Most preventive maintenance schedules die in week three. Not because the team lacks discipline — because the schedule is built backwards. Someone downloads a template with sixty tasks, pins it to the wall, and for two weeks the shop hums. Then a rush order lands, the sixty lines sit unchecked, and the wall art becomes a monument to good intentions.
A schedule with sixty tasks of equal weight has no priority. Everything is important, so nothing is owned. The schedule that survives covers the ten machines that can actually stop the business — and is honest about the rest.
1. Ten machines, not the whole plant
Rank every machine by failure pain, not age or price:
- Stops revenue if it dies — the oven, the espresso machine, the production printer. These are the core.
- Creates safety or liability — door closers, van brakes, the ladder everyone shares.
- Just annoys — the label printer with a spare in the drawer. These get a note, not a schedule: run to failure, spare on the shelf. Written down, that's a decision. Unwritten, it's neglect waiting for a Friday.
For each of the ten, record its peak dependency — the window where the business can't survive it being down. The bakery's deck oven is most critical at 5am Saturday, not 2pm. Build the schedule backwards from that window: risky work happens at 2pm Saturday, never 6pm Friday.
2. Five frequency words, not a matrix
- Daily (2 minutes, folded into the existing opening/closing checklist): clean, look, listen.
- Weekly (15–30 min, same slot every week): lube, filter, drain, one test run — in the dead hour, not before the rush.
- Monthly (about an hour): gaskets, belts, calibration, the panel nobody opens.
- Quarterly (half a day, booked like a job): professional service, wear-limit parts.
- Annual (the shutdown day): the strip-down you can only do when the doors close.
The rule that keeps tiers honest: if a tier holds more than six tasks, move tasks down a tier until it fits its slot. A weekly list of twenty items isn't a schedule; it's a wishlist with reminders attached.
3. Two lines per task
"Check the oven" is not a task — it's a vibe. A task is checkable by someone who wasn't in the room:
- What done looks like: Close the door on a dollar bill at four points — even resistance all round, no cracks at the corners. Replace if the bill slides.
- The parts line: part number, where the spare lives, reorder point. A PM that finds a cracked gasket with no gasket in the drawer hasn't prevented a breakdown — it has scheduled one with better lighting. The $30 spare on the shelf is what turns a three-day supplier wait into a twenty-minute swap.
And one owner per asset: a name, not "the team." "The team" owns nothing.
4. The downtime log that tunes the schedule
Four fields on a clipboard by the machine: date, machine, what failed, minutes down. That's the whole log.
Twenty minutes a month, three rules:
- Two failures on one machine in a quarter → raise its frequency a tier, or fix the design — the part that keeps failing gets replaced with a better part, not lubed a fourth time.
- Zero failures for a year → lower the frequency a tier. This is where the schedule starts paying for itself.
- Any failure inside a peak-dependency window → that job moves permanently to the off-peak slot, and its part joins the drawer.
The bakery that got its Saturdays back
A 22-cover neighbourhood bakery ran on a 1998 deck oven — single unit, no backup. Six unplanned events in Q1: two seized door hinges (each a 40-minute Saturday delay), a cracked gasket mid-bake (twelve batches remade), a drifting prover thermostat, and a mixer belt squeal that was ignored twice and then snapped mid-dough. Roughly nineteen hours of unplanned downtime in thirteen weeks, every hour inside the window the oven matters most.
They built the one-page schedule: daily clean-and-listen folded into the closing checklist, weekly slot Saturday 2pm, monthly gasket/hinge/thermostat line, quarterly tech visit booked like a wholesale order, four-field log on a clipboard, spares in the drawer from day one.
Q2: one unplanned event — the prover thermostat drifted again, caught on the monthly line, 25 minutes to fix instead of a lost morning. The spare gasket turned a would-be three-day wait into a twenty-minute swap. The oven carries roughly $2,800 every Saturday; the schedule costs about forty minutes a week.
The counterfactual is the Q1 path continued: the belt snaps during Saturday rush, $4,000 of dough goes to the bin, and the queue of regulars learns to buy their loaf somewhere else.
Five traps
- The wall of tasks — sixty lines because the template came with sixty. Ten owned beats sixty aspirational.
- PM without parts — finding the crack is not prevention; the spare in the drawer is.
- No owner — "the team" means nobody by March.
- The schedule that never tunes — January frequencies defended forever.
- Backfilled checkmarks — the Friday-ticked month destroys the only signal the log carries. An empty box on a missed week is information.
Kits — the paid tools behind the free advice:
- The First 30 Minutes — free incident quick-start checklist
- Ops Starter Kit — incident response for small teams — $14
- Ops Starter Kit Vol. 2 — advanced incident response & communications — $27
- Ops Mega Bundle — all 5 kits in one download — $49
Related on the site: the monthly ops report is the one page that makes last month legible; the incident post-mortem template is where a breakdown that reached customers gets its root cause.
Top comments (0)