I wished I could say that about all of them though, first success is definitely not a guarantee for a repeat performance, there are always new mistakes to make and some of those have the power to kill you.
Anecdotally, so, not my own start-ups but friends that have tried to start out, if they kept it simple (consultancy, contract work) they were mostly successful, if they went for the pot of gold at the end of the rainbow they mostly failed.
Business-to-business 100% success rate, business to consumer 1 in 5 or worse.
Well hopefully that holds true for me, I was planning on taking the simple route. The simple route seems to be a little more sustainable: if you can find customers to do consultancy or contract work for, perhaps those can be your initial customers for your start-up.
if you didn't lose anybody else's money, it's not a loss. experimenting is always a win.
To which I'd add: Solve a business problem. Charge businesses money for it.
The friends I've seen fail with 'startups' had no true value built in, and no capturing that value (e.g. not charging for it).
Or they were trying to sell vague services to a vague audience, who they didn't understand. (Namely: consumers, who pay for very little on the web, without understanding & compensating for that fact with what they chose to build.)