I've had a few Lot#s of Varget that differed in velocity from others by as much as 50-75 fps for a given charge weight with all the Lot#s for the other components (brass, primers, bullets) being identical. However, it's much more common to see Lot-to-Lot powder differences in the 5-10 fps range, or well within the velocity ES/SD values for the different Lot#s, meaning the burn rate of the new Lot# is statistically pretty close to the previous Lot#. The problem is, there is no way to know for sure with a new Lot# until you determine an average velocity for a known charge weight.
Because I use QuickLoad to expedite/facilitate the reloading process, this is generally not a problem. I initially use a reduced charge weight to "calibrate" the program, adjusting the burn rate factor (Ba) until the velocity predicted by the program exactly matches the average [measured] velocity from the reduced charge weight of the new Lot# of powder.
Fortunately, it is not necessary to use QuickLoad to do the same comparison, and it is quite easy, especially if you have the charge weight of a tuned load with previous Lot#s of the same powder. Simply choose a reduced charge weight as compared to the previous tuned load that will generate safe operating pressure even if the new Lot# of powder is markedly faster than the previous Lot#. I generally reduce the charge weight from a tuned load with a different Lot# of the same powder by 2-3%. However, you could increase that a little more for safety purposes if desired. Load a sufficient number of rounds for a few foulers and 5 additional shots for a velocity determination, then determine the average velocity with a known charge weight and record the velocity data in a data book for that rifle where you will always have it. If you use the same reduced charge weight for a given rifle/cartridge (i.e. a standardized value) to carry out this test every time you open a new Lot# of powder, you will eventually have a nice little database of Lot-specific velocity-per-charge weight values for future comparison. Yes, this approach requires an "extra" range trip, but it's only a few loaded rounds and the velocity data will allow setting up a charge weight test in a more informed and safe manner.
A common alternative to the focused approached outlined above is simply to always start at a sufficiently low charge weight in the initial charge weight testing, whether using a value from a reloading manual, QuickLoad estimates, or similar. If different manuals provide noticeably different starting charge weight values, my feeling is that going with the more conservative choice of which set of values to use is never a "wrong" way to go. If the more conservative starting charge weight ends up being far below the intended charge weight, so what? It's a few extra loaded rounds, which in my opinion is well worth it for the information gained when before doing the testing all you really had was a black box due to the unknown burn rate of the new powder Lot#.
As long as you start with a sufficiently low/conservative charge weight, the more important consideration using the second approach becomes the upper end of the charge weight range. In other words, how wide does the charge weight test window need to be? How high do you actually need to go? If the new Lot# of powder is faster than the previous Lot# and you go too high, you will be limited to using fairly gross observational data to decide whether some of the charge weights at the top end of the window are too hot and should not be fired. For example, many use hard bolt lifts or pressure signs on the brass as indicators. Unfortunately, they can work, but neither of those are really reliable indicators until pressures are already well above MAX. If you know the velocity of a previous tuned load using the same powder, it should not be too difficult to make an estimate of when it's time to stop going any higher. However, the first time out with a given powder making that decision may be a little more difficult. As always, erring on the side of caution at the upper end is the best approach.
