Launch quick and learn
It's now a well established principle in software development that an iterative approach works best for the majority of projects. Unless you're running life support systems or a nuclear power station, the best approach is to launch with a small but coherent set of functionality, learn from your users, and then launch another small update. That way you continually incoporate feedback from your users in your project and can adapt it to changing requirements as you go along.
It still surprises me how many people try and build community solutions and don't take an iterative approach. They have a vision of what their community want and they can often be so focused on that vision that they end up with something that is: - too big. If you have too many places to interact you'll dilute your community, often to a point where they feel no one else is there. Nobody wants to be first at a party. - too complicated. KISS is a good principle in many areas of life and community is no exception. A clever community solution will be beaten by a clear simple one every time. - too expensive. If you spend a ton of development effort on building your vision it's going to make the solution more expensive. Proving the value of community can be one of the big challenges you face, so why make it even harder with a bigger upfront cost than necessary? - too late. Building something big, complicated and expensive takes time that you may not have. What your community wants may have moved on considerably from what you originally thought if it takes you six or more months to build it. I always try to encourage customers to keep their initial implementation of community functionality small, focused and using out of the box functionality where possible. Those projects that have done this and followed an iterative approach have tended to be the most successful.
