One likely cause is variations in the case head to shoulder length, despite the "careful" sizing.
Even with new brass, there will be variations in this "case headspace" length. My experience has been that this is especially prominent in cases with a narrow-angled shoulder (.308 and .223). Other contributing factors are variations in the brass (alloy composition, annealing variances, thickness) and sizing issues. Note that you can vary the amount of shoulder setback quite a bit by using different amounts of case lube for each case. Test this yourself - lube up a few cases with only a small amount of lube, the lube a few with a larger amount of lube, then size them. You will see significant differences in the case head to shoulder length.
You indicated that you have fired the brass (at least) twice. That might explain everything. Variations in the brass (mentioned above) will start to show up, and each piece of brass might have differing amounts of springback after sizing. That will result in differing case head to shoulder length, thereby different OAL's since the Giraud indexes on the shoulder.
If you think that you need an extreme degree of precision in case length, you might try annealing the cases after two firings. Then try sizing samples of annealed and non-annealed cases to see if annealing reduces the variation in OAL.
I noticed that you measured the OAL to the nearest ten thousandth, or at least to the nearest .0005. I don't think that you can measure the case head to shoulder length that precisely, but you should be able to measure to the nearest .001 with a Hornady case headspace gauge. When you measure the case head to shoulder length, is there a variation among the cases? If yes, does that variation correlate with with variation in the OAL of the cases? If it is, then the OAL variation is in fact due to variation in the case head to shoulder length after sizing.
I measured case to shoulder several times and I'm pretty consistent in my technique. No variation in base to shoulder.
If no, then the variation is due to the trimming step with the Giraud trimmer. Are you using the same pressure to insert each case as it bottoms out? Are you holding the cases exactly perpendicular? I have noticed that there is some "play" in the Giraud's spring-loaded shell holder - it is possible to insert the cases at an angle rather than exactly perpendicular.
I feel that my technique with the Giraud is also pretty consistent.
Taking a different tack, why do you need the case OAL measurements to be that precise? As long as your longest case length is less that the corresponding portion of your chamber, it probably doesn't matter that some of your cases are up to .004 shorter than others. Reading between the lines, your use of .308 cases and your interest in extreme precision might mean that you are a Palma or F/TR shooter and need consistent neck tension. Do you think that variations in the length of the case necks might result in differing neck tension for each case? My guess would be that other factors would overwhelm this effect - factors like differing neck wall thickness (unless you turn the necks) or differing "springback" after sizing.
I want everything to be as consistent as possible. Yes I turn necks and yes I feel that neck length will effect neck tension.
Thanks for your input I appreciate it.
Randy