For me VCC is a lot like bus compression and it's just not the same added at the end. I don't think you can start the mix with SSL model then switch to Neve/RC etc near the end and expect it to work any other way but worse.
Agreed, although I tend to work a little differently from you perhaps.
I usually set up some basic eq and comp on each track first, like painting with broad strokes, then set up my groups and master.
Then I add VCC, usually at full drive, and tweak the gain to get a high or overloading signal depending on how it sounds, first an instance on each individual track, then to the groups, then to the master.
Doing it this way I tend to switch between the different saturation flavours on a per channel basis to find what works. This is where what I said on the OP comes into play, because I'd always like to use the SSL (because it's SSL), but invariably end up on Neve based on spot impression. Then I'll tweak.
I've been surprised on other threads I've read when people have said 'is VCC worth it - I don't hear a difference' because to me the difference is enormous and immediate, especially when adding it to something you've already comped.
Obviously if you were mixing on hardware this wouldn't be such a conscious part of the process because the tone shaping and artifacts are already there, but then I imagine SSL stuff works better than this in real life