Total Productive Maintenance (TPM): Pillars and a First 90 Days

By Mirko Peters · 2026-11-11

Total productive maintenance gets operators and maintenance to care for equipment together. Learn the eight pillars and a 90-day plan for one pilot machine.

Total productive maintenance (TPM) is a way of running equipment in which operators and maintenance staff keep machines in good condition together. The aim is no breakdowns, no defects and no accidents. This guide is for production managers who have a line with too many stops and want a plan they can start on Monday. You will see what the eight pillars are, what autonomous maintenance looks like, and how to structure the first 90 days on one pilot machine.

Key points

• Total productive maintenance (TPM) has operators and maintenance staff care for equipment together, with the aim of no breakdowns, no defects and no accidents.
• The word total stands for total participation, total equipment productivity and total life cycle, not for a maintenance programme owned by one department.
• The eight pillars vary by author, and the two to start with are autonomous maintenance and focused improvement on the biggest loss.
• A first 90 days works best on one pilot machine: restore it, measure its losses for at least two weeks, then fix the biggest loss.
• Measure OEE and log every stop with a reason, including small stops, or you will not know whether TPM is working.

Total productive maintenance: what "total" covers

The Lean Enterprise Institute (LEI) explains that TPM is a set of techniques first pioneered by Denso in the Toyota Group, to keep every machine able to do its required work. It uses "total" in three senses:

• Total participation: maintenance staff, line managers, engineers, quality people and operators all take part.
• Total productivity of equipment: the focus is on the six major losses, which are downtime, changeover time, minor stops, speed losses, scrap and rework.
• Total life cycle: maintenance practice changes with the age and stage of the equipment.

The difference from traditional preventive maintenance is who does the work. In classic preventive maintenance, the maintenance department does it all. In TPM, operators take part in routine care, improvement projects and simple repairs, and they take real ownership of their equipment. Daily cleaning, lubricating, tightening and inspecting are typical examples.

TPM is tied closely to OEE, the measure that shows how much of your planned time produces good parts at full speed. The six big losses map directly onto OEE: equipment failure and setup affect availability, idling and minor stops and reduced speed affect performance, and defects and start-up rejects affect quality. See the six big losses explained by OEE.com.

The benefit is practical. TPM moves a plant from reactive maintenance, where people fix a machine when it breaks, towards proactive work, where they find wear before it stops production. The result is fewer unplanned stops, more uptime and steadier quality. It does not remove the need for planned and predictive maintenance, such as condition monitoring. It gives those methods a stable base: a clean machine with known inspection points.

TPM pillars: the eight pillars and where to start

The All About Lean overview notes that the Japan Institute of Plant Maintenance (JIPM) expanded the original seven pillars to eight, and that this version is the most common in English texts. It also points out that there is no fixed canon, and that authors differ in names, count and content. Here is the common list, with a plain description.

1. Autonomous maintenance (Jishu Hozen): operators look after their own equipment.
2. Focused improvement (Kobetsu Kaizen): small project teams remove the biggest losses, one at a time.
3. Planned maintenance (Keikaku Hozen): maintenance work is scheduled and planned instead of reactive.
4. Quality maintenance (Hinshitsu Hozen): keeping equipment conditions that protect product quality.
5. Early equipment management (Shoki Kanri): learning from today's machines when designing and starting up new ones.
6. Training and education: building the skills that the other pillars need.
7. Administrative and office TPM: applying the same thinking to support departments.
8. Safety, health and environment: the target of no accidents.

The All About Lean overview of the eight pillars of TPM lists these with their Japanese names. For a first project you do not need all eight. Pick autonomous maintenance to change daily habits, and focused improvement to attack the largest measured loss. Add planned maintenance once you know which failures repeat.

