Before you build anything, find your customers and talk directly with them.
After talking with a bunch of customers, you'll know if there's a market for your product, how much you can charge, your USP, and where to market your product. This approach works much better than the "Field of Dreams" method.
I'd say, know who your first customer is before you even begin. I had the greatest success when my first customer had told me that they were buying and to just hurry up and finish already.
for me business rule no 1: buy low (products, knowledge, skills etc) then sell high in order to do that you might need to add value (packaging, story, place etc) in between first two steps.
The most important thing I can think of (especially in the early days) is this:
Some days, you'll wake up and feel like you're going to conquer the world. Some days, you'll wake up and think, "I must be f***ing crazy for even considering this".
On the former, you want to start work on that big piece of the software (or project) that you've been avoiding for a couple of days/weeks.
On the latter, write down all your ideas. These days, you're thinking about ways your service falls short, and, as a result, what you can do to add value to it. It's best to work on grinding out menial things on these days, and work for an hour or two longer than you normally would.
Also, if you don't wake up scared some days, you're not in the right frame of mind.