You won't learn much unless you actually start over.
By this I mean that you have to do seating testing independent of your powder node, so that you're not ALSO seeing a collapse of powder node(messing with seating results). I would drop powder by a full grain for the seating testing.
Might be a good time to do primer swapping as well.
While so far away from a powder node your grouping will be worse. That's great, the worse the better.
With this, you'll more easily see at least improvements -from seating alone.
Then with best primer, best coarse seating, and fully fire-formed to stable cases, increment back up to a new best powder load. If load density isn't good with best powder load, note muzzle velocity, adjust neck tension* and run powder testing again until your best load also fills the case (but not compressed).
Then fine tweak seating in a narrow window for tightest group shaping.
*With partial sized necks, tension is not interference fit. It's sized spring back(~1thou max) grip, against seated bullet bearing. So you adjust this tension through sizing LENGTH. Force X Area
If FL sizing necks then forget any tension adjustment as an option. It's already extreme and varying a great deal due to donut and base-bearing junction binding. You could try bigger/smaller bushings, but the first adjustment of consistent value would be to stop FL sizing necks.