Free your maintenance
Total Preventive MaintenanceTPM, its eight pillars, and how to actually start
Total preventive maintenance — known in industry as Total Productive Maintenance (TPM) — is a plant-wide programme that makes equipment reliability the shared responsibility of operators and maintenance, measured by how much productive capacity the plant loses.
It is not a maintenance schedule. It is a way of organising a factory around eliminating losses, of which unplanned downtime is only one. This guide covers where TPM came from, its eight pillars, how it differs from preventive maintenance, and what a small team can do with it without a consultant or a budget.
First, the name: productive, not preventive
“Total preventive maintenance” is how a great many people search for this topic, and it is worth being precise about why the phrase is slightly off. TPM stands for Total Productive Maintenance. The term was formalised by Seiichi Nakajima and the Japan Institute of Plant Maintenance (JIPM), and the approach is generally dated to Nippondenso — a Toyota supplier — around 1971, where operators already performing autonomous maintenance were folded into a full plant programme.
The distinction is not pedantry, because the word carries the idea. Preventive maintenance is about preventing failure. TPM is about protecting productive capacity — and failure is only one of the ways a plant loses it. The others are the reason the programme exists: setup and changeover time, minor stops, reduced speed, start-up scrap and in-process defects. JIPM groups these as the six big losses, and TPM measures them with OEE(Overall Equipment Effectiveness = availability × performance × quality).
So: if you came here for a servicing schedule, what you want is preventive maintenance. If you came here because someone said the plant should be doing TPM, read on — it is a bigger commitment and a different kind of thing.
The six big losses, and how OEE measures them
TPM does not ask “how often does this machine break?” It asks where the plant's productive capacity goes. JIPM groups the answer into six big losses, and each one falls under one of the three factors of OEE. The grouping is the useful part: it tells you which lever a given loss belongs to.
| Loss | OEE factor | What it is |
|---|---|---|
| Breakdowns | Availability | Unplanned stops from equipment failure. The loss everyone counts, and usually the only one. |
| Setup and adjustment | Availability | Changeover time, plus the trial runs and tweaks before the line is producing to spec again. |
| Idling and minor stops | Performance | Jams, misfeeds, sensor trips — each under a few minutes, none logged, and together often the largest loss on the line. |
| Reduced speed | Performance | The gap between the design cycle time and the rate the machine is actually run at, usually for reasons nobody wrote down. |
| Process defects | Quality | Scrap and rework produced during stable running. |
| Start-up losses | Quality | Yield lost between start-up and stable running — the warm-up scrap that gets treated as a cost of doing business. |
How OEE is calculated
Overall Equipment Effectiveness is the product of three ratios, each one bounded by a pair of the losses above:
- Availability = run time ÷ planned production time
- Performance = (ideal cycle time × units produced) ÷ run time
- Quality = good units ÷ units produced
OEE = Availability × Performance × Quality. The multiplication is what makes the number honest, and uncomfortable: three factors that each look respectable at 90% give an OEE of 72.9%. A line whose operators would describe it as running well is losing more than a quarter of its capacity, and no single measurement would have shown it.
Two cautions that decide whether the number is worth collecting at all. First, the headline figure is nearly useless on its own — an OEE of 65% tells you nothing you can act on, while 95 × 85 × 80 tells you the machine is available and the problem is speed and scrap. Always carry the three factors, never just the product.
Second, planned production time is a definition, not a fact. Whether planned maintenance, breaks and unsold capacity sit inside or outside the denominator changes the result by tens of points, and a plant that quietly moves the line every quarter is measuring its own reporting, not its equipment. Fix the definition once, write it down, and compare the machine only to itself.
The eight pillars of TPM
TPM is conventionally drawn as eight pillars standing on a foundation of 5S — sort, set in order, shine, standardise, sustain. The foundation is not decorative: several pillars depend on equipment being clean and accessible enough that an abnormality is visible at all.
1.Autonomous maintenance (Jishu Hozen)
Operators take on routine cleaning, lubrication, bolt-tightening and inspection of their own equipment. The purpose is not to save technician hours — it is that the person standing at the machine all day is the first to notice a change in it, and cleaning is how abnormalities get found.
