The Promise Every
Plant Manager Believes
You read the case studies. Toyota achieved 90% OEE. Nippondenso
reduced breakdowns to near zero. The literature is seductive: implement
Total Productive Maintenance, and your machines will run like Swiss
watches, your operators will become empowered equipment owners, and your
maintenance costs will plummet.
So you launch the initiative. You hire a consultant for a three-day
kickoff. You form a TPM steering committee. You print posters. You
schedule autonomous maintenance training. You feel the energy.
Eighteen months later, the posters are yellowed, the steering
committee hasn’t met in six months, the autonomous maintenance has
devolved into operators wiping down machines with oily rags, and your
breakdown frequency is exactly what it was before you started — possibly
worse, because the preventive maintenance schedule you disrupted to make
room for TPM activities never got back on track.
What happened?
The answer is the same answer it always is in quality management: you
adopted the form and abandoned the substance. You implemented the tools
and skipped the philosophy. You did the activities and forgot the
purpose.
What
TPM Actually Is (Before We Discuss How You Broke It)
Total Productive Maintenance was developed in the 1960s and 1970s,
primarily at Nippondenso (a Toyota Group company), as a systematic
approach to maximizing equipment effectiveness. Seiichi Nakajima is
generally credited with formalizing its structure.
The core idea is radical in its simplicity: equipment
reliability is not solely the maintenance department’s
responsibility. It is a shared ownership model where operators,
maintenance technicians, and engineers all contribute to keeping
machines running at their theoretical best.
TPM rests on eight pillars (though variations exist):
- Autonomous Maintenance — operators perform
cleaning, lubrication, inspection, and minor repairs - Planned Maintenance — a scheduled, predictive, and
preventive maintenance system - Focused Improvement (Kaizen) — cross-functional
teams attack major equipment losses - Quality Maintenance — using equipment data to
ensure output quality, not just machine uptime - Early Equipment Management — designing
maintainability and reliability into new equipment from the start - Training and Skills Development — building operator
and technician competence systematically - Office TPM — applying TPM principles to
administrative and support processes - Safety, Health, and Environment — ensuring zero
accidents and zero environmental incidents
The goal is to pursue zero breakdowns, zero defects, and zero
accidents — not as aspirational poetry, but as operational
targets with measurable progress.
The key metric is Overall Equipment Effectiveness (OEE), which TPM
aims to maximize by systematically eliminating what Nakajima called the
“six big losses”:
- Equipment failure (breakdowns)
- Setup and adjustments
- Idling and minor stoppages
- Reduced speed
- Process defects
- Reduced yield (startup losses)
If this sounds comprehensive, that’s because it is. TPM is not a
quick-fix program. It is a fundamental reorganization of how a
manufacturing organization thinks about its physical assets.
Where
TPM Goes Wrong: The Eight Ways You Sabotaged Your Own Program
1.
Autonomous Maintenance Became “Operators Clean Machines”
This is the most common and most damaging reduction. Autonomous
maintenance (Jishu Hozen) is supposed to be a structured progression
where operators:
- Step 1: Clean the machine thoroughly, identifying abnormalities
- Step 2: Eliminate sources of contamination and hard-to-access
areas - Step 3: Establish cleaning and lubrication standards
- Step 4: Conduct general inspection training
- Step 5: Conduct autonomous inspection
- Step 6: Standardize and organize
- Step 7: Full autonomous management
What most plants actually do: they tell operators to clean their
machines at the start of each shift. Maybe they add a lubrication
checklist. Then they declare autonomous maintenance “implemented.”
Steps 2 through 7 never happen. The operator never learns to inspect
the machine for early wear signals. The contamination sources — the
hydraulic leak that keeps dripping, the dust ingress point that was
never sealed — remain unaddressed. The cleaning standard becomes a piece
of paper taped to the machine that nobody follows after the first
month.
The result: you get clean-looking machines that break down at the
same rate as dirty ones. You also get resentful operators who feel
they’re being asked to do janitorial work, and alienated maintenance
technicians who feel their expertise is being undermined.
2. Planned
Maintenance Was Never Actually Rebuilt
TPM requires you to transform your maintenance strategy from reactive
(fix it when it breaks) to a structured combination of preventive
(time-based), predictive (condition-based), and corrective (eliminate
root causes) maintenance.
Most plants already have a preventive maintenance schedule before
TPM. It’s usually mediocre — based on manufacturer recommendations,
rarely updated, and frequently skipped when production pressures mount.
When TPM launches, the existing PM schedule isn’t overhauled; it’s
simply relabeled. The same inadequate intervals, the same incomplete
task lists, the same tendency to defer when the line is behind
schedule.
