It's far safer to make a product for a market that you know exists, rather than build a product and try to find a market. You differentiate yourself by providing more value than your competitors, whether it's a novel feature, better service, etc. There is obviously an element of risk/reward - if you go for a potentially unknown market, e.g. you have a cool idea and think people will buy it, then you could make a much better return as the only player. Or you find out that there's no market and you crash and burn.