A preventive maintenance (PM) schedule is a documented plan that assigns each asset specific maintenance tasks, inspect, lubricate, adjust, replace, on defined triggers: calendar time, usage such as runtime hours or cycles, or measured condition. Its purpose is to catch wear before it becomes failure, on your calendar instead of the machine's.

Almost every plant has a PM schedule. Far fewer have one that gets followed. This guide covers the three trigger types, a seven-step framework for building a schedule from scratch (or rebuilding one that died), and the compliance metric that separates a real program from a binder on a shelf.

What is a preventive maintenance schedule?

A preventive maintenance schedule is the master calendar of recurring maintenance work: which task, on which asset, how often, by whom, taking how long. It converts a maintenance strategy into assignable work orders. Without it, preventive maintenance is an intention; with it, PM becomes a plan you can staff, schedule, and measure.

A useful schedule has four properties. It is complete every asset that deserves PM has tasks defined. It is specific each task has steps, parts, tools, and an estimated duration, not just a title like MONTHLY PM. It is loaded honestly total PM hours fit inside the crew's actual capacity. And it is alive intervals get adjusted as failure history accumulates. Most schedules fail the third and fourth tests.

Where the schedule lives matters less than whether it generates work automatically. A spreadsheet can hold a schedule; a CMMS can hold it and issue the work orders, track completion, and show you compliance without anyone building a report by hand.

What triggers a PM: time, usage, or condition?

Every PM task fires on one of three trigger types, and choosing the right one per task is most of the design work.

The three PM trigger typesWhat fires the work order?TIME-BASEDcalendar intervalevery 30 / 90 / 365 dayse.g. quarterly beltinspectionsimple · ignores duty cycleUSAGE-BASEDmeter intervalhours · cycles · unitse.g. oil change every500 runtime hoursfairer · needs meter dataCONDITION-BASEDmeasured thresholdvibration · temp · pressuree.g. replace bearing whenvibration exceeds limitmost precise · needs sensorsleft to right: less waste per task, more data required
The three PM trigger types. Most plants run a mix: time-based for simple tasks, usage-based where meters exist, condition-based for critical assets.

Time-based triggers fire on the calendar: every 30 days, every quarter, every year. They are simple to administer and right for tasks where degradation tracks time regardless of use, seals drying out, filters loading in a dusty room, regulatory inspections. Their weakness: a machine that ran one shift and a machine that ran three get identical service.

Usage-based triggers fire on a meter: runtime hours, cycle counts, units produced. They match service to wear, which is fairer and cheaper across mixed duty cycles. The prerequisite is trustworthy meter data, a machine counter someone reads monthly, or better, a machine connection that streams it. Plants that have connected their equipment to a live data layer (the approach platforms like Harmony take, pulling from PLCs and sensors without rip-and-replace) get usage triggers essentially for free.

Condition-based triggers fire when a measured parameter crosses a threshold, vibration, temperature, oil chemistry, current draw. This is condition-based maintenance and at higher sophistication, predictive maintenance. It is the most precise trigger and the most demanding: sensors, baselines, and someone who owns the readings.

The right question per task is not which trigger is best, but which failure mode you are trying to intercept and what data you actually have. A schedule of 100% time-based PMs is a fine starting point. A schedule that is still 100% time-based five years in has stopped learning.

How do you build a PM schedule? The 7-step framework

Build the schedule asset-by-asset, starting from criticality, and load it against real crew capacity. Here is the sequence.

  1. Build the asset register. List every maintainable asset with an ID, location, and criticality rating. You cannot schedule what you have not named. Rank criticality by consequence of failure: safety, production stoppage, cost. A simple A/B/C ranking beats a stalled perfect one.
  2. Pick the assets that deserve PM. Not everything does. Cheap, redundant, low-consequence assets can deliberately run to failure, see the maturity ladder in our equipment reliability guide. Spending PM hours on a $400 fan with a spare on the shelf steals hours from the bottleneck line.
  3. Define tasks per asset from failure modes. Start with the manufacturer's manual, then adjust with your own failure history and your senior technicians' knowledge. Each task needs: steps, parts, tools, safety requirements (lockout points), skill level, and an honest duration estimate.
  4. Assign a trigger and interval to each task. Time, usage, or condition, per the section above. When in doubt, start with the manufacturer interval and tighten or relax it later based on what inspections actually find.
  5. Load-level against crew capacity. Add up the PM hours per week and compare to the hours your crew genuinely has after reactive work and meetings. If the schedule demands 300 hours and the crew has 180, the schedule is fiction. Cut low-value tasks or stretch intervals until it fits. This step is where most schedules are silently doomed.
  6. Generate and assign the work. Put the schedule in a system that creates work orders automatically on trigger, with the task detail attached. Paper calendars and memory do not survive vacation season. This is the core job of a CMMS and pairing it with good planning and scheduling practice is what turns generated work into completed work.
  7. Measure compliance and tune intervals. Track PM compliance weekly (next section). Review findings quarterly: PMs that never find anything are candidates for longer intervals; failure modes that keep surprising you need new or tighter tasks.
PM schedule build flow1 · ASSET REGISTERID · criticality2 · SELECT ASSETSPM vs run-to-failure3 · DEFINE TASKSsteps · parts · duration4 · SET TRIGGERStime · usage · condition5 · LOAD-LEVELfit real crew hours6 · GENERATE WOsauto from CMMS7 · MEASURE + TUNEcompliance · findingsfeedback: adjust intervals
The seven-step build flow. Step 5 (load-leveling) is where most schedules quietly fail; step 7 feeds back into trigger tuning.

