Just as a suggestion so others don't misunderstand your question, the proper abbreviation and terminology for describing seating depth measurements from case head to the tip of the bullet, is either "OAL" Over All Length, or COAL Casing Over All Length, I've also heard it referred to it as COL also. Just a friendly advisory, and not intended to embarrass or insult your integrity.
Are you using the "FCD" or "Factory Crimp Die"? Or are you just seating them with a standard seat and crimp die?
If your using the FCD, it might be the culprit. I've never used one that I can recall, but if I had a nickel for every time I've heard someone having problems getting them adjusted correctly, I would probably be a wealthy man by now.
Other wise, make sure the ring nut is tightened down good and snug. Lee dies have that rubber insert washer type set up, which if the washer isn't compressed real good and snug against the press, it can cause the entire die to push up when seating, and, or, crimping. This right here could cause a significant variance in "OAL" lengths.
Another possible cause could be that the crimp feature is creating the variance. I'll explain how this happens. If your seating and crimping in the same step, and the brass isn't all trimmed to the same length, the crimp ring will crimp will cause bullets to get seated at varying depths, thus causing variations in finished OAL's.
There are two different methods of correcting this problem.
1. You can trim your brass to the same lengths, then with the ram at full extension, and an empty piece of brass in the shell holder, adjust the die body down until it is just making contact with the case mouth. Then with the seating stem adjusted so it just barely seats a bullet, adjust the die body and seating stem down in small or necessary increments until your crimp and "OAL" are within desired spec.. This is a much more cumbersome and difficult method for most, but it has worked well for me, as I trim all my brass.
2. Or you can back the entire seating die out sufficiently far enough to ensure that the die isn't making any contact with an empty case mouth when the ram is at full extension, then back it out another turn or so to make sure it won't crimp even the longest piece of brass Then place a bullet in a case mouth, seat and adjust the seating stem ONLY until you have the desired "OAL". Then back the seating stem out far enough to ensure it won't can't make any contact with the bullets. Then begin adjusting the die body down until you are crimping to the desired spec.. But because not all the brass will be the same lengths, you'll still have to deal with some inconsistencies with crimps, that's just how it is when not using trimmed brass.
BTW, and just so it doesn't spur an argument about the right or wrong way to load rimless cartridges. Just about every reloader I know DOES NOT TRIM rimless AL-ing cartridges. It's more of a personal preference thing, IMO, rather than a right or wrong way of a reloading procedure. I just happen to think it's worth the extra time to trim, and there is no doubt some additional time involved. But it allows me to seat and crimp in one single step, vs dealing with inconsistent oal's, neck tension variations, and variations in crimps.
FYI, oal variations in particular with the high pressure 9mm cartridge and othe high pressure cartridges such as .40 cal. can result in significant, and even excessive to the extent of concerning pressure spikes, when oal varies by .010" or more. This fact alone has reinforces my decision to trim my brass for more than 30 years.
GS