I only pulled a few for concentricity over .002" but I pulled a pile of them for neck thickness variances by more than .0015". Just looking at my piles of brass, I'm guessing that I am keeping 1/3 of them and 2/3 are getting tossed or repurposed. I think I might run these through my sizing dies just to see what happens.
Neither concentricity nor runout can be credibly compared until brass is fire formed, but thickness variance does matter.
It's the first thing I check with all new brass, regardless of brand.
That you might discard 2/3 of a batch still leaves you with 1/3 that are good. As long as this leaves you enough it's fine,, just throw the rest away and save yourself the troubles with it.
Running it through a sizing die will not fix it in the long term, and will eventually amplify the issue.
Thickness variance at necks runs the full length of cases. So with any body sizing cycles the case will take a banana form and runout grows. Eventually the runout exceeds chamber clearance, and you get thrown shots.
If you neck sized only, then there is no action to cause bananas, and cases with thickness variance wouldn't matter so much. But most people cannot get away with neck sizing only (It takes a special plan).
There can be further culling, as cases have there own character.
After 3 fireformings with no body sizing, I compare H20 capacities, typically culling 1/4 away there.
And for any shot thrown, ever, I toss the offending case without hesitation.
With a WSSM project, Win Reloading brass, and my process, I managed to settle with 80 cases from 1,000.
I was and am satisfied with this as I only need 50 cases.
2.5 barrels later, I'm still using these 50, so it's not really a bad deal for me.
Someone mentioned brand blindness. I couldn't agree more.
With thickness variance, I don't see it resolved by brand.
The best, and the very worst brass I've measured, were different lots of brown box Lapua 6BR.