Guide: TPM in manufacturing
Total productive maintenance: a working guide for plant teams
Total productive maintenance (TPM) is a way of running a plant in which operators, maintenance and management all share responsibility for keeping equipment in good condition and removing the losses that stop it producing good parts. It helps production heads, maintenance managers and operators who want fewer breakdowns, shorter changeovers and steadier quality.
Where TPM came from
The Japan Institute of Plant Maintenance (JIPM) describes TPM as overall maintenance for manufacturing: a method that improves productivity by working toward zero losses and strengthening the foundations of production. JIPM's own history records that the concept of plant maintenance with total participation was proclaimed in 1971, and that TPM was first implemented at Nippon Electrical Equipments (now DENSO). JIPM revised its definition of TPM in 1989.
Seiichi Nakajima wrote the first English book on the subject, Introduction to TPM, published by Productivity Press in 1988. That book presented TPM to managers as a multi-year program, and it set out ideas that still anchor TPM training today: operator-led maintenance, the six big losses and overall equipment effectiveness (OEE).
The word that matters most is total. In a traditional plant, operators run machines and maintenance fixes them. TPM in manufacturing breaks that split. Operators take care of basic conditions on their own machines, maintenance moves its effort toward planned and improvement work, and engineering designs new equipment so it is easier to keep in good condition.
The eight pillars of total productive maintenance
JIPM's TPM Excellence Awards assess plants on activity across eight pillars. The names vary a little between books and consultants; these follow JIPM's assessment checklist.
1. Focused improvement
Small cross-functional teams attack specific, measured losses on a machine or line. JIPM's checklist looks for clear loss definitions and tools such as loss maps, why-why analysis and PM analysis.
2. Autonomous maintenance
Operators clean, inspect, lubricate and tighten their own equipment to standards they helped write, and spot abnormalities early. It is built in steps (see below).
3. Planned maintenance
The maintenance department runs a structured program of preventive, condition-based and corrective work, with clear division of duties between operators and maintenance.
4. Quality maintenance
Set and keep the equipment conditions that make defects impossible, and check those conditions instead of only inspecting the parts afterwards.
5. Early equipment management
Feed what you found from running existing machines into the design and purchase of new ones, so they start up faster and are easier to maintain.
6. Education and training
Build operator and maintainer skills on purpose, with a training system, one-point lessons and skill certification.
7. Administrative and indirect departments
Apply the same loss thinking to planning, purchasing, stores and office work that supports production.
8. Safety, health and environment
Aim for zero accidents and zero pollution. Many plants find that a clean, well-inspected machine is also a safer one.
Autonomous maintenance, step by step
Autonomous maintenance is the pillar most people picture when they hear TPM, and the one most often done badly. It is not handing operators a cleaning rag and a checklist. It is a staged program in which operators learn their machine well enough to keep its basic conditions themselves. JIPM's award criteria refer to it as a step-by-step program; a site applying for its first TPM Excellence award is expected to have completed step 3 and be active in step 4.
The seven steps commonly taught are:
- Initial cleaning. Clean the machine thoroughly and, while cleaning, find abnormalities: leaks, loose bolts, worn parts, missing guards. Tag each one.
- Countermeasures for sources of contamination and hard-to-reach areas. Stop the dirt, swarf or oil at its source, and make awkward points easier to clean and inspect.
- Tentative cleaning, inspection and lubrication standards. Operators write down what to check, how, how often and in how many minutes.
- General inspection. Operators learn how the machine's subsystems work (hydraulics, pneumatics, drives, electrics) so they can inspect them properly.
- Autonomous inspection. Operator and maintenance checklists are merged and tidied, so nothing is checked twice and nothing is missed.
- Standardization. Workplace organization, material flow and standards are extended beyond the machine itself.
- Full autonomous management. Teams run their own improvement against goals, with maintenance as a partner rather than a fire service.
The six big losses and how they connect to OEE
Nakajima grouped the main equipment losses into six, and linked each group to one factor of OEE. If you measure OEE without knowing which loss is behind the number, you have a score but no plan.
| OEE factor | Loss | What it looks like on the floor |
|---|---|---|
| Availability | Equipment failure (breakdowns) | A gearbox seizes; the line is down for three hours. |
| Availability | Setup and adjustment | Die change, product changeover, trial shots until the first good part. |
| Performance | Idling and minor stoppages | A sensor trips, a part jams, the operator clears it in a minute. Rarely logged, often large in total. |
| Performance | Reduced speed | The machine runs slower than its ideal cycle because of wear, habit or fear of defects. |
| Quality | Process defects | Scrap and rework during steady running. |
| Quality | Reduced yield (start-up losses) | Scrap from warm-up and stabilization after a start or changeover. |
Some plants count planned stops (meals, meetings, planned maintenance) as outside planned production time, others count them as losses. Pick one rule, write it down, and use it everywhere, or OEE comparisons between lines mean nothing.
The OEE formula
OEE turns the six losses into one percentage. Each factor is a ratio between 0 and 1.
OEE = Availability x Performance x Quality- Availability
- Run time / planned production time. Breakdowns and setups reduce it.
- Performance
- (Ideal cycle time x total count) / run time. Minor stops and slow running reduce it.
- Quality
- Good count / total count. Defects and start-up scrap reduce it.
A useful cross-check: OEE also equals (good count x ideal cycle time) / planned production time. If the two methods disagree, one of your inputs is wrong. Try the numbers yourself with the OEE calculator.
