Since I have started working on my stealthy project I have been generating business ideas at the rate of at least one per day. Everyday I discover a need and I jot that down in my org-mode TODO list. I have almost transformed from a programmer to a business analyst of sorts. Some of the ideas can be merged into my start up but others are organizational in nature and I can see a good market for them.
All the research into my potential clients' industries has given me allot of insight into their business to that point where I can imagine myself competing with them and winning with my technological edge. It's a good habit to look at a corporate website and ask yourself "how do they make their money"? It's an effortless subconscious activity for me now, like a resident system debugger, lurking in the background and observant of every last little detail.
Once you go past the "app" stage and actually dig into the business models of most B2B enterprises, there is nothing the great majority offer other than a telephone and a friendly voice. People are making fortunes out of relationship management and contacts; really, and a whole lot of them are resellers, integrators and what not. I have seen B2B shops that could be run with nothing but email, spreadsheet and sourceable domain knowledge.
You want to look into the "process improvement" angle of things. Just as mainstream software developers consume design and development methodologies like crack, the business community, too, is craving "processes". Anything you can do to give them a methodology with an associated software solution will be greatly appreciated. You might need to roll up your sleeves and become a coach of sorts, but you will also be able to create software (if that's what you really want, I know I do.)
If you come from a mathematical or hacking background the "complexity" of your average business will seem trivial. The main obstacles might be politics, inertia, and where applicable legislation (following weird laws in certain industries) but you can think of a few ways to brand your solution and target a small niche of early adopters, first.
For me the problem with mini little side-projects is "face conservation"; I don't want to commit my identity to one business and be known you as "the guy who does X". One idea I jotted down is recruiting corporate faces, a few MBAs on contract basis, and launching little B2B start ups backed with solid business intelligence, research, marketing and hype.
Even if your current startup never takes off the ground, think a little harder about "process improvement", ask yourself "what can I do to make my tasks easier", imagine you need to communicate your current actionable item to one, two, three, several people, etc. I can almost assure you your next successful project is lurking somewhere in whatever you're working on now. For me, the best thing I have learned was to break away from the "application" mindset; heavy emphasis on software (desktop vs web, modules, components, packages, threads, sockets, layers, javascript, reimplementing fun-thing-foo in boring-but-ubiquitous-bar, etc.). I was looking at the software and not seeing the solution.
Avoid solution fetishism. I spent many years solving, or at least trying, very hard computationally sexy problems. My first reaction to any problem is to start writing it out as Lisp comments (remember that pg etherpad replay?) and then flesh that out into some sort of "code". I have a whole directory full of tiny little Common Lisp files that I used to sketch some small budgeting thing, a short essay, maybe a floor plan. Guess what happens when you run to sketch software? you WILL write software, and that's not always a good thing. If you can solve a problem with a small change in habit, maybe better organization, then don't bother writing code.
A little effort with these guidelines and you too will be driven insane with daily influx of ideas.
Cheers! and oh, this is my first contribution here :-)
There, fixed that for you.
Giving serious thought to one idea per day seems somewhere on the line between hardly possible and a huge waste of time.
I can go to the local founders meetups and pick you out a couple dozen people that generate loads of ideas. The ones you'll want to meet though are the ones that have seen a couple of them through by getting behind them and pushing.