I doubt the stem is the problem - -
I have been measuring bullets for a couple of years now with a comparitor; the ones with the Granite base and fixtures, one on the indicator tip and one that holds the bullet and slides around on the granite base.
It is not all that un-common to find hand made bullets that will not give consistent seating depths, in fact, it pretty much the norm. Once sorted, they all shoot just fine and the sorting makes it very easy to load with consistent deating depths.
Occasionally, one finds a lot that are within a couple of thousants but that is UNUSUAL. Usually, most of the bullets will be within a couple of thou but the real problem is a quanity of them will be off by a lot. You wonder why that one went way over there and you didn't see it in the flags? Guess what, the flags didn't lie, the seating depth was too far off to be accurate.
I bought a box of bullets recently that were .021 different from one side of the dial to the other. the biggest number of them were .001 different BUT about half of them are the different ones and would KILL ones day.
I was recently given a box of 150 bullets that were made in 1998. I measured them and they were within .001, about 2/3- 1/3. The best I have ever seen was a box of samples sent to me ( 100 ) that were ALL exactly the same. NOW, if we could get that all the time, we could load randomly without any concerns at all. But ALAS - - -!
A word of caution; if one's necks aren't uniform, seatig depth will be affected by differing neck tension. One must be vigilent.