What does a schedule entry actually look like?

A schedule is a stack of entries like these, hypothetical, but the shape is the point. Note that every row has a trigger, an owner-level skill, and an honest duration; rows missing any of those are not schedulable, they are hopes.

AssetTaskTriggerSkillEst. time
Filler #2 (A-critical)Inspect and lubricate main drive chain; check tension against specEvery 250 runtime hoursMechanic45 min
Filler #2 (A-critical)Vibration reading, drive-end bearing; log against baselineMonthly routeTrained operator10 min
Air compressor #1 (A)Oil sample to lab; replace inlet filterQuarterlyMechanic30 min
Case packer (B)Full mechanical inspection per checklist; replace wear strips as foundEvery 90 daysMechanic2 h
Exhaust fan, warehouse (C)None, run to failure, spare on shelf

Two details worth copying. The vibration reading is assigned to a trained operator, not a technician, routine condition checks are exactly the kind of work autonomous maintenance moves to the people already standing at the machine. And the C-class fan appears on the schedule explicitly as run-to-failure: a documented decision, so nobody quietly adds it back with a monthly greasing task that buys nothing.

How do you measure whether the schedule is working? PM compliance

PM compliance is the percentage of scheduled PM work orders completed within their due window:

PM compliance = (PMs completed on time ÷ PMs scheduled) × 100

The commonly cited target, aligned with the metrics published by the Society for Maintenance and Reliability Professionals (SMRP), is 90% or higher with tighter expectations for the most critical assets. Two honest caveats. First, define the due window before you measure, a common convention is completing the PM within 10% of its interval (a 30-day PM done within 3 days of due). Second, 100% compliance on a bloated schedule is worse than 92% on a lean one; compliance measures execution, not schedule quality.

Watch the trend, not the single number. A schedule that launches at 95% and slides five points a month is telling you the load-leveling was wrong or reactive work is eating the crew, check your backlog in crew-weeks alongside it.

PM compliance over time (hypothetical example)PM compliance, first 12 weeks of a rebuilt schedule50%70%90%target 90%wk 1wk 6wk 12dip: overloaded weekafter load-leveling fix
A hypothetical compliance ramp after a schedule rebuild. The week-3 dip is typical: the first honest capacity check usually forces a load-leveling correction.

Why do PM schedules die, and how do you make one stick?

PM schedules die from predictable causes, and every one has a countermeasure.

Overloading. The schedule demands more hours than the crew has, compliance drops, and skipping PMs becomes normal. Countermeasure: load-level ruthlessly (step 5) and cut tasks that never find anything.

Reactive work always wins. Breakdowns feel urgent; PMs feel optional. Countermeasure: protect PM hours in the weekly schedule the way production protects changeovers, and track schedule compliance so break-ins are visible. Involving operators in first-line care through autonomous maintenance a pillar of TPM offloads the simplest tasks entirely.

Paper and tribal memory. If PMs live on a wall calendar and in one planner's head, the program dies with a resignation letter. Countermeasure: digitize the schedule and its task detail. Plants like CLS replaced paper logging with tablet-based capture and automated reporting, the CLS case study shows what that shift looks like in practice.

No feedback loop. Intervals set in 2019 and never touched. Countermeasure: quarterly review of PM findings and failure history; tune intervals with evidence. Your maintenance KPIs PM compliance, MTBF trend, emergency-work percentage, will tell you whether the tuning is working.

What does preventive maintenance actually pay?

The best public numbers come from the U.S. Department of Energy's Federal Energy Management Program guidance, maintained by Pacific Northwest National Laboratory:

The honest counterweight: PM is not free, and over-maintaining is a real failure mode. Every intrusive PM carries a small risk of maintenance-induced failure, and every unneeded task burns capacity. The goal is the right work at the right interval, not the most work.

Where does the schedule go from here?

A PM schedule that holds 90% compliance for six months is the foundation for everything above it on the reliability ladder. Usage triggers replace calendar guesses once machine data flows automatically. Condition thresholds replace fixed intervals on critical assets. The progression, and what each rung costs, is mapped in our guides to condition-based and predictive maintenance. And if your current blocker is that PM records live on paper while the schedule lives in a spreadsheet, that is a data problem before it is a maintenance problem, the kind of digitization work described on our platform overview.