Where I work, I've noticed several trends:
- Duplicate work; as long as the wiki doesn't replace existing processes, placing things there is simply additional work. Moreover, because it's additional, there's no guarantee that any desired piece of information will exist in the wiki.
- People don't like not having an "owner" to a document, especially regarding private edits and such.
- Wiki's make you search for a topic (i.e. they require work), whereas email threads are pushed to you by others motivated enough to CC. Unfortunately, my motivation in tracking an issue will never be the same as the person directly affected by the issue.
- Mostly, however, wiki's don't provide a substantial functionality improvement to what I'm now going to dub the "enterprise wiki": i.e., Word documents on a shared drive. Here, Windows provides structure (well, technically Solaris does, but it's accessed by the user via Windows Explorer), while Word provides the standard document editing interface (although we do work with many document types).
The main drawbacks are terrible search, manual (and content limited) revision control, and an inability to track updates.
The main upsides are a rich editing environment for documents (compare Word to a textarea), support for any document format (don't like doc files? Use something else), and the opportunity for private collaboration & edits without implementing temporary ACLs (i.e., copy a file to the local drive, and email back and forth until publishable).
Personally, I think wiki's are great for collaboration between a heterogeneous, dispersed population. However, at the large corporation where I work, Word + Explorer is a ironic example of the worse-is-better Unix approach (many tools, barely working together) clearly beating the wiki/web-based Monolithic design.