An equipment maintenance management program is four decisions, and most sites make the last one first. The decisions are: what is on the register, which of those things matter most if they stop, what each one needs and how often, and whether the people you have can do that much work. Programs usually begin at step three, with somebody applying a sensible interval across everything, and they fail at step four, when the arithmetic nobody did turns out not to work. This page takes the four in order, because the order is what makes a program survive its first bad month.
Finish the register, including the orphans
The assets that get missed are the ones between departments. A compressor everybody uses and nobody owns, a roof-mounted extract nobody has seen, the standby generator that belongs to whoever last needed it. These are disproportionately the ones that fail badly, because nothing has ever been scheduled on them.
Rank by what stops if it fails
Criticality is what turns a flat list into a program. What halts, how long a replacement takes, whether there is a spare, whether anybody gets hurt. High-criticality assets earn tighter intervals and spares; low-criticality ones may earn a deliberate run-to-failure decision, written down with the reason.
Set intervals per asset, from the manual, then on evidence
The manufacturer's figure for the duty it assumed, adjusted for yours. A pump in a dusty room on continuous duty is not the pump in the manual. Where you depart, write down why, so the next person can see it was a decision rather than a drift.
Cost it in hours before you commit to it
180 assets at four tasks a year at 1.1 hours is 66 technician hours a month, $3,168 at a $48 loaded rate on the worked example this site publishes. Against three technicians that is carryable; against one who also takes every call-out it is not, and no amount of discipline changes that.
Questions people ask about equipment maintenance management program
How long does it take to build a program?
The register is the long part and is usually weeks rather than days. The interval decisions go quickly once criticality is ranked, because most assets fall into a handful of patterns.
Should every asset be in the program?
Every asset should be on the register; not every asset needs scheduled work. Deciding that some are run-to-failure is part of a program, not a hole in one.
What is the first thing to measure?
Reactive share. It is the number that tells you whether the program is doing anything, and it needs no new instrumentation beyond honest work orders.