Integration of Bundling Product Recommendation System and Stock Management in Fast Fashion Retail Business

Fast fashion presents products with designs that change quickly according to trends but at lower prices. This condition forces fast fashion business actors to update their stock quickly. However, uncertain public demand often results in piles of stock in the warehouse. One way to overcome this is to provide discounts through bundling product sales. In order for bundling product recommendations to be right on target, it is necessary to consider the level of complementarity of product combinations with public interest. The complementarity level of product combinations by public interest can be obtained by analyzing sales data. In this study, a bundling product recommendation system was developed and integrated with the purchasing and sales information system. Bundling product recommendations are carried out using the apriori algorithm. The recommendation system was applied using sales data from the information system between December 27, 2023 and January 2, 2024. The minimum support used was 0.1, and the minimum confidence used was 0.4. The system produces three bundling product recommendations within 30 seconds and showing that most consumers are inclined to buy a cloth set (tops and pants) and prefer neutral colors, namely white and black. The time required by the system to generate recommendations is 97.6% faster than if done manually. In addition, four fashion store owners validated the developed information system and the recommendation results. The validation results stated that the system facilitates business administration, monitors stock, analyzes consumer behavior, and develops marketing strategies with bundling promos. Finally, integrating the information and recommendation systems can increase data consistency and reduce data input time by 50% compared to systems developed separately.