I haven't installed any because we live in a dark hole and the amount sun we get (especially in the winter) really did not pencil out.
That's basically the bottom line, you need to look at the specific location and see what the expected yield is and then look at what local incentives there are and do the math. There's no one size fits all. There are some tools floating around on the web that aren't bad for helping to figure this out (although most are at a rather coarse granularity unless you pay $$'s, the more reputable installers will run the payware version for you). Google Earth actually has a daylight simulator that takes into account the terrain that you can play back for days in the past that's actually pretty close - but doesn't deal with cloud cover which can be a huge factor. As to cloud cover, high clouds you still get some power (decreased) but low clouds/fog shut you completely down (a buddy up in WA has solar and his winter yield is close to zero a lot of time due to fog) so its hard to really get a solid number.
There are several different types of panels out there. All of the older ones would shut down the entire panel if a single cell was occluded which is not good if you have intermittent shade (birds, tree shadows, leaves, dust, blob-o-snow, etc..). Some/many (haven't looked real recently but they were getting more common ~2 years ago) newer panels continue to deliver full power for the non-occluded cells and generally yield a lot better. Anyway, look closely at the proposed panel technology and see what the performance characteristics are because they're not all the same.
On the financial side the price/(panel/watt) has been on a pretty continual drop and I'm betting that will continue for the next 5-10 years (based on known but as-yet-unreleased innovations in manufacturing and design). This means that you basically need to figure in the time value of money versus the expected payback and the ongoing decrease in costs in your calculations. With interest rates where they are, the cost of the panels is decreasing faster than the value of stored money appreciates (confused yet, I sure am
). Mostly its pretty simple if you just BUY the panels and the system, where it gets weird is if you do one of the panel rental deals. A co-worker did one of those through solar-city (who did do a nice job on the install and were reportedly easy to work with), basically the way it works is he pays his residual electric bill to them and they pay the power company. Some of us did the math and figured that if he sold the house at ~5~ years in he was actually at a net liability (that is it would be cheaper to have bought the panels new at 5 years than to continue paying back the rental cost), now the odds of most purchasers figuring that out is pretty low.. but it was an interesting exercise (this was of course at the relatively low local power rates and his specific utilization). My friends in HI are seeing something like a 2-3 year payback because of the crazy power rates there, you're a bit higher than us (~12c/kwh vs ~7c here) but no where near there (40+c/kw).
edit:
Oh yeah and I hope your roof is in good shape (that is has an expected lifespan past the panels) because now you have a bunch of stuff on it that needs an electrician to remove if you need to re-do the roof. I haven't seen any good number on the impact on roof lifetime but expect there has to be some (you've penetrated the roof seal - that's bad, you have less exposed roof because the panels act as a second roof on top - that's good). No idea, maybe its a wash?