There's no right or wrong want to do it Charles, as long as whatever you come up with, works. I'll give you some of the ways I do things.
Instrument busses: Anything that has multiple instances, always goes to an instrument bus for me. This way, once I get the individual tracks sounding the way I want them to, I can send them to a bus and have control with one fader. Examples of this...
Drums: Say I have 30 tracks of drums using hyrbid kits or whatever. Once I get them sounding good, I'll send them to a main drum bus. From there though, I may make little sub busses as well. So say we have 6 toms. I may send them to a tom bus and that tom bus will be send to my drum bus which is sent to my master bus. Everything in my mix gets sent to the master bus in some form. If I create a parallel compression thing for my drums, sometimes I create a bus for it, other times I can use the Sonar Pro Channel Bus comp right in the drum bus because it gives me a wet/dry. So on that drum bus, I may have a Fairchild compressor in the bin, and in Pro Channel, I may use the Bus comp for the actual parallel compression. So, all 30 drum tracks are sent to the bus. If there is a tom or cymbal bus created, the toms and cymbals will be sent to their own bus and that bus will be sent to the drum bus. From there, the drum bus is sent to the master bus.
Guitars: I like to layer my guitars. There are at least 4 at all times for dirty, driven chorus sections. These will all be processed and set up as they are using the tracks they were recorded on...but after I get them where I want them, I will create a guitar bus and all 4 get sent to that bus which is then sent to my master bus. This way, I can control the "guitar bus fader" and control the volume. I can even put a light compressor in this bus to compress the guitars as an entity.
Bass: I normally do not send my bass to a bus even if I use a DI and another form of bass input. 2 tracks of something normally doesn't get bussed in my realm, but it depends on the situation.
Back-up vocals: I do so many back-up vocals, they need to be sent to a bus. Same process as the guitars and drums. Get them where I want them in the project and when I'm happy, I send them to a "bax bus" and then that bus is sent to the master. I now have full control over the whole lot of them with one fader and can also tweak them as an entity.
Effects send busses: These are good for effects and usually give you a much cleaner sound instead of dropping an effect on your effects bin. Reverbs, chorus, delay, phase, flange, and any special effects usually benefit from having busses created that will be routed to channels. In our "instrument bus" situation, the channels are being routed to the busses. In THIS situation, our busses are being routed to our channels to deliver effects to those channels. Once the bus is created and the effects is in place, you just go to the channel of your choice in your project and insert the send of that bus.
The good thing about this also....you can run one reverb in an effects bus and send it to all tracks without using multiple instances. This was really good for people that used pc's that didn't have much power and is still used in a few instances today even though we have the power now.
For most situations in a serious recording, you normally do NOT want to send one reverb all over your project. We have enough power to create effects busses with different reverbs on each instrument that may call for a verb. For examples, it's not really good practice to put a gated reverb that you'd use on a snare, on your vocals and guitars. I mean you CAN do that if you want to and feel the song calls for that, but in my opinion it's not something that should be done. A vocal verb needs to be selected for the voice. It shouldn't just be a generic verb you'd use on everything, know what I mean? A gated verb on a guitar track would make the guitar appear to have a tail as well as a cut-off point. Not a good move unless there is a section that calls for something like that.
A good rule of thumb when using this method of creating effects busses, is when you insert the effect into the bus, to run the effect itself 100% wet with no dry signal. This does 2 good things for you.
1. It stops the track from gaining volume when you turn up the effects send in your track.
2. You already have your dry signal in the track itself...we want effects coloration here, so as soon as you start to turn up the effects send on the channel, you will hear effect. If you left "dry" signal in the effect plug itself, as you turned up the effects send on the track, you'd have to turn it up way more to hear effect and you'd also get a volume boost because it would also be counting the dry signal in there. So you'd get more of a boost in volume than you would actually hearing the effect present.
From there, your effects busses should be sent to your master bus unless you feel the need to route them somewhere else. If that "somewhere else" exists, it should be routed to the master bus as the end of the chain. Hope this helps....best of luck.
-Danny