Preventive Maintenance Checklists: Templates and Real Examples

I have reviewed a great many preventive maintenance checklists over 22 years of CMMS and EAM implementations, and the pattern is consistent enough to be depressing. A typical checklist reads: "Check belt condition. Check bearing. Check alignment. Check for leaks. Clean unit." Five tick boxes, a technician signature, done in eleven minutes. Nothing on that sheet tells you the belt tension, the bearing temperature, the current draw, or whether the person was even at the right asset. Six months later the unit fails, somebody pulls the PM history, and every single record says "satisfactory". That checklist did not improve reliability. It generated compliance paperwork. The difference between the two is entirely a matter of how the checklist is written, and that is a design skill you can learn in an afternoon.

The message up front: a preventive maintenance checklist earns its place only when each line asks for a measured value or an explicit acceptance criterion, not a tick. Replace "check bearing" with "record bearing housing temperature, degrees C, acceptance: below 80 and no more than 15 above ambient", and the same eleven-minute task starts producing trendable data, defensible evidence, and real findings. Everything else in this guide follows from that one change.

1. What a PM checklist is actually for

Before the templates, it is worth being precise about the job a checklist does, because most are written as if they have only one purpose when they in fact have four, and the four pull the wording in different directions.

• Instruction. It tells a technician, including one who has never touched this asset, exactly what to do and in what order. This is the purpose everyone remembers.
• Detection. It surfaces developing faults early enough to plan the repair rather than react to a breakdown. A checklist that produces no findings over a year is either inspecting a perfect asset or not inspecting anything.
• Data capture. It records measured readings that build a trend. One bearing temperature is a number. Twelve monthly bearing temperatures are a degradation curve, and that curve is the cheapest condition monitoring you will ever own.
• Evidence. It proves to an auditor, a client, a regulator or an insurer that the work was done, by whom, when, and to what standard. In a contracted facilities environment this purpose is often the one that gets paid for.

The tick-box checklist serves the fourth purpose and fails the other three. Tension between them is real: a technician on a route of forty air handling units will resist a checklist that demands twenty readings per unit, and they will be right to resist if half those readings are never looked at again. The design discipline is to demand a measured value only where the value will be trended or compared against a limit, and to let everything else be a genuinely quick visual confirmation. For the wider programme-level version of this argument, see the PM programme design pillar, and for the strategy layer that decides which assets get a checklist at all, the complete guide to preventive maintenance.

2. The six rules that separate a useful checklist from paperwork

These are the rules I apply when auditing an existing PM library or writing a new one. They are not stylistic preferences. Each one closes a specific failure mode I have seen repeatedly in live systems.

3. Anatomy of a checklist: the fields every template needs

Whatever the asset, the structure is the same. A checklist has a header block that establishes context and a task block that captures work. Get the header wrong and the data is unusable later even if every task was performed perfectly.

The isolation and permit line deserves emphasis. Any PM that involves opening an enclosure, breaking into a pressurised system or working on live electrical equipment should carry its permit reference in the same record, not in a parallel paper process. The integration pattern for that is covered in the permit to work integration pillar.

4. Checklist 1: air handling unit, quarterly

Air handling units are the single most common PM asset in commercial facilities and the single most commonly reduced to tick boxes. This is a quarterly mechanical and electrical service checklist for a typical belt-driven or direct-drive AHU with heating and cooling coils. Adjust the acceptance values to the specific unit design data; the point is that every line has one.

Notice how many lines produce a number. Twelve months of that AHU checklist gives you a filter loading curve, a bearing temperature trend, a current-draw trend and a coil delta T history, all from a routine visit you were already paying for. That is condition monitoring without buying a single permanent sensor, and it is the single highest-return change most facilities teams can make to their PM library.

5. Checklist 2: LV electrical distribution panel, annual

Electrical panel PMs carry the highest consequence and the weakest checklists, largely because most of the meaningful work requires either a permit or a shutdown, so it quietly gets skipped. Split the checklist explicitly into live-safe tasks and isolated tasks so the deferral becomes visible rather than invisible.

The line that matters most on this sheet is insulation resistance recorded as a value rather than a pass. A panel that reads 800 megohms this year, 400 last year and 120 the year before is on a clear trajectory even though all three readings pass. Pass or fail throws that away. This is exactly the argument for numeric fields over boolean fields, and it applies to every regulated test on the estate.

