Cluster analysis and single dominant factors

I often chat with researchers wanting to use cluster analysis to group cases. I just wanted to point out a common scenario where cluster analysis may not be a good way of proceeding.