The analysis was rigorous. The workshops ran for months. The report was signed off. Two years later the PM programme looks exactly like it did before. Here's where RCM studies actually die.
The study cost £180K.
Eleven workshops. Your best reliability engineer, two maintenance supervisors, an operations rep, and a facilitator who genuinely knew what they were doing. Functions, functional failures, failure modes, effects. A decision tree applied to every one. Tasks selected with justification. Intervals set from real failure data, not from the OEM manual's most conservative guess.
It was good work. Nobody disputes that.
The final report is a PDF. It lives on SharePoint. There is a folder next to it with 40-odd spreadsheets, one per equipment class, each with a tab that says TASKS_FINAL.
Now open Maximo. Find the PM records for those assets.
They're the same ones that were there before the study started.
The Part Nobody Budgets For
Every RCM methodology tells you how to run the analysis. JA1011 tells you what questions a process must answer to be called RCM at all. The training courses, the software, the consultants — they all concentrate on getting from "this pump exists" to "this pump needs a vibration check every 90 days, because of this failure mode, justified this way."
That's the hard intellectual work, and the industry is genuinely good at it.
Then the study ends.
And the output — hundreds of tasks, each with an interval, a justification, a failure mode it traces back to, a criticality that drove the decision — has to become records in a maintenance system. Job plans. PM schedules. Task lists with the right frequencies against the right functional locations.
Nobody budgets for that. It isn't in the statement of work. The consultant's deliverable was the analysis, and the analysis is delivered. What happens next is "internal implementation," which means someone in your team, alongside their day job, is going to retype it.
What Retyping Actually Looks Like
The spreadsheet has 1,200 rows. Each one is a task with an interval.
Maximo needs job plans, with tasks under them, linked to PMs, linked to assets or locations, with frequencies expressed in the right units, against asset numbers that match exactly.
So somebody starts mapping. And immediately hits the questions the spreadsheet can't answer:
The asset IDs don't match. The study used equipment tags from the P&IDs. The CMMS uses functional locations from the original build handover. Nobody maintained a crosswalk. Roughly 300 of the 1,200 rows are ambiguous.
Two failure modes produced two tasks with different intervals on the same component. In the study, both are correct — different failure modes, different detection methods. In the CMMS, someone has to decide whether that's one PM or two, and if one, which interval wins.
The task description is a sentence, not a procedure. "Inspect coupling for wear" is a valid RCM output. It is not a job plan. Somebody now has to write the actual work instruction, and that person wasn't in the workshop.
The justification doesn't travel. The spreadsheet's WHY column is 200 characters explaining that this task addresses a specific failure mode with a specific consequence. There is nowhere in the CMMS to put it. So it's dropped.
That last one matters more than it looks.
The Task With No Memory
Three years later, a maintenance planner is under pressure to cut PM hours. They pull the schedule and look for candidates.
They find a quarterly inspection on a gearbox. Forty minutes. No findings recorded in two years.
Should it stay?
The RCM study answered this. That task exists because of a specific failure mode — bearing degradation with a slow onset and a hidden failure characteristic — where the consequence of not catching it was a four-day outage. The interval was set from the P-F interval, not from convenience. The two years of no findings is the task working.
But none of that is in the CMMS. There's a PM record with a frequency and a description.
So the planner makes the only decision the available information supports. They extend it to annual, or they delete it.
Nobody is being careless. The reasoning was destroyed at the point of data entry, three years earlier, when a WHY column had nowhere to go. A task with no recorded basis is indistinguishable from a task nobody thought about. It will eventually be optimised away, and the study that justified it will not be consulted, because nobody remembers it exists.
This is how a £180K analysis becomes a PDF on SharePoint.
The Three Ways It Dies
Across the projects we see, the same three endings come up:
1. It never lands. The study finishes, the champion moves role or leaves, and the implementation phase never gets resourced. The report is genuinely excellent. It is also inert. This is the most common outcome by a wide margin.
2. It lands partially. Someone works through the highest-criticality equipment classes — maybe 30% of the study — and the rest stays in the spreadsheet "for phase two." Phase two is not scheduled. The programme is now a mix of RCM-derived tasks and legacy tasks with no way to tell which is which.
3. It lands and then decays. Everything gets entered. But the basis doesn't come with it, so within a few years of planner turnover and cost pressure, the intervals drift and the tasks get pruned. The programme reverts toward where it started, one reasonable-looking decision at a time.
All three have the same root cause. The analysis and the maintenance system are two disconnected worlds, joined by manual transcription, and the transcription drops everything except the task and the interval.
What Should Happen Instead
The chain should stay intact. That's the whole requirement.
A failure mode leads to a decision. The decision produces a task with an interval and a justification. That task becomes a job plan and a PM in Maximo or SAP PM. And from that PM record, you can still get back to the failure mode, the workshop where it was decided, and the person who approved it.
Not in a separate document that has to be maintained alongside the CMMS. In the same chain.
That means:
- The analysis is structured data from the start, not a spreadsheet — functions, failure modes, effects and consequences as records, coded to a standard taxonomy like ISO 14224 so the same failure mode means the same thing across sites
- Decisions run through a governed decision tree with the justification captured as part of the record, not as a free-text note that gets dropped at import
- Approval is enforced before anything is promoted — an approved study becomes an immutable snapshot, so you can always answer "what did we actually agree, and when"
- Accepted tasks promote into import-ready job plans and PMs, with the identifiers already reconciled against the target system, rather than being retyped against a mismatched asset list
- The audit trail survives the promotion, so the planner three years later can see why the task exists before deciding to cut it
That last point is the one that determines whether the programme is still recognisably yours in five years.
The Question to Ask Before You Commission a Study
Most RCM proposals are evaluated on the analysis: whose methodology, how many workshops, which facilitator, how it maps to JA1011.
Ask a different question.
"When the study is signed off, what exactly arrives in our CMMS, and who does that work?"
If the answer is "we deliver the analysis and your team implements it," you have just been told that the most failure-prone part of the project is unresourced and unowned. That's not a reason to cancel the study. It is a reason to scope the landing before you start, rather than discovering it eighteen months later when the champion has moved on.
And if you already have a completed study sitting in a folder — it is not wasted. The intellectual work is done and it doesn't expire. What's missing is the path from that folder into the maintenance system, with the reasoning still attached.
Final Thought
RCM has a reputation in some organizations as an expensive exercise that didn't change much. That reputation is usually unearned by the analysis and entirely earned by the handover.
The study wasn't wrong. It just never reached the only place where maintenance actually happens.
A maintenance strategy that lives in a report is a document. A maintenance strategy that lives in your CMMS, with its reasoning intact and its decisions traceable, is a programme.
Which one did you pay for?
AssetStage keeps the chain intact from failure mode to work order — JA1011-aligned FMEA with ISO 14224 coding, a governed decision tree with independent approval, and promotion into import-ready job plans and PMs for Maximo and SAP PM with the full audit trail attached. Book a demo to see it against your own study.
Related reading: The Excel Handover Every CMMS Project Dreads and The Hidden Cost of Ignoring Criticality in CMMS Data.