We present a new factor analytical technique for three-way data based on Triadic Concept Analysis. Previous works in this direction focused on reducing the complexity of triadic data by using triadic concepts. In this paper we propose a method of grouping these concepts into well-structured families such that the factors can be interpreted as many-valued ones. Thereby we lift the already existing theory of ordinal factorisations to the triadic setting.