It's nice to see Drupal mentioned in the context of "cool thing to do" but I don't think Cindi Trainor in her piece on Sacred Cows in Library IT gets exactly what you can do with it:
Experimenting with low-cost or no-cost tools like Twitter will only cost staff time, but implementing expensive (think federated search) or complex-but-free technologies (think Drupal) because it's the cool thing to do can be a very costly lesson for a library to learn, in terms of budget, staff time, morale and user satisfaction.
First, there's no impediment to 'experimenting' with Drupal any more than there is to experimenting with Twitter. The first implementation I ever dealt with was on my own laptop. I didn't even need a network connection!
Also, as far as complexity goes, what are we comparing it to? I mean, you can't run a website on Twitter so that's not an option.
It just so happens that an institution's website is a fairly complex organism. It's going to involve a considerable investment no matter how you choose to go about it.
The fact that Drupal can potentially make it less costly in terms of budget, staff time, etc. -- while being far more effective as a tool -- that's what makes it "cool" and why people choose it. Not the other way around.
P.S. It's kind of ironic that the above quote fell under the Sacred Cow, "Cutting-edge is better; bleeding-edge is best" -- considering that the piece grew out of a discussion on the oh-so-bleeding-edge "Google Wave".