Creating Transparency on Product Variety Through Data-driven Similarity Analysis

Abstract

In recent years, the number of product variants has steadily increased in numerous industries to accommodate customer-specific requirements. At the same time, rapid technological changes have led to shortening product lifecycles and greater volatility in product portfolios. In manufacturing, the high level of product variety is reflected in greater process diversity and material flow complexity, making it harder to exploit economies of scale and utilize resources efficiently. In order to stay competitive under increased cost pressure, manufacturing companies try to counter these effects, but often lack transparency on product variety to successfully implement modular systems and standardization measures. As existing methods for capturing product variety are labor intensive and impractical for larger portfolios, this paper presents a data-driven approach by computing similarities between products based on their geometrical features. To capture product variety and create transparency, the similarities are visualized through a dendrogram and multidimensional scaling methods.