For me, the plan has always been about immediate use and budget. Problems that I need to solve now, if I can afford to.
I know a lot of people work a room at a time, and that makes a lot of sense. In my case, though, it’s just about the actual problem being solved. If having one voice operated fixture in a room that has three gives me enough automation to solve the “Voice control of lights” need for that room, I often just do the one.
I also laid out my phase 1 budget in advance ($1500 for stuff that I would be willing to completely replace two years after I started), with a pretty good idea of what my phase 2 budget was going to be as well. And that has shaped a lot of my decisions.
I detailed this whole process in my project report, so I won’t spend anymore time on it here. 
Oh, and I agree with Mike – – I didn’t have a good idea of what would become my eventual top priorities. Voice control barely existed when I started, and now it’s a big deal. The smart watch changed a lot of things for me as well. I had a detection need I thought would be solved by geopresence and turned out not to be, but I beacons became a good substitute. So there’s also been a lot of learning as I go along.
One more thing – – one year in I completely changed my priority list and put reliability at the absolute top. I found I was willing to give up features to get that, in part because it was costing me a lot of money to have someone else do all the fiddly maintenance.
I’ve seen other people go completely the other way. They start out saying that reliability is one of their top priorities, but then they find that versatility is important enough that they’re willing to do a lot of handholding on the system and put up with occasional glitches.
So I think one of the other things that we learn once we actually have home automation installed is what our own tolerance for glitches is. 