2.Focused improvement (Kobetsu Kaizen)
Small cross-functional teams attack one measured loss at a time — a recurring jam, a long changeover, a chronic quality defect — and do not move on until the loss is quantified before and after.
3.Planned maintenance (Keikaku Hozen)
The scheduled, technician-led work: time-based and condition-based servicing, planned around production rather than squeezed between breakdowns. This is the pillar that maps most directly onto classical preventive maintenance.
4.Quality maintenance (Hinshitsu Hozen)
Equipment conditions are tied to product defects, so maintenance targets the machine parameters that actually produce scrap instead of treating quality as a downstream inspection problem.
5.Early equipment management (Shoki Kanri)
What the shop floor learned about the current machines feeds into the specification of the next ones — maintainability, access, standard parts — so a new line does not repeat the failure modes of the old one.
6.Education and training (Kyoiku Kunren)
Operators are trained to detect abnormalities and technicians to diagnose root causes. Autonomous maintenance without training is just a reassignment of chores.
7.Office TPM (Kanri-Bumon TPM)
The same loss-elimination discipline applied to administrative processes that feed the plant — procurement lead times, scheduling, spare parts ordering, work order paperwork.
8.Safety, health and environment (Anzen, Eisei, Kankyo)
A target of zero accidents and zero environmental incidents, treated as a pillar rather than a constraint. Equipment that is clean, accessible and predictable is also equipment that is safer to work on.
TPM vs classical preventive maintenance
The two are often presented as alternatives. They are not: planned maintenance is the third pillar of TPM, so a plant doing TPM is doing preventive maintenance inside it. The difference is scope and ownership.
| Dimension | TPM | Preventive maintenance |
|---|---|---|
| Scope | Whole-plant programme: operations, quality, engineering, safety, admin | The maintenance function |
| Who does the work | Operators and technicians share it | Technicians |
| Goal | Eliminate the six big losses; maximise OEE | Prevent failure through scheduled servicing |
| Headline metric | OEE (availability x performance x quality) | Schedule compliance, MTBF, PM completion rate |
| Trigger for work | Measured loss, plus schedule | Calendar interval or meter reading |
| Time to results | Multi-year cultural programme | Weeks — it is a schedule |
The practical consequence: you can start preventive maintenance next week, and you cannot start TPM next week. TPM changes who is accountable for equipment, which is a cultural change that fails when it is announced rather than built. Almost every successful programme starts with one line, one pillar, and a number.
How to run TPM without a budget
TPM has a reputation as an expensive programme — consultants, certification, an OEE system wired into the PLCs. None of that is required to start, and starting is the part most plants never get to. A realistic first six months, with no capital spend:
- Pick one line, not the plant. Choose the constraint — the equipment whose downtime actually costs you output. A pilot that succeeds visibly is worth more than a plant-wide launch that stalls.
- Build the asset register first. You cannot attribute a loss to a machine that is not in a list. Equipment, location, criticality. A spreadsheet import is fine — this takes an afternoon, not a project.
- Record stops by hand, honestly. Duration and cause, every stop, for four weeks. Manual logging is imperfect and it is still enough to rank your six big losses. Automated OEE capture can come later, once you know which machine deserves the sensor.
- Start autonomous maintenance with cleaning. A short daily operator checklist — clean, inspect, lubricate, report anything odd. Nakajima's point was that cleaning is inspection. Give operators a way to raise a fault in seconds, or they will not raise it.
- Schedule the planned maintenance you already know about. Manufacturer intervals, plus whatever your technicians already carry in their heads. Getting it out of heads and onto a recurring schedule is the single highest-return step here.
- Run one focused-improvement cycle. Take the top loss from step 3, form a small team including the operators, fix it, and publish the before-and-after number. That number is what buys you the second cycle.
Steps 2 to 6 are exactly what a CMMS is for: the asset register, the fault reports, the recurring schedule and the history you measure the improvement against. FreeMaint's Core tier is free with unlimited users — which matters for TPM specifically, because pillar 1 only works if every operator can be in the system. On per-user pricing, autonomous maintenance is the first thing cut. Operators can report a fault by scanning a QR code on the machine without an account at all.
Put the first two pillars in place this week
Asset register, operator fault reports and recurring schedules — unlimited users, $0 forever, no credit card.
Create your free account