How to build a preventive maintenance plan

Almost every company has a preventive maintenance plan. Far fewer actually follow it, and the reason is almost never lack of will: it’s that the plan is written somewhere that demands nothing.

If you’re looking for software to manage it, here’s what needs to be built —with software or without it— and in what order.

Step 1: which equipment goes in

Not all of it. Starting with the full inventory is the most common mistake, and the one that leaves the most plans half-finished: six months loading data before seeing any result wears out any team.

Criticality is decided with three questions. If this asset fails, does it stop production or service? Does it pose a risk to people? Is there a regulatory obligation behind it? Anything that answers yes to one of these goes into the first phase.

For each one you need the bare minimum: where it is, what model it is, its serial number, when it was installed and when its warranty ends. That last piece of data pays for itself, because it’s very common to pay for repairs on equipment that was still under warranty.

In asset management it’s worth grouping them by model and family from the start. That’s what lets you configure something once for fifty identical units.

Step 2: what needs checking

This is where the plan is won or lost, because “check the machine” means something different to every person.

What works is writing it down: what gets checked, in what order and against what reference value. In GMAO Cloud that’s checklists, with their fields, their label and —where it applies— their minimum and maximum value.

That last part is what separates a useful plan from a decorative one. A checklist of tick boxes says someone looked; one with values says what they saw. And when a reading falls outside range it’s logged as an anomaly on the spot, not as a note on the side that no one is going to reread.

Checklists resolve in cascade: if the specific asset doesn’t have its own, the system looks for its model’s, then its subfamily’s, then its family’s. That way you define it once per equipment type and only fine-tune where it’s needed.

Where the content comes from: the manufacturer’s manual, the standard when there is one, and above all the people who’ve been servicing that equipment for years. That third source is the most valuable, and the one that disappears when someone retires.

Step 3: how often

The initial frequency comes from the manufacturer or the standard, and gets corrected afterward with the history. No one gets it right the first time.

The same piece of equipment can have several frequencies at once, which is how reality actually works: the monthly check is one thing and the annual overhaul with part replacement is another, and neither replaces the other. Each frequency keeps its own record of when it last ran, so the system knows which one is due first.

Step 4: who does it, and on a day that actually exists

This is the step that turns the plan into real work. The frequency has to produce work orders with an owner and a date, not reminders: a task that belongs to no one falls through silently.

Preventive maintenance generates these on its own, and before it does it checks two things a spreadsheet ignores: whether that day is a holiday and whether the technician is available. Scheduling a check for August 15th or for someone on vacation isn’t planning, it’s manufacturing a delay in advance.

After that you need to be able to move it, because the plan changes every day: a breakdown comes in, the client won’t give access, a part doesn’t arrive. In the calendar you drag it to another day, and the Gantt view shows each technician’s load across the weeks. If updating the plan costs more than ignoring it, it gets ignored.

Step 5: make it happen where the work happens

If the technician has to go to a computer to log what they saw, they won’t log it. And maintenance happens in basements, warehouses and machine rooms, which is exactly where there’s no signal.

In the technician app they fill out the checklist —on preventive work orders they can’t close the order without it—, log time with a stopwatch, consume material from the warehouse, attach photos and collect a signature. It all works offline and syncs afterward; if an action fails, it’s flagged with its reason instead of being lost.

Step 6: close the loop on anomalies

The step that kills the most plans when it’s missing. A technician spots something during a check, logs it, and nothing happens. By the third time they stop logging it, and rightly so.

An anomaly needs to become an incident with its priority and its owner, or an explicit decision to do nothing. Both outcomes are valid; silence isn’t.

It’s also the channel through which a preventive plan actually starts reducing corrective work: not by checking more, but because what’s spotted during the check ends up as planned work instead of a breakdown three months later.

Step 7: check whether it’s being followed, and correct it

A plan that’s never reviewed stays frozen on the day it was written.

Reports answer the four questions that matter: what percentage of the annual plan has actually been executed, which assets concentrate the anomalies, how many hours go to preventive versus corrective work, and how much it costs to maintain each asset.

That’s what lets you correct the frequency. Equipment that fails between checks needs more frequent visits; one that never gives trouble is probably over-maintained, and that costs money too. Both situations coexist in almost every facility.

If mandatory maintenance is involved

Worth knowing before you go looking for a dedicated module: in GMAO Cloud legal maintenance is not a separate module. It’s done with this same mechanism, by attaching the check routine the standard requires to the asset, its model or its family, and defining its frequency.

The evidence comes from the order history, the completed checklists with their values, and the associated documentation with its dates. The software records and proves; the company is the one that complies with the standard.

If you’d like to set this up on your own equipment, you can request a demo or get in touch.