because most startups dont know anything, they have an idea that was shaped from their experience and assumptions about how the market will react to it.
Once you appreciate that these things arent much more than educated guesses, you can be a lot more focussed actually determining how people react to your product, as opposed to going off and coding what noone wants.
My advice (well, others said this before, but my experience confirmed it) is to build your first product the way you like it. If you are happy with the product and you really enjoy using it then you already have one customer :). And I am ready to bet others will come.
On the other hand trying to fulfill every customer wish and implement all features users suggest is not every time a good idea. Usually it leads to an over complex product that is not well adapted to "young markets" of "untrained" customers. Is like trying to sell them a dvr enabled TV set when they just learn how to change channels.
Once you appreciate that these things arent much more than educated guesses, you can be a lot more focussed actually determining how people react to your product, as opposed to going off and coding what noone wants.