1) What systems, if down for an extended period of time, would cause embarrassment to the company, prevent the company from doing business (making stuff/providing services), or would prevent the company from collecting or paying out monies. You don't have to do a bona fide risk analysis, though that's best practice.
List your top 10 apps by importance. Get business owners to tell you what their recovery requirements are: 0-48 hours after incident? Could they coast for a week until the data center is rebuilt? Your highest-priority apps should drive the plan.
2) What is your assumption? That there's data loss but you can still get into the data center? Or is the building off-limits due to fire, flood, or anthrax? You must have a clear idea of what you're planning for.
3) Based on that assumption, what will it take to recover those apps? Hot-and-spinning, replicated database colocated at a hot site in some far-away city or in a rent-a-data-center down the block? Or can you get by on your recently-made backup tapes that you store off-site, rebuild your infrastructure in somebody's basement, and restore functionality. Put another way, how much infrastructure must you rebuild in order to get the bare minimum required systems up and running? How will you get the machines and build the machines? There's your outline. That'll be fi'ty bux! :]
