Usability tests of new products require the use of methodologies based on how to validate a prototype on the basis of the user's subjective and even unconscious feelings. This is referred to as ``Affective Engineering'' or ``Kansei Engineering''. One of the main steps when applying this approach concerns the semantic description of the product domain. Semantic description of a product is represented by a set of semantic attributes (SA), a set of words describing the domain or the product to be evaluated. A major concern is to establish the importance level of each Semantic Attribute to fit the user preferences as well as possible. The present paper proposes a new methodology making it possible to select design semantic attributes (SA) and exploit data from usability tests to integrate the user's perception and identify the importance and interaction of the semantic attributes. Applied algorithm is based on the fuzzy measures, in particular the Choquet integral. The proposed approach was applied and validated in the design process of innovative insoles. Relevance to industry: The paper proposes a new methodology making it possible to select design semantic attributes (SA) and exploit data from usability tests to integrate the user's perception and identify the importance and interaction of the semantic attributes. Applied algorithm allows designers to validate and adjust product concepts through data resulting from usability tests.