The pre-programmed holds (as opposed to unplanned holds) exist so that if there were any unexpected temporary hitches in the launch sequence, there is still time to make the launch window. Imagine for example that you have a 1 minute launch window, and a slower than expected fuel pump means that it takes 5 minutes longer to top off the fuel tanks. Well, you just missed your entire launch. Stand everyone down, and put the spacecraft back into standby before trying again tomorrow. If you have a planned hold, you simply reduce the planned hold time, and you can still hit the launch window.

The hold is also used as a sync point between all of the different activities that have to be completed at certain points in the launch preparation. For example, at the t-4 hold for the Curiousity launch, they needed to pass the go/no-go from the entire team. If one team had been running late up till that point it wasn't a problem, the other processes just entered the hold earlier. Once the last hold has been cleared, all of the remaining processes have to stay synchronised.

Basically, holds are used to give the personnel some breathing room to make checks and ensure everything is okay (and presumably to relax a bit, go to toilet, and other human necessities). If something does not check out, a hold could be extended to debug the problem - on the other hand, when the clock is running, the schedule cannot be changed because many things are preprogrammed in the launch vehicle and the spacecraft to occur at certain clock ticks.

There are very specific things that have to happen at certain points during the countdown. I expect that the plan to have the 10 minute hold was created after the software was configured for all the last minute tests. Reconfiguring all the software would be impossible right before the launch, so instead, they stop the clock. Stopping the clock is not unusual, and all the software that works off the clock is built to deal with a hold.