OEE for one shift
Worked example (illustrative)
An illustrative press line has a planned production time of 480 minutes in a shift. It loses 45 minutes to a breakdown and 15 minutes to a changeover, so it runs for 420 minutes.
The ideal cycle time is 1 minute per part. The line produces 380 parts, of which 361 are good.
| Factor | Arithmetic | Result |
|---|---|---|
| Availability | 420 / 480 | 87.5% |
| Performance | (1 x 380) / 420 | 90.5% |
| Quality | 361 / 380 | 95.0% |
| OEE | 0.875 x 0.905 x 0.95 | 75.2% |
| Cross-check | (361 x 1) / 480 | 75.2% |
The biggest single loss in this shift is the 60 minutes of breakdown and changeover. A focused improvement team would start there, and the planned maintenance pillar would ask why the breakdown happened and what check would have caught it.
How to start TPM without a big launch
Plants that succeed with TPM usually start narrow and deep, not broad and shallow.
- 01
Get the plant head to own it
TPM changes who does what. Without visible support from the person who runs production, operators will not be released for cleaning and training, and maintenance will not be released for improvement work.
- 02
Pick one model machine
Choose a machine that matters and that people complain about. Run autonomous maintenance steps 1 to 3 on it with one operator team and one maintenance partner.
- 03
Measure the losses before you fix them
Log breakdowns, setups and minor stops on that machine for a few weeks. Without a baseline, you cannot show that anything improved.
- 04
Tag and close abnormalities
Every tag raised during initial cleaning becomes a tracked job with an owner and a date. Tags that sit open for months kill operator enthusiasm faster than anything else.
- 05
Write the first standards with operators
Cleaning, inspection and lubrication standards with photos, written in the operators' own language and timed so they fit in the shift.
- 06
Review weekly, then extend
A short weekly review on the activity board: losses, open tags, standards followed. When the model machine holds its gains, copy the approach to the next machine.
Why TPM programs stall
TPM has a long track record, and so do failed TPM launches. The failure modes are familiar to anyone who has been through one.
- Treating it as a cleaning campaign. Step 1 gets done for an audit visit, the machine shines for a week, and nothing changes in how it is inspected or maintained.
- Launching all pillars at once. Eight pillar committees, eight sets of meetings, and no one with time to do the work.
- Tags without follow-through. Operators find problems, maintenance never closes them, and operators stop looking.
- No loss data. Without breakdown, setup and minor stop records, improvement is guesswork and progress cannot be shown.
- OEE as a target, not a diagnosis. When OEE becomes a number to hit, people adjust the definition of planned time instead of removing losses.
- Maintenance feels replaced. If maintenance staff believe autonomous maintenance means fewer maintenance jobs, they will resist. Make it clear their work moves toward planned and improvement work.
- Paper that nobody reads. Checklists filled in at the end of the shift from memory give a false picture of machine condition.
Where MaintenanceIQ fits
MaintenanceIQ is an industrial CMMS, not a TPM program. It can carry the records several pillars depend on. Planned maintenance runs as recurring schedules by calendar, by usage meter, or whichever comes first. Operator and maintenance checklists can be typed (checkbox, number, photo) and are frozen once completed. Breakdowns are logged with start and end time, which gives MTBF, MTTR and the availability side of OEE.
OEE in MaintenanceIQ is Availability x Performance x Quality for each asset: availability is derived from logged breakdowns, and performance and quality come from shift logs picked up from your machine sensors, or entered by the supervisor where a machine has no sensor. That keeps the losses visible by machine without a separate MES project.
Sources
- What is TPM? (opens in a new tab), Japan Institute of Plant Maintenance (JIPM)
- History of JIPM (opens in a new tab), Japan Institute of Plant Maintenance (JIPM)
- 2026 TPM Excellence Awards Outline (opens in a new tab), Japan Institute of Plant Maintenance (JIPM)
- Introduction to TPM: Total Productive Maintenance, Seiichi Nakajima (Productivity Press, 1988) (opens in a new tab), WorldCat catalog record
Frequently asked questions
What are the eight pillars of TPM?
Focused improvement, autonomous maintenance, planned maintenance, quality maintenance, early equipment management, education and training, administrative and indirect departments, and safety, health and environment. JIPM assesses plants against these eight in its TPM Excellence Awards.
Who created total productive maintenance?
JIPM records that the concept was proclaimed in Japan in 1971 and first implemented at Nippon Electrical Equipments, now DENSO. Seiichi Nakajima wrote the first English book on it, Introduction to TPM, published in 1988.
What is the difference between TPM and preventive maintenance?
Preventive maintenance is one activity: servicing equipment on a schedule before it fails. TPM is a whole-plant approach that includes preventive maintenance inside its planned maintenance pillar, and adds operator care, focused improvement, quality maintenance, training and design input.
How long does it take to implement TPM?
There is no fixed timeline. JIPM's first-level award expects at least two years of activity after TPM is introduced. Most plants see early results on a model machine much sooner, if they measure losses before and after.
Is OEE the same as TPM?
No. OEE is a measure that shows how much of planned time produces good parts. TPM is the program that removes the losses OEE reveals.
Keep reading
OEE calculator
Work out availability, performance, quality and OEE for a shift, with each term explained.
ReadPreventive maintenance software
How to run the planned maintenance pillar on calendar and usage-based schedules.
ReadWhat is CMMS?
The full form, the core modules and how to roll a maintenance management system out one line at a time.
Read