"Knowing what to build" is fine for technical people who are building a simple web based tool which they'll use themselves. However, this doesn't generalise to

"Knowing what to build" is fine for technical people who are building a simple web based tool which they'll use themselves. However, this doesn't generalise to building software for complex industries. As an example, a few years ago I was advising a company who wanted to automate the calculation of tax. You would have thought that this was easy, but in fact it's the most complex thing ever and you need that non-technical specialist domain knowledge. (You know you're in for a fun ride when you ask a question and the answer is "we're not sure, but we've booked flights to Italy to go and ask the authorities there".) I think I have missed out a load of qualifiers in my previous comments. Here they are: * The skills required for a startup will probably be spread out over many people. * There's no hard and fast rule about which skills anyone will have. * Technical people might be good at sales. * "Non-technical" people might not be good at sales. * "Knowing what to build" and "able to sell" are skills which may or may not be found in the same person. * Knowing what to build might be a skill of any of your founders have, depending on their experience in the market. There's probably a few more. I suppose this is the trouble with trying to write concise generally true comments, you forget to cover all the edge cases.