Another thing that helps is using graphite lube inside the necks. I started using once I had issues with bullet weld, and it drastically reduces the resistance when seating. And if I'm trying to get zero variance, I take an extra second to rotate the case 180° and run it through again, and that produces almost exact specs.I measure mine base to ogive and still get them all over the board.
I'm still confused by how much variance you're getting, do you have the stem screwed too far out to where it's starting to crimp just before you reach the full stroke on the arm? Or are you purposely crimping?