Recently at a seminar I taught, we went off on a tangent about PPAP and all the other associated statistical forms and requirements to get a part accepted; but before I go on a rant, an illustrative story:
I'm a shooter. Occasionally at the range, we have a contest. We set an 8-inch target 100 yards down range. At that distance, it looks like a period typed on a piece of paper. The contest is who can hit the target the most times with 20 shots, using their off hand and open sights. Loser buys dinner. Rifles today are dead accurate. The only real variable is the shooter's ability.
Now the rant:
When molding parts, the 'target' is making an acceptable part consistently. You get no extra points for making perfect parts where every dimension is dead on at the mean dimension (the bulls-eye). Like the targets on the range, the object is to not miss the target (put every part within the specified tolerance). You get bragging rights only if you hit the bulls-eye; but nothing more....