It's hugely about luck. We can look back on any success story and identify what made it successful (sometimes) but it only has a little predictive power. Success in startupping ultimately comes from either trying a lot of things (amplifying your luck until it approaches 100%) or survivorship bias when you get lucky on your first try and then write about how smart you are.
It helps to have an idea of what might succeed, by studying things that succeeded before and the present business environment, but that increases each attempt's success chance to, like, 2% rather than 0.2%.
(There's nothing wrong with getting lucky, we just probably shouldn't plan around it being the normal case. It has extreme variance by definition.)
