Your first RCM analysis without a reliability department
The usual reason an RCM program dies is not the standard. It is the first analysis: the wrong system, a room with no operators, or a scope that cannot be finished. If you do not have a reliability department, you have to be more ruthless about all three.
Pick a bounded, visible system
Choose something with a clear function, a manageable boundary, and enough pain that people will show up — a compressor package, a boiler feed pump train, a critical conveyor, not “utilities” or “the whole packaging hall.” Write the boundary down. The worked BFP-2 example shows one such boundary on a representative system.
Staff the room honestly
You need at least one operator, one craftsperson who works on the asset, and someone who can decide what “good enough” means for the plant. A facilitator who can keep the seven questions in order helps; that can be an internal person with a kit, or an outside facilitator. Do not start without the people who know how the machine actually fails.
Finish one task list
Time-box the analysis. Publish a task list and a one-page brief, even if a few modes are still marked “evidence thin.” A finished first analysis teaches the organization more than a perfect unfinished one. Then decide whether the next system is worth the same treatment, or whether a PM optimization pass is the better use of the next month.