Autonomous maintenance in practice

Autonomous maintenance is the pillar that people notice first, because it changes what operators do each shift. An overview from Reliable Plant lists the typical steps:

1. Clean the machine to a baseline standard that operators then keep.
2. Train operators in routine inspection based on the machine's manual.
3. Document inspection points and wear-prone parts, and label set points.
4. Identify lubrication points and schedule lubrication during changeovers or planned downtime.
5. Standardise cleaning, inspection and lubrication, and create a checklist for operator tasks.
6. Train operators to report emerging problems to their supervisor, and audit the checklist at decreasing intervals.

The first cleaning matters more than it looks. When an operator wipes down a machine, leaks, cracks, loose bolts and missing guards become visible. Tag each find and track it to closure. The same source says TPM builds on the 5S system of organising, tidying, cleaning, standardising and sustaining. In other words, 5S is the foundation: a machine that is not clean and orderly cannot be inspected by its operator. Lubricating, tightening and inspecting then become short, standard tasks inside the shift.

TPM implementation: a first 90 days on one pilot machine

Reliable Plant describes five implementation phases: pick a pilot area, restore the equipment, measure OEE, address the major losses, and implement planned maintenance. The plan below turns that into three blocks of about 30 days. The split into days is our own planning aid, not a standard.

Days 1 to 30: choose and restore.

• Choose one pilot machine. Beginners are advised to start with equipment that is easy to improve. A bottleneck or the most problematic machine is also an option, but harder.
• Form a small team of an operator from each shift, a maintenance technician and a supervisor.
• Clean and inspect the machine thoroughly, fix the defects you find and start the first operator checklist.

Days 31 to 60: measure.

• Track every stop on the machine for at least two weeks. Give each stop a reason code, and include an "unknown" code so that nothing is lost.
• Include micro-stops. Short stops often go unrecorded but add up to large losses.
• Calculate OEE for the pilot machine and rank the losses.

Days 61 to 90: fix the biggest loss.

• Use a cross-functional team with root cause analysis, such as 5 Whys and a fishbone diagram, to attack the biggest loss. Then verify that it is gone.
• Set maintenance intervals based on run time rather than calendar time where the data supports it.
• Review the pilot at day 90: what did the OEE do, what did operators say, what would you change before the second machine?

Example: what the first measurement can show

This is an illustrative calculation, not a measured result. A pilot machine is planned to run 480 minutes per shift. In the first two weeks of tracking, the team finds an average of 50 minutes of breakdowns and 30 minutes of changeover per shift. That leaves 400 minutes of run time, so availability is 400 divided by 480, which is 83 per cent.

Operators also report many short jams. The team counts 40 jams per shift, about 1 minute each, so 40 minutes are lost without anyone writing a reason. During the remaining run time the machine also runs slower than its rated speed. Assume this costs another 20 minutes of full-speed time.

In this example, the biggest single loss is not the breakdowns. It is the 40 minutes of jams, plus the reduced speed. Without measurement, the team would have started with the breakdowns. With it, focused improvement goes after the jams first. The numbers are made up, but the lesson is real: loss tracking changes the order of work.

Common mistakes

• Treating TPM as a maintenance programme. If the maintenance department owns it alone, operators stay passive and the "total" is lost.
• Handing over tasks without training. Operators need clear standards, time in the shift and a way to call for help.
• Starting everywhere at once. A pilot on one machine teaches you more than ten half-started lines.
• Measuring only big breakdowns. Slow cycles and small stops are often larger sources of lost time, so record them too.
• No time for cleaning and checks. If the schedule leaves no minutes for the checklist, the checklist dies in a month.
• Declaring victory at day 90. The pilot shows the method. The habit takes longer to build.

How software can help

A simple spreadsheet works for the first pilot. After that, software can help by capturing stops automatically from the machine signal, letting operators add the reason at the line, and ranking losses in a chart. A computerised maintenance management system (CMMS) keeps the work orders, spare parts and intervals in one place, and shows which failures repeat. Whatever tool you use, keep the reason list short and the data visible on the shop floor.

Next step

Choose one pilot machine this week and agree a date for the first joint cleaning with its operators and a maintenance technician. That is how total productive maintenance starts: one machine, one team, and a measured baseline.