Clustering with low silhouette scores

Question

I tried two clustering algorithms (Kmeans and Kmedoids) on my dataset. The silhouette scores are normally used to determine the amount of clusters