Take a fired case, with the neck partially resized. Put a long flat based bulllet in nose first, and chamber the round. When you extract it, it ought to reflect the distance from the bolt face to place where the rifling begins (the leade). Put that bullet where it won't get lost of banged around. Take another bullet from the same box and insert it as square as you can into the muzzle until it contacts the rifling, push on it a bit and twirl it around so the rifling engraves a mark on the ogive of the bullet. Now seat this bullet so that the distance between the base of the case and the ring around the bullet are the same length as the first dummy round you made. Set up your seating die using this second round. Now any bullet seated by the die will touch the rifling in the rifle. Turn down the seating stem until it will seat the bullet of the second dummy a bit further into the case. Measure the base to line distance. However much shorter it is will be the difference between the bullet and the rifling. I generally start a 0.20" bullet to land distance and shoot a bunch of these loads. If I'm not satisfied, I seat the next batch of bullets a bit deeper, and so on, and so on, until I'm happy. Then I save the second dummy round with the bullet info scratched on the case. I repeat the process for every bullet I load for, keeping the dummy rounds to simplify setting up the seating die when I want to load some more rounds with that bullet.
Or you can by the Sinclair oal tool and follow the directions.

As for the divergence in the bullet strikes on the target I can't help you each rifle barrel vibrates differently with different bullets. Perhaps you can find a way to dampen the barrel vibrations, like the BOSS system. If you don't sell these dampeners to your friends and neighbor's you won't be violating the patent.
Lastly, good luck.