consumer behavior and outperforms traditional machine learning models in churn prediction. The framework is designed as a scalable analytics pipeline, supporting real-time decision-making for retention management. Additionally, it is adaptable to a wide range of e-commerce platforms and online marketplaces, making it a practical and generalizable solution for churn prediction in digital commerce.
Language: English
Subfield: Marketing
Open Access status: diamond
OpenAlex ID: W4410696446
FWCI: 0.902
Cites: 51
Cited by: 8
Related to: 10