A lot of it is that the people build things they know will work (rather than things they think will work) tend to succeed more often. (At least in my own admittedly anecdotal experience) The hard work of research, initial sales and market validation is later retroactively glossed over as "vision" by onlookers.