This one surprised me too when I first tried it. The way I like to do it (much has already been covered):
- Dry room for flexibility in post
- Multiple mics for more sound difference (a condenser with a dynamic might be good - different takes)
- Recording samples at different distance from the mic (I found this was minimal in effect in a dry room)
- Make sure your timing ISN'T perfect! This is a big one. You need to record (or shift in post production) some takes a little early, some a little late. If you're a good muso and your claps are all perfect, it sounds thin and too snappy (unless you want that)
- Lots of takes
- Different samples for each clap (as Jeff pointed out) is very important to not sound stale and robotic
- Pan individual claps differently
- Group bunches of claps into close, mid and far where the close ones will have minimal stereo spread, a short reverb pre-delay and low mixing, but the far ones will have a longer and less bright verb with more pre-delay and higher verb mixing. You might also want to put the majority of your samples into this section. Eg put 4 claps in front, 10 in mids, and 20 claps in the rear.
- Clap differently. Palm of hand, on the fingers, flat hand, cupped hand etc
- Layer with sampled claps
- Lots of compression may help thicken it out a lot, but I tend to find if the rest isn't right, all this really does is bring up the noise floor without adding too much to the weight of the clap.
Anywho, a combination of those things tends to work well for me.