Preventive maintenance: building a program that earns its place, with a checklist template
What preventive maintenance is and is not, where tasks come from, the four kinds of task and how the kind is chosen, how intervals are set and corrected, what a checklist has to ask for, how the program is run and measured, and how it is reviewed so it gets smaller and better. With a free checklist template.
Key takeaways
- Preventive maintenance is work done on a schedule to stop a failure or to find one before it matters. It earns its place only when each task can name the failure mode it manages.
- There are four kinds of preventive task: fixed interval service or replacement, inspection, condition monitoring and failure finding. The kind is chosen from how the failure behaves, not from habit.
- An interval has a basis: a wear rate, a failure history, a P-F interval, a statutory period or, as a last resort, a vendor figure. The basis is written down so the interval can be corrected.
- A checklist asks for readings against limits, never ticks. A finding becomes a work order the same day, or the inspection was a walk.
- The program is reviewed on evidence. A task that has found nothing in years, or a failure that keeps arriving between tasks, is telling you the task is wrong.
Preventive maintenance is work done on a schedule to stop a failure from occurring or to find one before it has a consequence. Every site has a program of it. What separates a program that keeps the plant running from one that keeps the tradespeople busy is whether each task can say which failure it exists to manage, whether the interval has a basis, and whether the findings ever change anything.
This guide sets out what preventive maintenance is and is not, where tasks come from, the four kinds of task and how the kind is chosen, how intervals are set and corrected, how the program is built and loaded, what a checklist has to ask for, how it is run and measured, and how it is reviewed so it gets smaller and better rather than larger every year. A checklist template comes with it.
Free download
Preventive maintenance checklist template
A checklist tab with a failure mode and a pass criterion against every task, a task library by equipment class to start from, and a review tab that asks whether each task is earning its place.
What preventive maintenance is, and what it is not
The defining feature of preventive maintenance is that it is planned in advance of a failure, on a trigger that is time, usage or condition. It covers four kinds of work: servicing or replacing a part at an interval, inspecting for visible degradation, measuring condition to catch a failure in progress, and testing protection that would otherwise fail unseen. Corrective maintenance is the other side: repairing a failure that has already occurred, whether found by a preventive task or by a breakdown.
Two confusions are worth clearing early. Predictive maintenance is not an alternative to preventive maintenance. It is one of its kinds, the condition-based one, and it is planned on a schedule like the rest. And the spelling: preventive and preventative mean the same thing. The standards use preventive, so this guide does.
What preventive maintenance is not is a list of everything the vendor manual suggests, done at the intervals the manual gives. Vendor intervals are written for a warranty across every duty the equipment is sold into. They are a starting point where the site knows nothing yet, and most sites know a great deal.
Where tasks come from
A preventive task exists to manage a failure mode. That is the test every task has to pass, and it is the reason the analysis starts with failure modes rather than with tasks. For critical equipment, FMECA or a reliability centred maintenance study lists the failure modes, their consequences and the kind of task that manages each. For the long tail of lower criticality equipment, a template by equipment class does the same job faster, provided the template was itself built from failure modes.
The criticality assessment decides how much analysis each asset deserves. The crusher gets a full study; the sump pump gets the pump template. Both end in the same place: a list of tasks, each tied to a failure mode, each with a kind, an interval and a basis for that interval.
A task that cannot name the failure it manages is a cost without a purpose. Trace it or remove it.
The four kinds of task
The kind of task is chosen from how the failure behaves, not from what the crew is used to doing. Wear that follows hours or tonnes suits a fixed interval. Degradation that is slow and visible suits inspection. A failure that gives measurable warning suits condition monitoring. A failure that is hidden, where the device sits idle until called on, suits a failure-finding test. The P-F curve is the reasoning behind the third, and the fixed plant guide shows the regimes applied across a processing train.
| Kind of task | Suits | Example | Interval comes from |
|---|---|---|---|
| Fixed interval service or replacement | Wear or ageing that follows hours, tonnes or cycles closely enough to predict | Replace screen media at 1,200 operating hours | Measured wear rate or failure history, set short of the expected life |
| Inspection | Slow, visible degradation on items that cannot be instrumented cheaply | Walk the conveyor route, check idlers, skirts, cleaners and structure | How fast the defect grows, with margin to plan the repair |
| Condition monitoring (on-condition) | Failures that give measurable warning: vibration, temperature, oil, current | Vibration readings on the mill motor bearings every four weeks | The P-F interval, so at least two readings fall inside it |
| Failure finding | Hidden failures of protection: trips, interlocks, standby equipment, relief valves | Test the crusher low lube pressure trip every six months | The unavailability that can be tolerated for the protection, and the failure rate of the device |
Setting the interval, and writing down why
Every interval has a basis, and the basis is recorded beside it. A measured wear rate. A failure history from the CMMS. A P-F interval from condition data. A statutory period. A vendor figure, marked as such so it is the first to be challenged. The basis is what lets the interval be corrected later without an argument, because the question becomes whether the evidence has changed rather than whose opinion wins.
