The Promise That Sold You
Total Productive Maintenance arrived in your plant wrapped in the
authority of Japanese manufacturing excellence. Seiichi Nakajima
developed it at Toyota’s supplier base in the 1960s and 1970s, and by
the time the rest of the world caught on, the results were compelling:
equipment effectiveness climbing above 85%, breakdown rates dropping by
90%, and operators who actually understood the machines they ran. The
promise was straightforward — instead of waiting for machines to break
and then scrambling to fix them, you would maintain them proactively,
involve the people who operated them every day, and systematically
eliminate the six big losses that robbed your production capacity.
The eight pillars of TPM sounded rigorous and complete. Autonomous
maintenance would put basic care in operators’ hands. Planned
maintenance would schedule interventions based on actual equipment
condition, not arbitrary calendar dates. Focused improvement would
target chronic losses with structured problem-solving. Early equipment
management would design maintainability into new installations. Quality
maintenance would establish the zero-defect relationship between machine
condition and product output. Training and education would build the
skills your people needed. Safety, health, and environment would be
integrated, not bolted on. And office TPM would extend the philosophy
beyond the shop floor.
It was a comprehensive system. And like most comprehensive systems
that arrive with a consulting contract and a three-day kickoff workshop,
your organization managed to hollow it out within six months while
keeping all the visual indicators of compliance.
The Six Big Losses
Nobody Actually Tracks
The theoretical foundation of TPM rests on identifying and
eliminating six categories of equipment loss: breakdowns, setup and
adjustment time, idling and minor stops, reduced speed, process defects,
and reduced yield during startup. Together, these losses determine your
Overall Equipment Effectiveness, and the premise of TPM is that most
plants are operating at 40-60% OEE without knowing it because they have
never measured these losses individually.
This is where most TPM implementations fail first. Your plant
probably installed OEE monitoring dashboards during the kickoff phase.
They look impressive on the conference room wall. But when you ask the
maintenance manager what percentage of minor stops are being captured,
the answer is usually silence — because minor stops, the
three-to-five-second interruptions that happen dozens of times per
shift, are almost never logged. The operator hits the reset button and
keeps going. The shift record shows full production. The OEE number on
the wall is calculated from whatever data was easiest to collect, which
means it is wrong, which means every improvement target derived from it
is arbitrary.
The six losses framework requires granular, accurate data capture at
the machine level. Most organizations are unwilling to invest in the
automated monitoring or the operator discipline needed to get this data.
Instead, they estimate. They round. They exclude shifts where “things
were unusual.” And within a few months, the OEE number on the dashboard
becomes a negotiated figure that everyone agrees on rather than a
measured fact that drives action.
If you cannot measure the six losses accurately, you cannot
prioritize them. If you cannot prioritize them, your TPM activities are
guesswork. And if your TPM activities are guesswork, you are not doing
TPM — you are doing maintenance theater with a Japanese name.
Autonomous
Maintenance: The Custodial Inversion
Autonomous maintenance is the pillar most organizations implement
first, because it seems simple and it produces immediate visible
results. The idea is that operators perform basic maintenance tasks —
cleaning, lubrication, tightening, inspection — and in doing so, they
develop an intimate understanding of their equipment’s condition. They
notice the unusual vibration, the slight temperature change, the warning
signs that precede a breakdown. They become the first line of
defense.
What actually happens in most plants is that autonomous maintenance
becomes a cleaning schedule. Operators are given checklists with boxes
to initial. They wipe down their machines at the start of shift. They
initial the boxes. They go back to running production. The checklists
are collected by a supervisor who files them in a binder. Nobody reviews
the data. Nobody analyzes trends. Nobody uses the inspection findings to
plan maintenance interventions.
The fundamental misunderstanding is treating autonomous maintenance
as a task to be completed rather than a skill to be developed. The
cleaning is not the point. The cleaning is the context in which
operators learn to see abnormalities. When your autonomous maintenance
program measures compliance — did the operator initial the box? —
instead of learning — can the operator identify an abnormality that a
trained maintenance technician would catch? — you have built a paperwork
system, not a capability.
I have visited plants where autonomous maintenance checklists had
been running for three years and not a single operator could explain
what an abnormal lubrication condition looked like on their machine.
They could tell you where the lubrication points were. They could tell
you how often they were supposed to check them. But they had never been
taught what they were looking for, what constituted a warning sign, and
what to do when they found one. The checklist asked: “Lubrication OK?”
