There is a plethora of books and talks about starting companies; Sam Altman (of YCombinator) offers the BEST around. Seriously worth your listening. Below are my 11 favorite lessons from his first lecture.
Every idea should sound totally crazy but actually be defensible with a short sentence.
If you are not an actual user of the product that you should surround yourself with the customer as much as possible
Initially it should look like your market is extremely small and that you pursue a monopoly of it. The key thing is that there is a huge market for your service and 10 years from now. Great start ups take 10 years.
You must begin with a mission. If you have several ideas that you're working on, should pick the one that you think about most when you try to not think about work.
The purpose of a start up to solve a problem that you're really passionate about.
What's more important than figuring out how to come up with ideas his building relationships with possible cofounders.
It is a great thing when other people don't like your idea. It is a sign that they are not likely to do it.
Better to build something that a few users LOVE than that a lot of users like.
Way easier to go from small number of customers who love it… to lots who love it… than from lots who like it to lots who love it.
Find a small set of customers: understand them extremely well… serve them extremely well.
Companies build what the CEO measures.
Check out his entire lecture series here.