This case study examines how NovaMart, a large omnichannel retailer, leverages artificial intelligence and machine learning to transform customer management from a largely intuition-driven activity into a data-driven and predictive process. The case focuses on the complementary application of unsupervised and supervised learning. NovaMart first integrates customer information from transactions, digital platforms, loyalty programmes, and customer-service systems to create a unified analytical data foundation. Using K-means clustering, the company identifies five behavioural customer segments: Premium Loyalists, Promotion Seekers, Occasional Explorers, Digital Enthusiasts, and At-Risk Customers. Classification models are subsequently developed to assign new customers to these segments and predict outcomes such as customer churn and campaign response.
The case highlights how analytical insights can be embedded into customer relationship management through personalised marketing, targeted retention interventions, and differentiated customer treatment. It also examines challenges involving data quality, variable selection, model interpretation, class imbalance, model drift, privacy, fairness, explainability, and organisational acceptance. The case demonstrates that successful AI/ML adoption requires more than algorithmic performance; it depends on integrated data infrastructure, governance, cross-functional collaboration, continuous model monitoring, and managerial judgement. Overall, NovaMart illustrates how clustering and classification can jointly enable scalable, predictive, and responsible customer management.
Data sets for the case study
Tech Talk 