Predicting online shopping cart abandonment with machine learning approaches

Abstract

Excessive online shopping cart abandonment rates constitute a major challenge for e-commerce companies and can inhibit their success within their competitive environment. Simultaneously, the emergence of the Internet’s commercial usage results in steadily growing volumes of data about consumers’ online behavior. Thus, data-driven methods are needed to extract valuable knowledge from such big data to automatically identify online shopping cart abandoners. Hence, this contribution analyzes clickstream data of a leading German online retailer comprising 821,048 observations to predict such abandoners by proposing different machine learning approaches. Thereby, we provide methodological insights to gather a comprehensive understanding of the practicability of classification methods in the context of online shopping cart abandonment prediction: our findings indicate that gradient boosting with regularization outperforms the remaining models yielding an F1-Score of 0.8569 and an AUC value of 0.8182. Nevertheless, as gradient boosting tends to be computationally infeasible, a decision tree or boosted logistic regression may be suitable alternatives, balancing the trade-off between model complexity and prediction accuracy.