Because .308 has more brass capacity variation than most other chamberings except .300 Win Mag, the American Rifleman did some tests awhile back that found about 6,000 psi peak pressure variation with different makes of brass they tested with the same load. Same with swapping primers around.
As mentioned, brass is made from different alloys.
This article on x-ray spectrometry of different brands of brass shows this. So even if the outside dimensions were identical, you could still get a 3% variation in case capacity due to alloy. The densities of the various brasses used are below.
Copper:Zinc / Density at 68°F / Notes
60:40 / 8.39 gm/cc / (aka, Muntz Metal)
70:30 / 8.53 gm/cc / (aka, Cartridge Brass)
80:20 / 8.67 gm/cc / (aka, Low Brass)
It needs to mentioned that the outside dimensions of the case head, the exact extractor groove angle cut and the rim and rim chamfer and extractor groove tolerance can account for some weight variation that doesn't affect the internal capacity. Other than the primer pocket, only the brass that is forward of the head affects capacity in proportion to the brass weight. If the head is light, for example, a case would have less internal capacity than an identical weight case with a wider, thicker rimmed head. You can see the
.308"s head dimension tolerances in the SAAMI drawing. The last time I weighed and measured water overflow capacity of several different headstamp .308 and 7.62 cases, I found weight difference predicted capacity difference ±20%, so the head dimension variation is real. Even the difference in headstamp impression volume has some part of the effect.