In software development we have a term for spinning your wheels during the initial tentative phases of a project; Analysis paralysis occurs not because of lack of direction, the project is stuck because there are too many good ideas. Everyone has a solution to the problem and without a true strategic direction every one of these solutions can create a delay in the project outcome. Each avenue and opinion creates a path or possibility to consider before moving forward. The biggest problem with analysis paralysis is there is a lot of activity without much progress.
Traditional software development goes in a straight line with one phase leading into the next; Planning, Analysis, Design, Testing, Implementation. Also know as waterfall method this form of managing complex projects has all but been replaced by more agile methodologies.
Rapid development, early release with constant iteration is very successfully used by technology leaders such as Google.
It means putting out beta releases for review and consumer testing before the full product is fully materialized. For multi-faceted applications that require consensus from large and diverse user communities this approach allows for adjustments to be made frequently to satisfy more and more users.
Design, develop, deploy, and repeat. This is the way to move forward on large projects. http://www.firelily.com/opinions/cycle.html
Complex multi-year initiatives must be broken up into smaller projects each with manageable deliverables. Parsing large tasks into smaller manageable tasks makes it possible to advance. Small teams with decisive leaders make project results obtainable.
Breaking analysis paralysis can be done my taking on a simple part of the project and completing it, test it with real users. Review the results, gather observations and feedback and then move forward. Don`t waste time trying to document or design every possible contingency, take the project apart. Take each piece and produce it to the best of your capability now. Then review and revise.