Flerkriteriebaserad produktklassificering i högmix-lågvolym-miljö : En empirisk jämförelse mellan traditionell ABC-analys och flerkriteriebaserad klassificering som beslutsstöd för differentierad lagerstyrning

Sammanfattning

Inventory management in high-mix-low-volume (HMLV) environments is often complex due to large product variety, intermittent demand patterns, and varying lead times within the article portfolio. Traditional ABC classification is commonly used as decision support for inventory management, but it is typically based on annual consumption value and therefore risks providing an overly simplified representation of the articles’ relevance for inventory control. The purpose of this study is to empirically evaluate a multi-criteria classification model as decision support for differentiated inventory management in an HMLV environment. The study examines how the classification results differ from traditional ABC classification and how the results can be used as a basis for inventory management decisions. The study was conducted as a quantitative case study at the company Elpress. Article data from the company’s ERP system were used for the product classification, where both a traditional ABC classification and a multi-criteria model based on weighted linear aggregation were applied. The criteria included demand, value, and lead time, where the weights were calculated using the entropy method. The quantitative analysis was complemented by a semi-structured interview to assess the practical relevance of the results. The results show clear differences between the models, where the traditional ABC model produced a more concentrated classification with a large proportion of articles placed in class C. The multi-criteria model distributed more articles into classes A and B, as articles classified as C in the traditional model received higher priority when multiple criteria were considered. The study shows that multi-criteria classification can provide a more nuanced decision basis than traditional ABC classification, but also that the model should not be used as an automatic control rule, rather as support for further analysis and differentiated inventory management.