Predictive maintenance — vibration analysis, oil analysis,
thermography, ultrasonic testing — is discussed in the kickoff
presentation and then quietly shelved when someone realizes it requires
investment in sensors, software, and training. The condition-monitoring
data that should feed your maintenance planning never materializes.
You cannot achieve TPM’s goals with a reactive maintenance culture
wearing a preventive label. The breakdowns will continue because the
underlying maintenance strategy never changed.
3.
Focused Improvement Teams Attacked the Wrong Losses
The focused improvement (Kobetsu Kaizen) pillar is where TPM delivers
its most visible short-term wins. Cross-functional teams analyze a
specific machine’s losses, identify the biggest contributors, and
systematically eliminate them.
The trap is in the loss analysis. Most plants don’t have accurate
data on their six big losses. Their OEE calculation (if they have one)
lumps everything into vague categories. So the improvement team attacks
what it can see — the obvious problems, the ones everyone already knows
about — while the actual dominant losses remain invisible.
A team spends three weeks optimizing a changeover that accounts for
2% of the total equipment loss while a chronic minor stoppage problem
that accounts for 18% goes unaddressed because nobody has the data to
prioritize correctly. The team celebrates a successful kaizen event. The
machine’s overall effectiveness barely moves.
Without rigorous loss accounting — measuring and categorizing every
minute of equipment time — focused improvement becomes a lottery.
Sometimes you pick the right target. Usually you don’t.
4.
Quality Maintenance Was Confused With Quality Control
Quality Maintenance (Hinshitsu Hozen) is the most intellectually
demanding pillar and the one most commonly skipped entirely. Its
premise: equipment conditions directly determine product
quality, so controlling equipment conditions is a more effective
strategy than inspecting product output.
This requires you to map every quality characteristic of your product
back to the specific equipment parameters that influence it. What
bearing wear tolerance affects dimensional accuracy? What hydraulic
pressure drift causes surface defects? What thermal variation in the
injection barrel causes sink marks?
This is hard, analytical, engineering work. It requires understanding
the physics of your process at a level most organizations never
develop.
What happens instead: plants conflate Quality Maintenance with their
existing SPC program. They already have control charts on product
characteristics, so they check the box. But SPC detects problems after
they occur; Quality Maintenance is supposed to prevent them by
controlling the equipment conditions that cause them in the first
place.
The difference is proactive versus reactive. Most TPM implementations
stay reactive — and then wonder why their defect rates don’t
improve.
5. Training Was a
One-Time Event, Not a System
The Training pillar is supposed to establish a systematic skills
development program: define the skills operators and technicians need,
assess current skill levels, close the gaps through structured OJT and
classroom training, and validate competence through practical
demonstrations.
What actually happens: a two-day TPM training session at launch.
Maybe a follow-up refresher six months later. No skills matrix. No
competence assessment. No structured progression from basic to advanced
equipment knowledge.
Your operators were never taught how their machines actually work.
They don’t understand the hydraulic circuit diagram. They can’t
interpret a vibration trend. They don’t know what a bearing sounds like
when it’s starting to fail. And because they don’t know these things,
they can’t perform autonomous maintenance at any meaningful level, they
can’t contribute to focused improvement, and they can’t catch the early
warning signs that prevent catastrophic breakdowns.
Training isn’t a pillar you implement and forget. It’s the foundation
every other pillar stands on. Skip it, and the whole structure is
hollow.
6. Early
Equipment Management Was Never on the Radar
This pillar addresses a problem most plants don’t even realize they
have: they buy equipment that is inherently difficult to
maintain, then spend the next fifteen years paying for that
mistake.
When a new machine is specified, the purchasing decision is typically
driven by the engineering team (who focus on process capability) and the
finance team (who focus on price). Maintainability is rarely a selection
criterion. Access points are difficult to reach. Lubrication points are
hidden behind guards that require special tools to remove. Critical
components require machine disassembly to inspect. Spare parts are
proprietary and expensive.
Early Equipment Management says: involve maintenance and operations
in the equipment specification process. Build maintainability
requirements into purchase specifications. Run reliability tests during
commissioning. Capture lessons learned and feed them into the next
purchase.
Almost no plant does this. So they keep buying machines that are
maintenance nightmares, and then they launch TPM programs to deal with
the consequences — which is like trying to fix a bad diet with
antacids.
7. You Measured
Activity Instead of Results
This is the metric problem that plagues every quality initiative, and
TPM is no exception. Plants track:
- Number of operators trained in autonomous maintenance
- Number of kaizen events conducted
- Number of TPM boards installed
- Number of one-point lessons created
- Percentage of machines with cleaning standards
These are activity metrics. They measure effort, not outcome. They
tell you whether people are doing TPM things, not whether those things
are producing results.
The metrics that matter:
- Breakdown frequency (how often does each machine fail?)
- Mean time between failures (how reliable is each machine?)
