In my view, there are two kinds of projects: front—end and back—end. Front—end projects are more user experience—related and lighter on the technology. Change.org is an example of this: it’s not the tech that matters; it’s iterating on the feature set and user dynamics enough to understand what will really take off and what won’t. For this kind of effort, build a prototype quickly (one month), instrument it such that you can gather data on what users do and don’t do, launch, get as many users as you can, and then try different things and see what moves the needle on user onboarding, adoption, retention, virality. Based on the data, tweak it (perhaps for a number of years) until you get the hockey stick curve you are looking for.