Actually, if you pull military surplus ammo, you often see large charge extreme spread of a couple of grains. People talk about staying within .1 grains, but they use that same number whether they are throwing 2.7 grains of Bullseye in a .38 Special wadcutter load, or if they are using a magnum rifle case holding 70 grains or more. Obvious, as a percentage, there is a big difference. A rifle will let you see smaller percentages of variance on a target than a pistol will, but if this is a hunting rifle, I think you'll find at the load levels you are using, if your charge has an extreme spread of .5 grains, you'll be hard put to tell the difference from one charge to the next on paper. 0.1 grains may matter to a bench rest shooter, but those of us shooting service rifles or hunting rifles just can't see those small differences when shooting from positions that don't include a bench. Besides, outdoor conditions constantly change. You don't really want a load so touchy you can see .1 grains difference in charge. I like
Dan Newberry's load development method that identifies charge-insensitive loads on purpose.
That said, you'd still like to do better than nearly two grains, ES in charge weight, I'm sure. With a stick powder, tapping is not a good idea, IME, unless you have some way to make it extremely uniform, such as a motorized vibrator like a fish tank pump that has a timer to operate it exactly the same amount of time for each throw. It is too easy to make stick powder settle at different bulk densities. Try using a long drop tube some time to put powder in a case, as compared to no drop tube. Even if you put two identically weighed charges in, the drop tube charge will not come up as far inside the case. It has packed down its bulk density.
Powder baffles, indeed, two of them at 90 degrees to one another, seem to do best for me. You can download a set of instructions and templates for making them for different ID powder hoppers at
my file repository.