Most of the time you've got to create a schedule before you know exactly what the work entails, more often still you're given a deadline and have to make the production fit.
In the latter case, time can help dictate production methodology, but the day to day schedule might not meet your prediction.
For example, in our recent motion graphics for Apartment Finder we predicted we'd storyboard everything in three days (which we did) and follow that up with color models proceeded by motion tests. Turned out we went right into animating the first segment. It was just more sensible.
That could have been a disaster if the client hated the work. We were sure they'd like it and that this would be simplest presentation. This method allowed us to produce the work over a week ahead of schedule although the daily workload bore no resemblance to the initial schedule.
The rule of thumb for drawn animation is that it will take on animator one week to produce 15 seconds of footage. You can usually take that same amount of time for art production.
These days drawn animation mixes so much with digital animation that scheduling gets hazy. Things that would be done in as pencil tests or through complex art production procedures (photocopied zooms, multi-layered pans, flipped or reused drawings) are done in conjunction with the animation stage usually with final art.
Then if you get into Flash or utilizing the keyframe elements of AfterEffects scheduling becomes even more nebulous. Once you're into to production its predictable, but before the work starts only experience can give you a general idea of how long it will take.
We like to use good ol' Microsoft Excel to build our schedules. It's brutish, but infinitely expandable. Thirteen segments of a film can line up in order and you can match that up with the pipeline of a thirty second commercial, some web shorts, and some kids' interstitials to make certain the proper staff can be predicted and your artists can flow from one production to the next.