How to build a preventive maintenance program that actually sticks

Almost everyone agrees preventive maintenance (PM) is worth doing. Far fewer organizations have a PM program that survives contact with a busy month. The failure is rarely the idea — it's the rollout. Here's how to build one that's still running a year from now.

Start with the assets that hurt when they fail

Don't try to schedule everything on day one. List the equipment whose failure is expensive, dangerous, or disruptive — boilers, rooftop units, elevators, generators, pumps, fire systems. That short list is 80% of your risk. Get those on a schedule first and ignore the long tail until the core is humming.

Pick intervals you can defend

For each asset, three sources set the interval, in order of priority:

1. Code and warranty: some services are mandatory (fire, elevator) or required to keep a warranty valid. These aren't optional.
2. Manufacturer guidance: the nameplate and manual usually specify service intervals. Start there.
3. Your own failure history: if a pump keeps failing at 10 months, a 12-month PM is too slow. Let real data tighten or loosen the interval over time.

Beware the opposite failure too: over-maintaining. Servicing something monthly that only needs it quarterly burns labour and breeds the “these PMs are pointless” attitude that kills programs. Match the interval to the actual need.

Write checklists a new tech could follow

A PM that says “service boiler” is useless. A PM with a checklist — inspect, test, measure, record the reading — is repeatable, trainable, and auditable. The checklist is the program's memory. It also turns a PM into a data-collection event: trending a measured value (vibration, temperature, pressure) is how you catch failure before it happens.

Automate the scheduling, or it won't happen

The reason paper PM programs die is that someone has to remember them. Software fixes this by auto-generating the work order when a PM comes due and putting it in the same queue as everything else. If your PMs live in a calendar nobody opens, they don't exist. This is the whole point of a CMMS: the system remembers so your team doesn't have to.

Measure two things, relentlessly

• PM compliance: what percentage of scheduled PMs were completed on time? Below ~80% and the program is slipping.
• Reactive vs. planned ratio: as PM matures, the share of emergency/reactive work should fall. If it isn't, your intervals or checklists are wrong.

Keep it alive

Programs decay without ownership. Review the schedule quarterly: kill PMs that never find anything, tighten the ones that keep catching problems late, and add the assets you deferred at launch. A PM program is a living thing, not a one-time setup.

A simple rollout plan

1. List your 10–20 highest-risk assets.
2. Set a defensible interval for each (code > manufacturer > history).
3. Write a real checklist per PM.
4. Put them in software that auto-creates the work orders.
5. Track PM compliance and the reactive ratio.
6. Review and adjust every quarter.

Do that and preventive maintenance stops being a binder you feel guilty about and becomes the quiet reason things stop breaking. See how Arlo handles PM and asset history, or estimate the payoff with the ROI calculator.