The “When” of ScopeWhen do you have to know scope? At the end of a project, scope is very well known, inasmuch as it is the delivered system. That, of course, is a little late. But you certainly do know more and more as you proceed through the life cycle. So to make the question more specific, how much of the above do you need before proceeding to actually architect your solution? As already noted, Scope and Solution Concept are inextricably bound. A project needs to iterate between these, with all necessary stakeholders at the table, until an explicitly stated balance of expected value, feasibility and risk is achieved, before proceeding further. This means that how much scope definition is necessary at the concept stage is not black and white. It is a matter of how much risk you are willing to assume, including the risk of not knowing what all your risks are. You owe to yourself to at least confront this head on, rather than stumble over it later and take a nasty fall. At a minimum, assess your project against all these factors and decide for yourself. Do you know the objectives of your project? How well and how much of a risk is that? Do you know its boundaries? And so on. Otherwise you are likely, among other unpleasant outcomes, to be defining architecture for elements you will never build, and gathering and modeling more explicit requirements than you have any hope of satisfying, a famous sore spot with business stakeholders.