I have done workup where I started at SAAMI and moved all the way into the lands. I have done workups where I started at the lands and worked back to SAAMI.
What I do these days is call the bullet maker (if I have never worked with that bullet before), ask them, and work a little in front of, and behind what they say.
For instance: Hornady recommends their ELD -M to be 0.02" off the lands. I spoke with the tech for about 30 min on the phone....and I am from Nebraska , been to their plant many times to buy cheap plinking factory seconds and talked with them.
Anyway, I start at 0.020, I move back to 0.030ish in 0.003 increments and up to 0.010. So far in .308, .270, .300wm, and .223 their recommendation has worked very well.
I had to switch from 200 gr accubond in my .300wm to 200 gr partitions because the AB was too long for my Model 70 magazine and the best accuracy required me to single shot it. The partition is almost 0.15" shorter so I was able to move the bearing surface closer to the rifling. ...and it's a hunting rifle so it needs to feed from the mag.
However, with sierra match kings I have found seating depth not to help/hurt much (within reason).
Some bullets seem to have a jump that works best at all powder charges. Some like the match king really doesn't. For this reason, I tend to choose tangent ogive bullets because I load for too many rifles and just don't want to wear a barrel out with bi-variate pairwise testing.
So I subscribe to the Chris long/uncle Nick barrel time theory, and use tangent ogive bullets.
I may lose some BC against the VLD's, but I don't have to chase lands as much, and load development is much faster.