The paper presents a new method for identifying indicator species and species assemblages that characterize groups of sites. The method combines species relative abundance with their relative frequency of occurrence in site groups, resulting in a symmetric indicator value. The significance of these values is assessed using a randomization procedure. Unlike TWINSPAN, this method is independent of other species' relative abundances and does not require pseudospecies. It can be applied to any hierarchical or non-hierarchical classification of sites and provides criteria for comparing typologies, identifying where to stop dividing clusters, and determining main levels in hierarchical classifications. Species assemblages are grouped based on their indicator values at each clustering level, preserving the heterogeneity of species observed in each site. The method is illustrated using a dataset of carabid beetle distributions in Belgium. The paper discusses the limitations of TWINSPAN, including its assumption of a dominant gradient and arbitrary cutting points. It also highlights the importance of considering species' ecological breadth and the need for standardized methods in assessing representativeness. The new method is compared to classical approaches and is shown to be more effective in identifying indicator species and assemblages. The results demonstrate the method's ability to detect species characteristic of specific habitat groups and to identify hierarchical structures in species distributions. The method is applied to a dataset of carabid beetles in Belgium, revealing species associated with different habitat types and highlighting the importance of considering ecological factors in species distribution analysis. The paper concludes that the new method provides a flexible and asymmetrical approach for identifying indicator species and assemblages, which is essential for ecological monitoring, conservation, and management.The paper presents a new method for identifying indicator species and species assemblages that characterize groups of sites. The method combines species relative abundance with their relative frequency of occurrence in site groups, resulting in a symmetric indicator value. The significance of these values is assessed using a randomization procedure. Unlike TWINSPAN, this method is independent of other species' relative abundances and does not require pseudospecies. It can be applied to any hierarchical or non-hierarchical classification of sites and provides criteria for comparing typologies, identifying where to stop dividing clusters, and determining main levels in hierarchical classifications. Species assemblages are grouped based on their indicator values at each clustering level, preserving the heterogeneity of species observed in each site. The method is illustrated using a dataset of carabid beetle distributions in Belgium. The paper discusses the limitations of TWINSPAN, including its assumption of a dominant gradient and arbitrary cutting points. It also highlights the importance of considering species' ecological breadth and the need for standardized methods in assessing representativeness. The new method is compared to classical approaches and is shown to be more effective in identifying indicator species and assemblages. The results demonstrate the method's ability to detect species characteristic of specific habitat groups and to identify hierarchical structures in species distributions. The method is applied to a dataset of carabid beetles in Belgium, revealing species associated with different habitat types and highlighting the importance of considering ecological factors in species distribution analysis. The paper concludes that the new method provides a flexible and asymmetrical approach for identifying indicator species and assemblages, which is essential for ecological monitoring, conservation, and management.