Predicting Accounts Receivable with Machine Learning: A Case in Malaysia

Abstract

Accounts receivable plays a major role in credit to cash conversion cycle which involves collection management, payment management and debtors aging. Lack of visibility on accounts receivable management limits efficiency in collection management leading to long aging debtors. Aging payments eventually turns into bad debts leaving a negative impact on the cash flow. The case study company, PangCo Jaya is a wholesale consumer goods distributer located in Malaysia. The objective of this study is to predict payment timing of PangCo Jaya’s customers. The outcome of this case study enables collection team to plan for proactive debt collection. This study applies Machine Learning techniques to address accounts receivable challenges faced by PangCo Jaya. The tools used in this study is Python in Jupyter Notebook. A supervised Machine Learning classification model for predicting invoice payment prediction has been developed. This solution enables proficient payment collection by reaching out to targeted customer with potential payment delay prediction. In future, these trends can be applied to evaluate if there is a necessity in revising customers’ credit policies and estimate potential receivables that are under risk of turning into bad debts.