How to Write PM Checklists Techs Actually Follow

The two ways checklists fail

Most bad PM checklists fall into one of two failure modes, and they fail for opposite reasons.

The vague checklist. “Inspect conveyor.” Inspect what, exactly? The belt? The rollers? The drive motor? And pass by what standard — how worn is too worn, how loud is too loud? A vague item outsources the entire inspection to whatever the tech happens to know and happens to notice that day. Two techs will do two completely different inspections, and neither one is wrong, because the checklist never said what right looks like.

The novel. The opposite failure: a 40-item wall of text, usually copied from an OEM manual, that covers every conceivable step in exhaustive detail. Nobody reads it. The tech does the PM the way they always do it, then checks every box at the end in one pass. By item 12, the checklist has stopped being an inspection tool and become paperwork. Pencil-whipping isn't a character flaw — it's the rational response to a checklist that doesn't respect the tech's time.

The fix for both is the same: write fewer items, and make each one specific enough to be worth doing.

Anatomy of a good checklist item

Every good checklist item has three parts: an action verb, a specific component, and an acceptance criteria — the standard that separates pass from fail. If a new tech can't tell from the item alone whether the machine passes, the item isn't done yet.

Here's what that looks like in practice:

Bad: “Check belt.”
Good: “Check drive belt tension — deflection under thumb pressure should be about 1/2 inch; replace if cracked or glazed.”

Bad: “Inspect hydraulics.”
Good: “Check hydraulic fluid level in sight glass — should be between MIN and MAX lines with the ram fully retracted; top off with ISO 46 if low, and note any milky or dark fluid.”

Bad: “Grease bearings.”
Good: “Grease both pillow block bearings on the main shaft — 2 pumps each with the grease gun on the wall; stop if grease purges from the seal.”

Bad: “Check safety devices.”
Good: “Test the E-stop — press it while the machine is running; the machine should stop within 2 seconds. Fail the item if it doesn't.”

Notice what the good versions have in common. They name the exact component, they tell the tech how to perform the check, and they define the pass/fail line — a measurement, a visual standard, a behavior. The tech doesn't have to guess, and the result means the same thing no matter who did the PM.

Add measurements where trends matter

Pass/fail is the right format for most items, but some failures don't announce themselves with a clean pass or fail. They drift. A motor that drew 12 amps last year draws 13 this month and 14 next month — and every one of those readings, taken alone, looks fine. By the time it fails a pass/fail check, you're replacing a motor instead of a bearing.

For those failure modes, ask for a number instead of a checkmark:

• Temperatures — bearing housings, motor casings, panel interiors. An infrared thermometer reading takes ten seconds.
• Pressures — hydraulic system pressure at idle and under load, compressed air at the regulator, filter differential pressure.
• Amp draws — a clamp meter reading on each motor leg. Rising amps mean rising load or failing insulation.
• Vibration and noise notes — even a subjective “smooth / slight vibration / rough” scale recorded consistently will show a trend.

A number recorded every month catches drift that pass/fail misses. You don't need a condition-monitoring program to get this benefit — you need two or three measurement items on your existing PMs and a place where the readings accumulate so someone can see the line moving.

When the measurement comes from a gauge or instrument that also needs its own due date and certificate history, use a separate calibration tracking workflow for intervals, reminders, certificates, and out-of-tolerance follow-up.

Right-size the list

For a routine monthly or quarterly PM, aim for 5–12 items. That's long enough to cover the failure modes that matter and short enough that every item gets real attention. If you find yourself past 15, you're either padding the list with items that never fail, or you're trying to cram an annual overhaul into a monthly PM.

For the big annual jobs, don't write one 40-step monster. Split the overhaul into separate work orders per subsystem — one for the drive train, one for hydraulics, one for electrical, one for the frame and guards. Each one gets its own focused checklist, its own time estimate, and its own completion record. Shorter lists get followed; they also make it obvious which subsystem is eating your maintenance hours.

One rule with no exceptions: safety and lockout steps come first, always. “Lock out and tag the main disconnect; verify zero energy before opening the guard” is item one on every checklist that involves opening a machine — not a note at the bottom, not an assumption. If the checklist is the procedure, the procedure starts with lockout.

Write for the newest tech, not the 20-year veteran

It's tempting to write checklists for the person who currently does the PM — the veteran who knows that “check the tensioner” means the tensioner behind the left guard, the one with the trick spring. But tribal knowledge leaves in a two-week notice. The day your veteran retires, every vague item on the checklist becomes a mystery, and the machine's reliability walks out the door with them.

The test for every item: could a competent tech on their first week execute it correctly with no one to ask? If not, add the missing detail — where the component is, which tool to use, what the setpoint is.

And use photos. A picture of the correctly tensioned belt, the sight glass at the right level, or the grease fitting hidden behind the bracket beats three paragraphs of description. Attaching a reference photo to a checklist item takes a minute and answers the question every time, forever.

Close the loop

Here's the part most teams skip, and it quietly makes the whole program pointless: every failed checklist item should spawn a follow-up work order. An inspection that finds a cracked belt and doesn't generate a work order to replace it hasn't prevented anything — it's just documented the failure in advance. If failed items go into a notes field that nobody reads, your PMs are finding problems that nobody fixes.

Make the path automatic if your system supports it, or make it a hard rule if it doesn't: fail an item, create a work order, before the PM gets closed.

Finally, review your checklists yearly against your failure history. Pull the breakdowns from the last twelve months and ask two questions of each one: was there a checklist item that should have caught this, and did the machine fail anyway? If a component keeps failing and no item covers it, add one. If an item has passed on every PM for three years and the failure mode has never occurred, consider cutting it. Checklists aren't written once — they're tuned against what actually breaks.