Comparator inserts from different manufacturers can have different sized holes. For example, the Hornady aluminum comparator insert holes are noticeably smaller as compared to the steel inserts sold by Sinclair. As a result, the Sinclair steel inserts seat on the bullet ogive very close to the top of the bearing surface, whereas the Hornady inserts seat noticeably farther out on the ogive. Either type of insert can work, but it is critical to use the same insert consistently, whichever you choose. That way, the readings will be consistent.
@JoeE, the bullet ogive is the the curve that corresponds to the edge of the ogive (nose). The larger the radius of that curve, the longer and more "pointy" the nose will be, generally. A smaller ogive radius will usually mean a short and rather squat bullet nose, like a pistol bullet nose, for example. It's the same as the radius of a circle. The larger the radius of a circle, the less it will be appear to be curved over a given arc length. Likewise, bullet noses have different lengths and ogive radii. Thus, the caliper insert may seat at noticeably different longitudinal points on the ogive radii of different bullets. If used correctly, the caliper insert will always seat at a point of the same
diameter, but that diameter may occur at different points along the ogive radii of different bullets, or at different points relative to the top of the bearing surface.
The key to taking measurements with a tool such as the Hornady OAL gauge is consistency. You do not want to use excessive force on the bullet via the plastic rod when seating the bullet to "touching" because you can actually begin to push it into the lands. Once you have a consistent feel for just barely "touching", you should be able to reproduce the exact same cartridge base-to-ogive (CBTO) and cartridge overall length (COL) measurements consistently/repeatedly. If you can do that, then I would trust your measurements. If the bullet you're using is substantially different (i.e. in terms of weight and/or ogive radius) than was used in the commercial ammunition, it is not at all surprising that the measurements at "touching" would be different. In your situation, you do not know with certainty how close to "touching" the ogive of the commercial ammunition bullet is seated. For that matter, it might even be seated into the lands. If you had some loose bullets of the exact same type as used in the commercial ammunition, you could find the distance to "touching" using the loose bullets with your comparator and then by comparison determine where that bullet was seated in the commercial ammunition relative to "touching".
However, none of that is necessary with your reload bullets as long as you can make consistent measurements with the OAL gauge. If you can, I would trust your measurments and go with what they are telling you, regardless of the relationship of those measurements to the commercial ammunition.
You can seat bullets jumped (i.e. seated off the lands), at "touching", or jammed (i.e. seated into the lands). All are done by various reloaders and all can work in the right circumstances. Each has its own considerations. For example, one consideration with
jammed bullets is that the farther they are seated into the lands, the higher the pressure "spike" will be at initiation, or the early part of the combustion process as the bullet moves into the lands and the rifling engraves the bearing surface. A consideration with bullets seated at
"touching" is that with even as little as .001" variance in seating depth (CBTO) between loaded rounds, some bullets may actually be just off "touching", some actually seated at "touching", and some seated just barely into the lands. Some bullets do not behave well in this situation. For that reason, I rarely seat bullets at "touching", but sometimes they can work well there. Finally, a consideration with
jumped bullets is that the farther the bullet is seated away from the lands, the deeper is will be seated in the case. Thus, pressure can increase if the bullet is seated far enough from the lands, which reduces the internal "effective" case volume. This usually isn't an issue unless one were to
abruptly alter bullet seating depth by something like .030" to .050" deeper in the case without reducing the charge weight first.
Starting with a new bullet seated at .015" to .020" off the lands (jumped) is exactly what I would recommend if you don't already have a good idea where it might want to be seated. You can do your charge weight testing at that seating depth, starting low and working up carefully. Once you move on to optimizing seating depth, I would start by covering a seating depth range from perhaps .003" off the lands to about .030" off the lands in .003" increments. Because the charge weight was optimized with the bullets seated in approximately the middle of this range, it will not usually be necessary to re-adjust the charge weight again after seating depth optimization. In other words, you probably won't have moved the bullet far enough in the neck to cause a change in effective case volume large enough to noticeably alter pressure/velocity. Most of the bullets I have ever loaded have tuned in somewhere within this range, so that's where I would start.
One final thing - keep meticulous records including any measurements you take, seating die micrometer settings, etc. Having good records can make troubleshooting far easier, and it can also help minimize "re-inventing the wheel"; i.e. having to completely re-determine some measurement or setting because you didn't write it down the last time.