| Resumo: | In the context of retail analytics, market basket analysis serves as a powerful technique to extract valuable knowledge about consumer preferences and shopping habits. Through its application to a Portuguese retailer, the following study examines purchase transaction data and clusters it based on the product categories in consumers’ baskets. With the goal of mining product relationships, this study compares a heuristic and an association rule-based approach for a cluster-based identification of product substitutes and complements. The paper concludes that for finding substitutes, the heuristic fares the best results. For the discovery of product complementarity, an association rule learning-based approach is suited best. Beyond its theoretical contribution, the insights gained through the analyses are utilized to increase customer satisfaction and sales by providing recommendations on managerial decisions, ranging from determining the timing of product promotions, possible improvements to the store’s design, informing product placement decisions, to suggestions regarding customer communication.
|
|---|