Seeing this question made me think of an interesting one: all the answers are based on the assumption that there are users. But what if there are no users at all — how should tasks be prioritized? Should you continue developing new features or improving existing ones to attract users, or should you start a new project instead?