For a fixed interval task, the interval sits short of the expected life by a margin that reflects the spread in that life and the consequence of running past it. For a condition task, the interval is set so that at least two readings fall inside the P-F interval, which gives one to see the trend and one to plan the repair. For a failure-finding task, the interval follows from the unavailability of the protection that can be tolerated and the failure rate of the device, and for safety functions this is set by the safety study rather than by maintenance.
Intervals drift towards too short, because a task that finds nothing feels safe and a task that is skipped feels dangerous. The review stage below is where that drift is corrected.
Building the program
The program is the loop in the figure. Failure modes produce tasks. Tasks are loaded into the CMMS as maintenance plans against the asset hierarchy, and grouped into routes for the running tasks and shutdown packages for the stopped ones, so that one isolation and one access serve many tasks. The plans generate work orders on their intervals, the work is done and recorded, and the findings and failures feed a review that changes the tasks.
The loading is where programs most often go wrong in the CMMS. A plan on the wrong level of the asset hierarchy puts the history in the wrong place. A plan with no job plan behind it generates a work order with a title and nothing else. A plan whose counter is never updated generates on calendar time while the interval was meant to be in tonnes. The master data has to be right before the program will be.
1Failure modes
What can fail, and what happens when it does?
The failure modes worth managing, by criticality
2Tasks and intervals
Which task manages each, and how often?
A task per failure mode, with an interval and its basis
3Program
How is it loaded and grouped?
Maintenance plans in the CMMS, grouped into routes and shutdown packages
4Execute and record
Was it done, and what was found?
Readings and findings on the work order, defects raised as work
5Review
Is the task earning its place?
Intervals corrected, tasks removed or added, from findings and failures
→ Each stage feeds the next
↻ Back to the tasks
| Stage | The question it answers | What it produces |
|---|---|---|
| Failure modes | What can fail, and what happens when it does? | The failure modes worth managing, by criticality |
| Tasks and intervals | Which task manages each, and how often? | A task per failure mode, with an interval and its basis |
| Program | How is it loaded and grouped? | Maintenance plans in the CMMS, grouped into routes and shutdown packages |
| Execute and record | Was it done, and what was found? | Readings and findings on the work order, defects raised as work |
| Review | Is the task earning its place? | Intervals corrected, tasks removed or added, from findings and failures |
What a checklist has to ask for
A checklist is the task written down so it can be done the same way every time and reviewed afterwards. The weakest checklists are lists of items with a tick box beside each: they come back complete, say nothing, and cannot be reviewed because there is no reading to compare. The strongest ask for a reading or a result against a limit, and turn a failed limit into a finding and a work order on the spot.
The columns below are the ones in the template. The failure mode column is the one most checklists lack and the one that makes the review possible. The pass criteria column is the one that turns a walk into an inspection.
| Column | What goes in it |
|---|---|
| Tag | The functional location or equipment number the task belongs to |
| Equipment | As it says |
| Task | What is done, stated so a tradesperson new to the site could do it |
| Kind | The kind of preventive task |
| Failure mode managed | The failure this task exists to prevent or find. A task with nothing here is a candidate for removal |
| Interval | The number, with the unit beside it |
| Unit | As it says |
| Running or stopped | Whether the task needs the equipment out of service |
| Pass criteria | A reading, a limit or a condition. Never a tick on its own |
| Reading or result | What was measured or seen. Filled in when the task is done |
| Pass | As it says |
| Finding | What was found if it did not pass, in enough words to raise a work order |
| Work order raised | The corrective work order number for a finding |
| Done by | As it says |
| Date | As it says |
- State the task so a tradesperson new to the site could do it from the words alone, with the tools and the spec named.
- Make the pass criterion a number or a condition. Seal drip rate below the vendor limit, not seal OK.
- Record the reading even when it passes. The trend across readings is where the P-F interval and the next interval correction come from.
- Raise the work order for a finding the same day, with the finding's words on it, and write its number on the checklist.
Running it: the due window and PM compliance
Preventive work orders are scheduled first in the weekly cycle, because they are the only work that can be seen coming, and because a program that is not done cannot be prevented from failing. Each has a due window, often a tenth of its interval either side of the due date, inside which it counts as done on time.
PM compliance is the share of preventive work orders completed inside their window, and it is read beside the corrective backlog. High compliance with a shrinking backlog means the program is working. High compliance with a growing backlog means the program is being done and is addressing the wrong failures, which sends you to the review. Low compliance means the program is being traded for corrective work, and the backlog will grow next. The maintenance KPIs guide sets out the measures that read together.
PM compliance says whether the program was done. Only the corrective backlog says whether it was the right program.
Reviewing the program
A program that is never reviewed only grows. Every incident adds a task and no task is ever taken out, until the crew cannot complete the program and starts skipping the ones that feel least urgent, which is how the important ones get skipped. The review is where the program gets smaller and better, and it runs on evidence the checklists and the failure history already hold.