The operator checked “Yes.” Every day. For three years. Until the
bearing failed catastrophically and shut down the line for eleven
hours.
Planned Maintenance: The
Calendar Trap
The second pillar, planned maintenance, is where organizations
discover that scheduling maintenance by time interval is expensive and
often counterproductive. The TPM approach advocates condition-based
maintenance — monitoring actual equipment health and intervening when
the data says intervention is needed, not when the calendar says it is
due.
But condition-based maintenance requires sensors, data collection,
analysis capability, and the organizational discipline to act on what
the data reveals. It is significantly harder than calendar-based
maintenance, which only requires a spreadsheet and the discipline to
shut down the machine on schedule.
So most organizations default to calendar-based maintenance and call
it planned maintenance. They create preventive maintenance schedules
based on manufacturer recommendations — the same generic recommendations
written for every installation regardless of operating conditions,
environment, load, or duty cycle. They follow these schedules
faithfully. And they discover two things: first, that they are
performing maintenance on equipment that does not need it yet, which is
wasteful; and second, that they are not performing maintenance on
equipment that needs it sooner than the schedule predicts, which is
catastrophic.
The worst version of this trap is the plant that converts from
reactive maintenance to calendar-based preventive maintenance and then
declares victory. “We’ve implemented planned maintenance!” the manager
announces. The maintenance budget doubles. Equipment availability does
not improve. Unplanned breakdowns continue at roughly the same rate
because the calendar schedule addresses the wrong failures at the wrong
time. And the organization concludes that TPM does not work, when what
does not work is their specific implementation of it.
Condition-based maintenance is not optional. It is the core of the
planned maintenance pillar. If you are not monitoring equipment
condition and using that data to drive maintenance decisions, you are
not doing planned maintenance in the TPM sense. You are doing scheduled
maintenance, which is a different and inferior approach.
Kobetsu Kaizen: The
Improvement That Isn’t
Focused improvement, or kobetsu kaizen, is the pillar that targets
chronic losses — the persistent, low-level problems that everyone has
learned to live with because they seem too small to justify a major
investigation and too entrenched to eliminate with a quick fix.
The TPM methodology for focused improvement is structured: identify
the loss, analyze it, develop a countermeasure, implement it, confirm
the result, and standardize the new condition. It is essentially the
scientific method applied to equipment losses. And it works — when it is
actually applied.
What usually happens instead is that focused improvement becomes a
suggestion scheme. Operators submit ideas. A committee reviews them
monthly. The easy ones get implemented. The hard ones get filed. The
chronic losses — the ones that truly limit equipment effectiveness — are
rarely addressed by operator suggestions because they are chronic
precisely because they have complex, systemic root causes that cannot be
solved with a five-minute idea.
The best focused improvement teams I have worked with were
cross-functional, had dedicated time away from production, and were
given authority to make changes without seeking permission through
multiple layers of management. This is expensive. It means taking people
off the line, which means temporarily reducing output, which means the
plant manager has to be willing to invest in improvement at the expense
of short-term production targets. Most plant managers are not willing to
do this. The focused improvement meetings happen during lunch breaks.
The teams never get traction on the real losses. And the suggestion box
fills up with ideas for better lighting in the break room.
The OEE Obsession
Perhaps the most damaging distortion of TPM is the obsession with OEE
as a single metric. OEE is the product of three ratios: availability
multiplied by performance multiplied by quality. It is a useful
composite indicator when calculated correctly from accurate data. It is
a dangerous distortion when used as a target.
Goodhart’s Law — when a measure becomes a target, it ceases to be a
good measure — applies to OEE with devastating precision. When OEE
becomes the metric on which a plant manager’s bonus depends, the three
components start getting optimized in the easiest ways possible, none of
which involve actual improvement.
Availability can be improved by reclassifying planned downtime as
“scheduled” rather than “lost.” Performance can be improved by running
equipment at a slower, more conservative speed — the opposite of what
TPM intends. Quality can be improved by redefining what constitutes a
defect, or by inspectiting less rigorously, or by reworking product
before it gets counted as a reject. All of these actions improve the OEE
number without improving the actual effectiveness of the equipment.
I have seen plants with OEE dashboards showing 85% — world-class by
most standards — where the shop floor was producing scrap at a rate that
would have horrified the customer, but the scrap was being reworked
internally and reclassified as good output. The number was beautiful.
The reality was not.
