I can't concentrate on a startup, other than 2-3 hours a day because I have to work to make a living. In practice, an idea will take me 3 years to implement at this pace which makes it meaningless to begin with. I can't quit my job, as I can't survive without income.
How do people overcome this? I'm pretty sure it's not such a rare case.
You need to do two things.
Consulting at a higher daily/hourly rate to free up more time is how we're doing it. Although building a good consulting business is as hard, if not harder, than a software business we're finding (more touch points, etc.). And finding good clients is not always the easiest either.
Either that or save all the money you can and quit and go for broke (mint.com method).
Here are two more things you could do:
I agree with Gary E., you have no idea what it will take. Get in motion. Set deadlines for yourself, unreasonable ones. Ship something soon and create the habit of shipping regularly.
I can't concentrate on a startup, other than 2-3 hours a day because IWhat about weekend days?
have to work to make a living.
In practice, an idea will take me 3 years to implement at this paceI disagree. Everything depends on your demands.
which makes it meaningless to begin with.
I myself don't think a 3 year development time is unreasonable if you wind up with a profitable venture that returns money to you for years beyond that. People will do a a 4 year bachelors, a 2 year Masters, a 5-15 year Ph.D., etc., but 3 years to single-handedly create a profitable software venture is "meaningless"?
(Along these lines, as Anthony Robbins wrote, "People will greatly overestimate what they can accomplish in one year, but they will greatly underestimate what they can accomplish in 10 years." Or something close to that).
That said, if launch date is critical based on things like trends or competition, then either you hire out, quit your job, bag your project, get a small team of co-founders who will pitch in for sweat equity, change your idea, or hope your assessment of that launch date criticality is incorrect. Or some mixture of all of these.