Determining combination of items for product bundle offering using market basket analysis

Abstract

Determining the combination of items for product bundle offerings is important to increase consumer buying interest. By analyzing historical data on sales transactions, purchasing patterns can be identified to find the right product combination that has a high probability of selling. This study aims to determine consumer purchasing patterns on large transaction data using market basket analysis with the help of RStudio Software. In this study, the pattern of purchasing fashion products is revealed from sales transaction data throughout 2020 using the a priori algorithm. The sales transaction data contains 262 fashion products that have been successfully transacted by consumers. Each product has a type of size ranging from medium, large, and extra-large. The transaction data contains 816566 completed transactions. This transaction data will then be processed by the a priori algorithm with a minimum support value of 0.003 and a minimum confidence level of 0.8, found 15 product combinations. Based on this combination, this study suggests that for each product combination it is recommended that there are black, white, maroon, and navy products. This study might be helpful for the fashion producer to increase customer buying interest by offering product bundle promotion.