- Mean time to repair (how quickly do you recover?)
- OEE and its three components (availability, performance, quality
rate) - Maintenance cost as a percentage of equipment replacement value
- Percentage of maintenance that is planned versus unplanned
If these numbers aren’t moving in the right direction, your TPM
program is failing — regardless of how many kaizen events you conducted
or how many cleaning standards you posted.
8. Leadership
Treated TPM as a Department Program
TPM requires fundamental changes in organizational behavior:
operators taking ownership of equipment condition, maintenance shifting
from firefighting to prevention, engineers designing for reliability,
and managers prioritizing long-term capability over short-term
output.
These changes require sustained leadership commitment. Not a kickoff
speech. Not a steering committee that meets quarterly. Not a consultant
who visits monthly.
The plant manager needs to visit the gemba regularly, observe
autonomous maintenance in practice, ask operators what they’ve learned
about their equipment, review the OEE trends personally, and make
resource decisions that demonstrate TPM is a priority — even when
production pressures scream otherwise.
When leadership treats TPM as “the maintenance department’s program,”
that’s exactly what it becomes: a maintenance initiative that operators
don’t support, production doesn’t prioritize, and engineering doesn’t
contribute to. And a TPM program without cross-functional ownership is
just a maintenance budget with better branding.
The OEE Trap Within TPM
Since TPM uses OEE as its primary metric, it inherits all the
problems of OEE measurement — and TPM implementations make these
problems worse.
When OEE becomes the scorecard for TPM success, the incentive
structure distorts immediately. Teams learn to calculate OEE in ways
that make it look good:
- Minor stoppages under a threshold are excluded
- Changeover time is classified as “planned” and removed from
availability - Reduced speed is accepted as the machine’s “normal” operating rate
rather than a loss - Defect rates are measured at final inspection rather than at the
process step, masking equipment-related quality losses
A plant can report 75% OEE using generous assumptions while its true
OEE — calculated against theoretical ideal cycle time with every loss
category included — might be 45%. The TPM program looks successful on
the dashboard while the equipment losses it was supposed to eliminate
continue unabated.
This is not a theoretical concern. It is the lived reality in most
plants that claim to have implemented TPM. The numbers are comfortable.
The equipment is not.
What Genuine TPM Looks Like
To be clear: TPM is not a failed concept. When implemented with
discipline and sustained over years, it produces extraordinary results.
Plants that do it properly achieve:
- Breakdown rates 90% lower than industry average
- OEE improvements of 20-30 percentage points
- Maintenance cost reductions of 30-50%
- Defect rates that drop because equipment conditions are controlled,
not just monitored - Operators who genuinely understand their equipment and catch
problems before they become failures
But these results come from organizations that:
- Progress through autonomous maintenance steps
systematically — not stopping at cleaning, but building true
operator competence in equipment inspection and diagnosis - Invest in condition monitoring technology —
vibration, oil analysis, thermography — and use the data to drive
maintenance decisions - Base focused improvement on rigorous loss data —
not opinions or visibility, but measured and categorized loss
accounting - Connect equipment conditions to product quality —
mapping the causal chain from machine parameter to dimensional or visual
defect - Build skills systematically — with defined
competencies, validated assessments, and career progression tied to
equipment mastery - Involve maintenance in equipment specification —
and refuse to buy machines that are designed to fail - Track outcome metrics religiously — and report them
with the same rigor as financial results - Demonstrate leadership commitment daily — through
gemba presence, resource allocation, and personal accountability for TPM
results
The difference between these organizations and the ones that fail is
not budget, not consultant quality, not even starting condition. The
difference is depth of commitment to the philosophy behind the
tools.
The Real Question
If your TPM program has been running for two years and your breakdown
rate hasn’t dropped significantly, your maintenance cost as a percentage
of asset value hasn’t decreased, and your OEE — honestly calculated —
hasn’t improved, then you don’t have a TPM program. You have TPM-themed
activities.
The activities feel productive because they generate artifacts:
checklists, boards, training records, kaizen reports. But artifacts are
not outcomes. And until your organization develops the honesty to
distinguish between the two, you will keep running a TPM program that
produces everything except what TPM was designed to deliver: equipment
that runs at its theoretical best, operated by people who understand it,
maintained by a system that prevents failures instead of reacting to
them.
Stop measuring whether you’re doing TPM. Start measuring whether your
equipment is getting better. The answer to that question will tell you
everything you need to know about your program — and most of you won’t
like what it says.
About the Author: Peter Stasko is a Quality
Architect with over 25 years of experience transforming manufacturing
operations across automotive, electronics, and precision engineering
sectors. He has led TPM implementations — both successful and
instructively unsuccessful — on three continents, and writes about the
gap between quality theory and shop-floor reality because someone needs
to tell the truth.
