Symokurka is worried about exactly the right things, I think.
I'm offering a bit of what worked for me - so feel free to take or leave it (we all know what free advisement is worth!)
One thing I'm stunned that nobody picked up on, is that I've never zoned commercial/residential *at all* myself, and did fairly well, though some areas favour one use over others. In fact, the huge key here is a sort of flexibility.
micronation Thought Experiment: Confederation of Democratic Lawyers (the CDL!)
Let's say this CDL divides their land up into:
40% commercial
50% residential
10% infrastructure
within themed areas of:
40% Bavarian
40% Roman
20% MontyPython
And say it's all distributed like this:
=================================================
commercial: 20% Bavarian, 10% Roman, 10% MontyPython (sums to 40% of all land)
=================================================
residential: 20% Bavarian, 10% Roman, 10% MontyPython (sums to 40% of all land)
=================================================
infrastructure: 10% Bavarian, 5% Roman, 5% MontyPython (sums to 20% of all land)
=================================================
Okay, so infrastructure you don't have to worry about much; it's not going anywhere.
But let's say Bavarian commercial land is a hit, and MontyPython residential land is really popular too.
And say Roman residential is a total flop (these are just examples).
Well, depending upon how you designed things, you can be either be:
a) doing great, because you could shrink the flops and expand the successes geographically, or
b) be totally screwed, because some areas are bursting at the seams while others sit empty.
I'd seen this sort of thing in Caledon early on, planned for it, and did pretty much okay. The classic example is the forest area. I get people who love to be forest dwellers, and they come, build a community in the forest, and it gets... more foresty. This attracts the like-minded, and they 'grow the forest' into areas that were sorta foresty, but not blatantly urban around the forest. In time some forest people leave, and the forest-edge folk move in to the heart of the forest if they can. The more shop-minded people expand nearby towns into the town/forest buffer zone... all the while, this helps maintain 100% occupancy because Caledon is designed to ebb and flow with spot-market conditions. I get armchair analysis from people all the time, yet nobody seems to pick up on this. It's Business 101 stuff: if widgets move better than gadgets, put more widgets on the diminishing gadget shelfspace.
From the little table above, you can see how if each neat little subtheme is geographically overplanned, or worst-case 'walled in' you will get spots of unfillable low occupancy whenever conditions change. Which is all the time. Say not many want Roman Residential (or whatever) - if it's walled in and stuck as Roman Residential - you'll be burning tier on it. The approach that seems to work is have say a Roman Residential 'core' area that you might hold steady, but let it fade into the other themes geographically per market demand.
Of course, you may have the luxury of scary-crazy-high demand, but... if ever you don't, it's great not to have painted oneself into a corner. Some regions simply "have to be a certain way" - Michel's Al Andalus is one of these, where if you don't have areas A, B, and C done a certain way simply it doesn't work as a concept. But if you can avoid that, you'll have better occupancy long term, I think.
* * * * *
Incidentally, shops and merchants are key - they are a tie-in to the strength of the overall SL economy. Which makes the CDS and Caledon look like specks of dust by comparison. Sure you could do like Dana did and make a beautiful, nearly 100% residential area, but residential areas tend to value quiet and privacy. A whole lot of "leave me alone please" in an environment like that. And probably a lot less laws to argue over - no fun for you guys! 
With shops, you'll get most of SL dropping in and buying stuff that the merchants make. Once it stabilises out, neither the region owner nor the merchant is paying the tier - it's the happy shoppers. This is a wickedly powerful sustaining force - essentially, the grid economy is now floating most of your boat. In fact, some resident merchant-barons easily float their private lands off their shop earnings too. Think of all this as 'gross national product' - and believe me, the merchants can do well. So well, that they could easily float an entire sim or three on their own if they felt like it.
I hope this was constructive!
Des
). I do like Seattle's streetcar neighborhoods -- each one with a small commercial center, schools, and open public space. Not everything need be in a central core.
