cms tend to fail because people don't bother to understand the culture within the organization, as if (since it's not customer-facing) it somehow doesn't count. it would be interesting to discover what the success factors of content management might be if some sort of user research took place before the CMS tool or system or requirements were chosen.

my thinking is that if you look at the culture, the people, their tasks and information, you might actually come up with a way to manage the content appropriately for the company. otherwise, it's the age-old problem of throwing software at a problem. and that, as we well know, is nothing new.