It depends where your rifle, bullet, and charge land in terms of forgiveness. If you do follow the philosophy of "ladder testing" or "optimal charge weight" or "forgiveness envelope" then you should be loading with a charge that will forgive you for a lot more than 0.1grn variability. Alternatively, if you're shooting outside of that window, it might make much more difference for you, or if you're on the ragged edge of that window, then changes in temperature might push you out and into more variability.
Almost all of the electronic measures out there throw +/- 0.1grn, and for what it's worth, even our beam balances (they're balances, not "scales") aren't perfect, such that frictional losses can provide fractional grain deviations that I've found to be in the 0.05 to just under 0.1grn ballpark when they show the same on the indicator.
Reality is that for most of us, we're not shooting consistent internal volume cases, let alone elasticity, nor consistent enough bullet weights (nor metallurgical consistency) such that 0.1grn will not produce as much variability as these other issues.