OEE is a diagnostic tool, not a performance target. The moment you
tie rewards to it, you create the incentive to manipulate it. The
organizations that benefit most from TPM are the ones that use OEE to
identify where to investigate, not to evaluate whether someone is doing
a good job.
The Training Gap
The training and education pillar of TPM is usually the first to be
cut when budgets get tight, which is ironic because it is the pillar
that determines whether every other pillar works. Autonomous maintenance
requires operators who understand their machines. Planned maintenance
requires technicians who can interpret condition data. Focused
improvement requires teams who can analyze problems systematically. None
of these capabilities exist without deliberate, sustained investment in
training.
Most organizations approach TPM training as a one-time event. They
bring in a consultant for the kickoff. Everyone attends a two-day
workshop. They receive a binder. The binder goes on a shelf. The
consultant leaves. And the organization is left with people who know the
vocabulary of TPM but lack the skills to execute it.
Real TPM training is ongoing. It involves structured skill
development, mentoring from experienced practitioners, hands-on practice
with real equipment, and assessment to confirm that learning has
actually occurred. It takes months, not days. It costs money. And it is
the single most important investment in the entire program, because TPM
is fundamentally a human capability system. Without capable people, the
eight pillars are just words on a poster.
The Culture That TPM
Requires
The deepest failure mode of TPM implementation is attempting to
deploy it in an organizational culture that cannot support it. TPM
requires a culture of transparency, where people are willing to report
problems without fear of blame. It requires a culture of learning, where
mistakes are treated as opportunities to improve rather than reasons to
punish. It requires a culture of ownership, where operators feel
responsible for the condition of their equipment and are empowered to
act on what they observe.
Most manufacturing organizations do not have this culture. They have
a culture of production targets, where the overriding priority is to hit
the number, and anything that interrupts production — including
maintenance, inspection, and improvement activities — is seen as an
obstacle. In this environment, TPM becomes a burden that production
people tolerate during audits and ignore during normal operations.
Changing organizational culture is harder than implementing any
methodology. It requires consistent leadership behavior over years, not
months. It requires leaders who model the behavior they expect — who go
to the gemba, who ask questions instead of assigning blame, who invest
in prevention even when it means accepting short-term production losses.
Most organizations are not willing to sustain this level of leadership
commitment. They want the results of TPM without the cultural
transformation it requires.
What TPM Done Right
Actually Looks Like
The plants where I have seen TPM produce genuine, sustained
improvement share several characteristics. First, they measure
accurately. They invest in automated data collection for the six losses.
They do not estimate, round, or negotiate OEE numbers. The data is the
data, and it drives every decision.
Second, they develop people before they deploy tools. Operators are
trained in inspection technique, not just inspection procedure.
Maintenance technicians are trained in analysis, not just repair.
Engineers are trained in loss elimination methodology, not just
problem-solving frameworks.
Third, they protect time for improvement. Focused improvement teams
get dedicated time, not lunch breaks. Autonomous maintenance gets
supervised practice time, not a checklist handed to an operator between
production orders. Planned maintenance gets analysis time, not just
execution time.
Fourth, they use OEE as a compass, not a scorecard. The number tells
them where to look. It does not tell them whether someone is performing
well. Individual and team performance is evaluated on capability
development and problem-solving effectiveness, not on whether a metric
moved in the right direction this month.
Fifth, leadership is engaged and patient. The plant manager visits
the shop floor regularly, not for audits, but to learn. The maintenance
manager participates in improvement activities, not just reviews them.
And senior leadership accepts that TPM is a multi-year transformation,
not a quarterly initiative that should show ROI by the end of Q2.
The Diagnostic Question
If you are running a TPM program and you want to know whether it is
real or theater, ask one question: when was the last time an operator
stopped production because they noticed an abnormality during autonomous
maintenance, and what happened to that operator afterward?
If the answer is “I can’t remember” or “that never happens,” your
autonomous maintenance is not autonomous, your operators are not
inspecting, and your TPM program is decorative. If the answer is “they
got praised, maintenance was called, and the potential breakdown was
averted,” you might actually be doing TPM.
If the answer is “they got yelled at for stopping the line,” you have
your answer about your culture, too.
Peter Stasko is a Quality Architect with over 25
years of experience in manufacturing quality systems. He has implemented
and assessed TPM programs across automotive, aerospace, and heavy
industry sectors on three continents. He believes that most TPM failures
are not methodology failures — they are leadership failures disguised as
implementation challenges.
