Mix bus compressor, and usually a program EQ, get inserted right off the bat, but in bypass until I get rolling along.
Interestingly, I have found a a couple "poor man's" SSL G384s that I'm using lately: One is the bus compressor on the Audient Sumo, the other is the dbx 162SL (that new purple one), both of which sound, or can be made to sound identical to the famed SSL.
I have kind of given up on multiband compression. Turns out I LIKE the pumping or ducking artifact when the kick-drum taps the gain-reduction down. I just like that brick-wally smack. I'm a rock, roots, blues, country, surf engineer, so that sound works much of the time.
Anyhoo, I limit the kick and snare in mix with dbx 160x, overheads with a stereo opto of some sort (which is often done going to tape and doesn't need any more in mix), bass with a distressor, vocals with any number of classy mono compressors, my favorite being an LA-3A, lead guitars also like the LA-3A with API 560 graphics, BGV with an Aphex Compellor.
So...pretty standard modus operandi circa 1990. I don't do drum compression sub-groups, which seems to be the thing lately, but, it's not something I ever did.
On the mix bus I get the comp to pop when the kick and snare hit, just going down a couple db on the meter, max, with a 2:1, 3:1 OR 4:1 RATIO.
My bus EQ is quite often an API 550B.
NOW HERE'S A QUESTION FOR DEBATE: ON MIX BUS, DO YOU COMPRESS BEFORE EQ OR AFTER?
I'm not saying which way I go.