Product-describing attributes possess a hierarchical structure ranging from concrete to abstract attributes. For marketing managers to successfully position a product, they must link the product's concrete attributes to the desirable outcomes or values that consumers seek when consuming the product. The most commonly used techniques for analyzing attribute hierarchies are the laddering technique and hierarchical cluster analysis. However, the laddering technique, a qualitative analytical method, has the drawback of relying heavily on the analyst's subjective judgment, thereby lacking objectivity. Moreover, hierarchical cluster analysis suffers from the problem of losing information related to asymmetry—which represents the hierarchical nature among attributes—during the data transformation process. This paper introduces a method that can objectively identify both the horizontal and vertical (hierarchical) structures of attributes simultaneously by employing asymmetric multidimensional scaling, which adds parameters representing the causal levels of attributes to conventional symmetric multidimensional scaling. This method is then compared with existing hierarchical structure methods based on the laddering technique and cluster analysis through illustrative examples. Unlike the laddering technique, this method enables the construction of hierarchical maps based on a quantitative and objective model, and unlike hierarchical cluster analysis, it has the advantage of allowing data analysis without losing the causal relationship information (asymmetry among attributes) inherent in the raw data.