Creating a Safe and Sustainable Omni-Channel Retail Ecosystem Using Predictive and Prescriptive Machine Learning
Abstract
The retail ecosystem has significantly evolved from simple neighborhood stores to large omni-channel retail systems. With razor-thin margins, the retailers need to constantly innovate to reduce losses from fraud, shrink, waste, lost sales opportunities, out-of-stock and many other scenarios. Recent advances in omni-channel retailing have also led to the rise of a new group of fraudsters who abuse the gap between the online and offline channels and take advantage of the relaxed retail company policies. Fraud can be related to payment, account take over, refund and cancellation abuse, associate collusion, marketplace, and even call-center abuse. Waste can be due to over-production, sub-optimal pricing or discounts, damaged products, throwaways, availability issues, packaging waste.
The paper tries to identify the need for better research on improvement in fraud detection systems for sparse data, improvement of the customer experience and reduction of returns. The paper discusses how the advanced predictive and prescriptive ML models can help in building tools and recommendations which reduce the losses and make the retail ecosystem safe from fraud and sustainable against wastage.
The data analysis is performed on omni-channel retail transactions over a period of two years across multiple countries, and feedback is collected from total loss analysts and store associates through interviews, surveys, and interactive tools. Machine Learning methods of supervised and unsupervised learning including an ensemble of anomaly detection, graph analytics, classification models with gradient boosting, reinforcement learning, can be used to highlight the high priority cases to focus upon, and then identify the root causes of such losses. Then, Causal Discovery models can be used to identify root causes and provide prescriptive recommendations. The ecosystem of ML assisted solutions can support in multiple aspects of running retail store operations starting from inventory management, predictive replenishment, shrink reduction, effective disposition, waste management, and finally combating different malicious fraud and abuse activities, all of which lead to different types of loss to the retailers. The paper provides a comprehensive evaluation of the challenges in omni-channel retail ecosystem and the proposed machine learning tools which can help resolve the issue at large. It also proposes an effective impact measurement solution based on Causal Inference techniques which can easily measure the impact of the changes made on important KPIs like savings, wastage, returns, etc. and tailored to both online and offline retail channels.
The implementation of ML techniques in retail total loss management can lead to multi-million-dollar savings through better fraud detection, inventory management and waste management. Performance of supervised models improves to close to 98% over the course of months by using feedback loop.
