I would plan for this machine to machine variability. In fact, even the same machine cutting a few days apart will also generate much variability. And since you are manually focusing it, that introduces much variability that can also affect the kerf.
The thicker your material is, the greater your kerf is likely to be (you have to put more laser power and hence more heat into your part to cut it). Also, remember that the laser beam is conical and will cut a tapered kerf. This is more visible on thicker materials. I think this is one reason why people sometimes focus in the middle (depth) of their material instead of the top face.
When I have parts whose dimensions are critical I cut a 1" square and measure the kerf from that. Then I diddle my design to account for that measured kerf and I cut it right away. This is a serious PITN … I think there might be some design programs that allow you to enter a kerf but I don’t happen to have one of those.
BTW, .006 - .010" kerf is fairly typical - at the high end on thicker pieces, at the low end on thinner pieces. If I can live with a couple thousands tolerance I just guess - I decide which is worse; oversized or undersized and then I just use 0.008.
Remember, however, that the lasers don’t necessarily cut the same in the upper LH corner of the bed that they do in the lower RH corner … they usually take more power in the lower RH corner. I don’t know how this relates to the kerf you will see. I could guess … Anyway, cut your 1" kerf test in the same spot where you will cut your “real” part.