1. Is there any need for the client to be able to operate “off the net”? Might the client be used to capture data when there is no connectivity? If so, this has to be a rich client.
2. Is there any need to operate in limited bandwidth areas with large amounts of data (e.g., large drop-down lists of personnel, statutes, parts inventories, etc.)? If so, you may need to at least have local “widgets” to cache and handle the lists – i.e., at least part of a rich client.

Quite often, I see web client applications written that “assume” massive amounts of communications capacity (“It worked great in our in-house 100 Mbps network!”). These applications then fail miserably when they encounter real-world conditions outside the lab.

Before choosing an architecture for an application, the business use cases for the application need to be carefully considered.

You may also want to look at Microsoft’s Windows Presentation Foundation (WPF) and Windows Communications Foundation (WCF) layers. Applications properly developed with WPF and WCF components can be deployed in either web client or rich client configurations, but it requires understanding and developing with that in mind.