6. Checklist 3: centrifugal pump and motor set, monthly

Pump and motor sets are where a well-written checklist earns the most, because the dominant failure modes (bearing wear, misalignment, seal degradation, cavitation) all give weeks of detectable warning through readings a technician can take in fifteen minutes. This checklist assumes a running set and a duty or standby pair.

The duty and standby rotation line is one of the most commonly omitted tasks in facilities PM libraries and one of the cheapest to add. Unrotated pairs end up with one pump at 40,000 hours and its twin at 900, which means the standby is effectively untested when the duty finally fails.

7. Checklist 4: standby diesel generator, monthly and annual

Standby generators are unusual in that the asset spends almost all its life not running, so the checklist has to prove readiness rather than assess a running machine. A generator PM that consists of a no-load monthly test run and nothing else is the classic example of a checklist that produces a clean history and a failed start when it matters. Reference the relevant standard for your jurisdiction; in many facilities the governing document is NFPA 110 for emergency power systems.

8. Checklist 5: generic asset PM, for anything without a bespoke sheet

Every asset register has a long tail: small fans, dampers, ancillary tanks, door sets, minor electrical items. Writing a bespoke checklist for each is not a good use of anyone's time, and it is how PM libraries bloat into thousands of unmaintainable job plans. A single well-written generic checklist, applied with judgement, covers the tail properly.

Two lines in that generic sheet do disproportionate work. The asset tag check gradually cleans up a register that nobody ever scheduled a data-quality project for, and the register data verification turns every annual visit into a free data-integrity audit. Over two or three cycles a generic checklist applied across the tail will find more bad master data than any dedicated cleansing exercise.

9. Choosing the frequency without guessing

The most common question after "what goes on the checklist" is "how often", and the most common answer in practice is "monthly, because that is what the last system had". Frequency should come from three inputs, in this order.

• Statutory and insurance obligation. Where a test interval is mandated, that is the floor and it is not negotiable. Record the governing reference on the job plan.
• OEM recommendation, adjusted for duty. Manufacturer intervals assume a duty profile. A generator run twice a year and a generator on weekly duty are not on the same schedule even when the manual says otherwise.
• Detection interval versus warning period. For condition-detecting tasks, the inspection interval must be shorter than the interval between a fault becoming detectable and the asset failing. A fault with a six-week warning period inspected quarterly will still surprise you.

Then adjust with evidence. If a quarterly task has produced no findings in three years across a population of similar assets, that is a candidate for extension. If an asset class keeps failing between visits, the interval is too long or the checklist is looking for the wrong things. Criticality drives how conservative you are in both directions, which is why the asset criticality classification work should precede the frequency decisions rather than follow them. The broader time-versus-meter-versus-condition trade-off is covered in the PM strategies pillar.

For building services specifically, a maintained industry schedule library such as SFG20 gives you a defensible starting position and saves writing several thousand task lines from scratch. Treat it as a baseline to tailor against the actual asset and duty, not as a library to import wholesale and forget, which is how organisations end up with PM volumes their labour capacity cannot deliver.

10. Turning a checklist into a CMMS job plan and task list

A checklist on paper or in a spreadsheet is a document. A checklist in the CMMS is an operating mechanism, because it generates work, captures typed data, enforces rules and builds history. The vocabulary differs by platform but the structure is the same everywhere.

The configuration decisions that actually determine whether the checklist works:

• Make readings typed numeric fields, not text. If bearing temperature is captured in a comment box, no trend, no alarm and no report will ever exist. Every value with a limit belongs in a meter or measurement point.
• Set the limits in the system, not only on the sheet. Upper and lower action limits configured against the measurement point let the CMMS flag the exception automatically rather than relying on the technician to notice.
• Make exception handling mandatory, not discretionary. An out-of-limit reading should require a failure code and generate a corrective before the PM can be completed. This single rule converts inspection findings into planned work instead of forgotten observations.
• Version the job plan. When a task set changes, revise rather than overwrite, so historic records still make sense. Auditors ask this question and unversioned libraries cannot answer it.
• Use routes for repetitive small assets. Forty similar fan coil units on one floor should be one route work order with forty sets of readings, not forty separate work orders that swamp the backlog.
• Design for the mobile screen. Anything the technician cannot complete on a phone in the plant room will be completed later at a desk from memory, which is how "satisfactory" gets typed forty times in a row.

