Machine Learning in Business: Product Bundling Strategy and Customer Segmentation via Market Basket Analysis Algorithm

This study employs the Market Basket Analysis (MBA) algorithm to uncover associations between product categories and item names. MBA aims to discern customer purchasing patterns and segmentation across regions. A company selling building materials in Pontianak, West Kalimantan, Indonesia, has never analyzed sales history data to enhance promotional strategies or gain insights into customer segmentation across regions. This study employs an experimental quantitative study with a company in Pontianak as the study object, using primary data such as sales history, customer database, and product database. The study population comprises 12,600 sales transactions, with a sample of 3,462 transactions focusing specifically on the Onda brand from January 2 to December 30, 2023. The results of the MBA algorithm will be evaluated based on support, confidence, and lift values. From the analysis results, associations between product subcategories and names are identified, providing insights for determining bundling or cross-selling strategies based on consumer purchasing patterns, such as combination angle valve JF 11 with basin tap Y 321 C. Customer segmentation based on consumer interests in each region is also obtained, which can inform the implementation of advertising strategies on social media platforms to bolster product sales and raise awareness.