If it's your first start-up, have a business model so you can bootstrap - raising any significant money will be difficult without a reputation. [Though being in YC/TechStars will get reputation/connections quickly I guess.]
If you are doing it in your spare time, plan on it taking all of your spare time.
Marketing is difficult, free publicity (from blogs) is surprisingly difficult to get and when you do fewer people than you think sign up / click though.
Go to as many networking events / conferences as possible meet lots of people and be sure to meet key people like press, bloggers and other companies
Spend some time working out what the smallest, least feature rich version of your app is, with only the features in it would make the product unusable without them. I often find frill features only get 5-10% take up from users, so don't put them in the critical path to releasing the first version - put them in later (or wait for users to demand them).
Do something that is night and day better than the existing solutions currently out there, because if it's only slightly better it's very hard to sell it to people and by the time you've written yours the others will have got better.
Read HN [you got that one already], watch Jason Calacanis's http://thisweekinstartups.com/ every week [on later today], read all of Paul Graham's essays
All feedback is valuable, don't get offended and reply quickly to encourage more feedback.
If you are doing it in your spare time, plan on it taking all of your spare time.
Marketing is difficult, free publicity (from blogs) is surprisingly difficult to get and when you do fewer people than you think sign up / click though.
Go to as many networking events / conferences as possible meet lots of people and be sure to meet key people like press, bloggers and other companies
Spend some time working out what the smallest, least feature rich version of your app is, with only the features in it would make the product unusable without them. I often find frill features only get 5-10% take up from users, so don't put them in the critical path to releasing the first version - put them in later (or wait for users to demand them).
Do something that is night and day better than the existing solutions currently out there, because if it's only slightly better it's very hard to sell it to people and by the time you've written yours the others will have got better.
Read HN [you got that one already], watch Jason Calacanis's http://thisweekinstartups.com/ every week [on later today], read all of Paul Graham's essays
All feedback is valuable, don't get offended and reply quickly to encourage more feedback.