Many factors could come into the variation in case weight before an observable difference in case volume becomes obvious - as I mentioned earlier the average plant producing cases will have a lot more than one machine producing a type of case to get the volume they need to fill orders. Mil brass especially (going back to my Lake City milsurp experience).
It would not take much of a difference in punch and die set ups to create a minute difference in case head thickness, each machine would have a different roll of feedstock metal that is being made into the cases (slight difference in metal composition will give you a different density of metal), and the extractor groove cut and rim chamfer would all play a part in this. The main thing here is one grain of metal is a hell of a lot less than one grain of powder so you might be able to see a big difference in batch weight of cases without seeing a large difference in internal volume.
What that could do is give you cases which resize and age differently than others as the factors that can change if cases are coming off different machines include things like different annealing cycles so you are starting with different hardness and ductility in amongst your batch of cases giving you different neck tension and different obturation and springback amounts potentially.
We can't really control this though, and most people I think would not be using big enough sample sizes to get a handle on if there are weight groups in the batch of cases - about 500 of the LC milsurp cases showed 5 distinct groups where the case weight was within 2gr in the batch, but there were if I remember right something like 4 or 5 grains at least between batches. It was noticeable at any rate (all cases cleaned, trimmed and resized with the primer pocket uniformed/decrimped prior to weighing).
My main point is internal volume is only one thing that might affect case performance.