Do you think multiple profiles are needed to use suspect linking?

While I don't know if absolutely necessary, we use multiple profiles today (we are still on 4.x though, not sure if that would change when we update).  The reason we do this today is that different users care about different artifact changes.  For example, our business folks care about changes between business rules and use cases or stories, while our agile teams care about changes between stories, embedded functional requirements, and work items).

 From an enterprise perspective, a business unit wants to monitor for impact analysis (changes to rules that may affect multiple applications) whereas they do not necessarily care about discrete functional requirements for a single application and the changes between them.

Hope that made sense :).