"Make up numbers" glosses over a bunch of actually-useful techniques, though. You might not have any data to begin with, but you can certainly get some.
The best technique from that set, I usually find, is betting (with actual money) on results: setting up a intra-company prediction market for when your software will ship can tell you more than you ever wanted to know about how much time it'll truly take.
For example, since a time estimate can trivially be converted into probability estimate for something happening on time, all of this is relevant: http://lesswrong.com/lw/3m6/techniques_for_probability_estim....
The best technique from that set, I usually find, is betting (with actual money) on results: setting up a intra-company prediction market for when your software will ship can tell you more than you ever wanted to know about how much time it'll truly take.