The signals below are the ones a review looks for, task by task, starting with the equipment that costs the most. Each has a usual meaning and an action. The action is recorded with its basis, so a change that the history later proves wrong can be reversed. The maintenance plan template on the fixed plant guide carries a basis and a last-reviewed date against every task for this reason.
| Signal | What it usually means | Action |
|---|---|---|
| No finding in years of doing the task | The interval is too short, the failure mode does not occur, or the task cannot see it | Extend the interval or remove the task, after checking failure history |
| Failures between tasks | The interval is too long, or the task does not address the failure mode that is occurring | Shorten the interval or change the kind of task, from the failure data |
| Task is done, failure happens anyway | The task is the wrong kind: an inspection cannot see a random failure | Replace with a condition task if the failure gives warning, or accept run to failure |
| Task has no failure mode written against it | It came from a vendor manual or a template, not from analysis | Trace it to a failure mode or remove it |
| Checklist comes back all ticks, no readings | The pass criteria are not measurable, so the task cannot be reviewed | Rewrite the criteria as readings and limits |
| PM compliance high, corrective backlog growing | The program is not addressing the dominant failures | Run the review on the worst actors first |
Where programs fail
- The vendor manual loaded as the program, so the plant does warranty maintenance for twenty years.
- Tasks without a failure mode, which cannot be reviewed and are never removed.
- Tick-box checklists that come back complete and say nothing.
- Findings recorded on the checklist and never raised as work, so the inspection found the defect and the breakdown found it again.
- Counters not updated, so tonnage-based intervals run on the calendar.
- Intrusive fixed interval tasks on equipment that gives warning, which introduce the failures they were meant to prevent.
- No review, so the program grows until it cannot be completed and the crew chooses which tasks to skip.
Common questions
What is preventive maintenance?
Maintenance done at a planned interval, in time, usage or condition, to stop a failure from occurring or to find a failure before it has a consequence. It includes fixed interval servicing and replacement, inspections, condition monitoring tasks and the failure-finding tests done on protective devices. It is distinguished from corrective maintenance, which repairs a failure that has already happened.
Is it preventive or preventative maintenance?
Both spellings are in use and mean the same thing. Preventive is the form used in the standards, including ISO 14224 and the SMRP metrics, and is the one this site uses. Preventative is common in Australian and British everyday usage. Search for either and you will find the same material.
What is the difference between preventive and predictive maintenance?
Predictive maintenance, also called condition based maintenance, is a kind of preventive maintenance. A fixed interval task is done at a set time whether or not the equipment needs it. A predictive task measures the equipment's condition on a schedule and triggers work only when a reading crosses a limit. Both are planned in advance, which is what makes them preventive rather than corrective.
What should a preventive maintenance checklist include?
For every task: the equipment it belongs to, what is done in words a new tradesperson could follow, the failure mode the task manages, the interval and its unit, whether the equipment has to be stopped, and a pass criterion that is a reading or a limit rather than a tick. When the task is done: the reading or result, whether it passed, the finding if it did not, and the work order raised for it.
How often should preventive maintenance be done?
At an interval with a basis. For wear that follows hours or tonnes, short of the measured life. For inspections, inside the time a defect takes to grow from visible to serious. For condition monitoring, so that at least two readings fall inside the P-F interval. For failure finding on protection, from the unavailability that can be tolerated. A vendor figure is a starting point where nothing else is known, and the interval is corrected as findings and failures come in.
What is PM compliance?
The share of preventive work orders completed within their due window, over all that fell due in the period. The window is usually set as a fraction of the interval, ten per cent being common. It is a leading measure: a program that is not being done cannot be prevented from failing. A high PM compliance beside a growing corrective backlog says the program is being done but is addressing the wrong failures.
How do you reduce preventive maintenance workload without adding risk?
By review, task by task, against evidence. Tasks with no failure mode written against them are traced to one or removed. Tasks that have found nothing in years have their interval extended or are dropped, after the failure history is checked. Intrusive fixed interval tasks on equipment that gives warning are replaced by condition tasks. Low consequence items with a spare on the shelf are moved to run to failure. Each change is recorded with its basis, so it can be reversed if the history proves it wrong.
Key terms
Plain-language definitions from our glossary for the concepts this article leans on.
Standards and further reading
- Best practices, metrics and guidelines for maintenance and reliability (SMRP)
- ISO 14224:2016 Collection and exchange of reliability and maintenance data for equipment (ISO)
- SAE JA1011 Evaluation criteria for reliability-centered maintenance processes (SAE International)
- IEC 60300-3-11 Dependability management: application guide, reliability centred maintenance (IEC)
- Operations and maintenance best practices guide (US DOE Federal Energy Management Program)
- ISO 55001:2024 Asset management system requirements (ISO)
Related case studies and tools
- Reliability engineering and asset strategy program (case study)
Related reading
Working through something like this?
See how we approach these initiatives, or tell us what you are dealing with.