I'm working on that 11-Second-Club animation right now, and I thought I'd say something (ha!) about lip-syncing. We only really had one or two classes at AIS involving lip-syncing, and I would argue that was good, because I do think it's takes a back seat role to physical body animation.
With that in mind, a lot of what I know about lip-syncing I figured out by actually animating lip-syncs, and I've found a lot of the stuff I learned coincides with what you're supposed to be doing. Here's one thing I've found to be truly important about lip syncing (and, blah, I always remember it 100 frames into animating...). With fast-talking dialogue, you have to understand how it slurs, because the slurred form is how you'll end up setting keys.
Right now I'm animating the following line: "And you want to call it a health problem." To mark this phonetically every time a unique syllable comes up in writing is ludicrous. You'd be doing this:
"Aah-nn-duh Yee-oo Ww-ah-nn-tuh tuh-ook Caaah-ll..." so on and so forth.
But when you slur it a bit, it comes out looking much more natural, because in fact we say 90% of things slurred.
"An Oooh Ah NN Ahk Ah it," so on and so forth.
The jaw is especially important. It opens and closes far less than we think. If you do it every time a new letter shows up, you'll get speed-racer talk.
So in conclusion, just like most animation, less keyframes is better, counterintuitive though that may be. Now hopefully by writing all this down, I'll remember it on frame 1 next time...