Air layering timing preference

[Teacup bonsai]

As I understand it is typically faster to keep the section to be layered attached to the main tree until it is thickened as it can make use of a larger root and foliage system.

My question for nuts who have turned air layers into long term projects.

Do you feel like you have been more satisfied with the tree in the end if layered earlier (branch is relatively thinner than ultimate goal) or later (when the branch is close to your desired ultimate thickness for the layered tree)?

 

[Empty Mountain]

I see air layering as a way to compress the time it takes to develop a bonsai so I would usually only air layer a piece that already has some value (thickness/movement etc.)

If a branch isn't thick enough I would just move on and find some better layering material. Waiting for a branch to thicken up on a tree doesn't seem like a great plan as you have less control over a full sized tree in the ground, maybe it won't put resources into the branch you want or might just drop that branch altogether. If there aren't parts of the tree that are worthwhile to air layer off but it's a desirable species/unique cultivar I would just start from small cuttings.

 

[Wires_Guy_wires]

I have trouble getting juniper cuttings to take, mostly because wood lice get in the soil and eat all the emerging roots before they can turn woody. It's an issue I've been struggling with for the past decade.
So I have to resort to air layers most of the times.

I have a couple branches that I'm developing as bonsai that are later to be air layered off. They need to thicken a little bit first, so I keep developing them until they're ready.
For other trees, like my arakawa maple, I prefer doing air layers on whatever thickness can hold one. Simply because I want them on their own roots and to develop a good nebari, it's better to start them young. Maples can grow quite fast if they're given the space, so I'm not as worried about thickness or size as I would be with slower growing plants.

I wish I layered my juniperus rigida though, because I didn't and it died over the summer. I'm now stuck with twenty tiny cuttings that will need a decade of development. However, this type is difficult to find and I think it will pay back the effort by about 200%. But still.. A decade..

Thin trunks and branches are easier to wire, so it can be good to start smaller if you want to do more extreme bends. The keyword is easy.
Thick trunks and branches require less development so it can be good to start larger if all the desired features are there. The keyword is uncertainty; older branches might not layer as easy as younger ones and cutting a ring of bark might send it to the afterlife.

All the roads lead to Rome here, and it's a matter of situation and desired outcome.