CMMS: the asset register and the work orders that turn maintenance into history

The asset hierarchy is the schema for every question you will ask later

How equipment is registered determines what can be analysed. A hierarchy that stops at the machine cannot show that failures cluster on one drive unit; one that descends too far becomes unmaintainable and technicians pick whichever level appears first in a search. Functional location coding, where the position in the process has an identity separate from the equipment currently installed in it, is what allows history to follow the location when a unit is swapped, and to follow the unit when it is refurbished and installed elsewhere. Getting this structure right before loading thousands of records is worth weeks of argument, because restructuring afterwards orphans the history.

A job closed without a cause teaches nobody anything

The difference between a maintenance record and maintenance knowledge is the closure detail: what failed, how it presented, what was found, what was done, what parts were consumed, how long the equipment was down as distinct from how long the technician worked. Free text alone cannot be counted; codes alone cannot be understood. The practical combination is a short coded set that fits on one screen and reflects the language technicians actually use, plus a description field. Long code lists produce whatever entry sits at the top, which is why so many plants discover their most common failure cause is the first item alphabetically.

Calendar-based schedules drift away from the machine

Preventive tasks loaded at commissioning tend to be inherited from a manual and never revisited, so the plant services a machine on a fixed interval regardless of how hard it has run. Where usage data exists, triggering on running hours, cycles or throughput fits the equipment better; where condition can be measured, inspection findings should be able to raise the next job. The metric to distrust is completion of the preventive schedule, which measures compliance rather than reliability and rises comfortably while breakdowns continue. Reviewing whether each recurring task has ever prevented anything is an uncomfortable exercise most maintenance functions have never done.

Spares are where maintenance meets the storeroom

The system needs to know which parts fit which equipment, which of those are critical because the machine cannot run without them, and what is actually on the shelf. Without the equipment-to-part link, a technician searches a catalogue by description and orders something that nearly fits. Without accurate stock, planned jobs are scheduled against parts that are not there and the crew arrives to find a stripped machine and no seal kit. Kitting parts against a planned job in advance is the practice that converts planning into completed work, and it depends entirely on both links being maintained.

Technicians only feed a system that gives something back

Data entry after a shift, on a desktop in an office, at the end of a day spent fixing things, produces the minimum record that closes the job. Access at the machine, on a device that works where the signal is poor, showing previous failures on that equipment and the manual for the assembly in front of them, changes the exchange from an administrative demand to a tool. Where that is not achievable, the honest alternative is a shorter mandatory record and a planner who interviews technicians on the significant jobs, rather than a long form that returns nothing usable.

Frequently asked questions

How much of the asset register should we load before going live?

Load the equipment that generates work and the structure you intend to analyse, not everything the plant owns. A register padded with items that never receive a work order makes searching slower and adds no history. Start with production equipment, critical services and anything under a statutory inspection regime, then extend as need appears. The important discipline is that the coding scheme is settled first, because adding assets later is easy and restructuring existing ones is not.

Should breakdown calls go through the system or straight to the technician?

Urgent calls will always be made by radio or by walking to the workshop, and trying to prevent that wastes credibility. What matters is that a record is created, either by the caller or by the technician afterwards, so downtime and cause are captured. Plants that insist on system-first reporting for emergencies typically end up with a large volume of unrecorded work and a downtime figure that flatters them badly.

What does a maintenance system need from production systems?

Two things above all: equipment running hours or counts, so schedules can be usage-based, and downtime events with their production impact, so the cost of failures is expressed in output rather than in labour hours. A link to the production schedule is also valuable, because planning a shutdown into a window when the line was going to be idle is far cheaper than negotiating access. None of these require deep integration, but all require agreeing which system holds the master equipment identity.