Whether the PM raises a follow-on corrective, and how that corrective is classified and prioritised, is part of the wider work-order taxonomy. If PM findings are not separable from breakdowns in reporting, you cannot prove the programme is working. See the work order types pillar for that structure, and the failure codes pillar for the Problem, Cause, Action coding that makes findings analysable. If you are still selecting a platform, the capability to define typed measurement points with limits inside a job plan is a genuine differentiator worth testing in the demo; the CMMS buyer shortlist covers how to structure that evaluation.

11. Evidence, photographs and the honest audit trail

Evidence requirements are where checklist design collides with trust, and it is worth being candid about that. Mandating a photograph of the asset tag, the filter face and the panel interior does two things at once: it gives a reviewer a second opinion on condition, and it makes it materially harder to complete a PM without attending. Both are legitimate, and technicians generally accept the first reason more readily than the second, so lead with the first.

What works in practice:

• Be selective. Three mandatory photographs per visit is sustainable. Fifteen is not, and the result is deliberately poor images taken to satisfy a counter.
• Specify the shot. "Photo of filter face after replacement" produces something useful. "Attach photos" produces a picture of a floor.
• Prefer a scan to a photo for identity. A QR or barcode scan is faster for the technician and better evidence than an image of a nameplate.
• Capture the reading, not a photo of the reading. A photo of a gauge is not trendable. Type the number and photograph only what a number cannot express.
• Review the evidence. If nobody ever opens the attachments, the requirement decays within a quarter. A short weekly sample review by the planner is enough to keep it alive.

12. The checklist mistakes I see most often

• Tick boxes with no acceptance criterion. The root of nearly every other problem on this list.
• Readings captured as free text. Unusable for trending, alarming or reporting. The number exists but the data does not.
• Compound task lines. "Inspect, clean and lubricate" hides two-thirds of a partial completion behind one result.
• Findings recorded as prose with no failure code. The information is in the record and invisible to every report.
• Regulated tests recorded as pass or fail. Discards the degradation trend that was the whole value of taking the measurement.
• Imported vendor libraries never tailored. Thousands of task lines, no relationship to the actual assets, immediate capacity failure.
• No isolation or permit linkage. The safe system of work lives in a parallel paper process and the audit trail cannot be reconstructed.
• Checklists nobody reviews. A PM library that has not been revised in five years is describing an estate that no longer exists.
• Measuring completion rather than findings. A 99% PM completion rate with almost no defects raised is not excellence, it is a strong signal that the checklists are not asking for anything. Track findings per PM alongside completion, as covered in the FM KPI framework.

The idea to walk away with

A preventive maintenance checklist is a data-collection instrument disguised as a set of instructions. Written as tick boxes it collects nothing, and a decade of "satisfactory" records will tell you absolutely nothing about how your estate is degrading. Written as measured readings against explicit acceptance criteria, with mandatory failure coding on exceptions and a small amount of specified evidence, the exact same visit by the exact same technician produces trends, findings, planned corrective work and a defensible audit trail.

The five checklists above are starting points, not finished documents. Take them, replace the acceptance values with your own design data and OEM figures, cut the lines that do not apply to your assets, and build them into the CMMS as versioned job plans with typed measurement points and configured limits. That is the whole method, and it requires no new software, no sensors and no budget approval.

Final thoughts

If you want one action out of this article, take the ten most critical assets on your register and rewrite their checklists so that every line either produces a number with a limit or asks a question a technician could not answer from the car park. Do not attempt the whole library. Ten assets, done properly, will produce findings within two cycles, and those findings are the argument you will need when you ask for the time to do the next hundred.

The rest is discipline: review the readings, act on the exceptions, revise the job plans when the evidence says the interval or the task set is wrong, and resist the pull to add task lines without removing any. A small, sharp, well-instrumented PM library beats a large one that only collects signatures, every time, in every organisation I have worked with.

Related reading: Preventive maintenance: the complete guide, PM programme design: quality over quantity, PM strategies: time vs meter vs condition, Failure codes: Problem, Cause, Action, Work order types in a CMMS, Permit to work integration.