Every plant I visit has a preventive maintenance schedule. It is usually a spreadsheet or a CMMS report, printed in colour, pinned to the maintenance office wall. It shows every piece of equipment, every PM task, the frequency, and the last completion date. It looks impressive. And in about seven out of ten cases, it is fiction.

The schedule exists. The work does not happen. Or it happens late, partially, or with such minimal execution that it provides no preventive value whatsoever. And then the machine fails, the maintenance team works through the night, the production team loses a shift, and the quality team deals with the nonconforming parts produced in the window between PM deferral and catastrophic failure. This cycle repeats, and the PM schedule continues to claim that maintenance is under control.

Preventive maintenance, in most plants I have audited, is reactive maintenance with better marketing. Here is why.

The compliance illusion

When I audit a PM system, the first number I ask for is PM compliance — the percentage of scheduled PM tasks that were completed on time. Most plants report a number between 85% and 95%. This number is almost always wrong.

The problem is in the definition of "completed." In one automotive plant, PM compliance was reported at 91%. When I dug into the CMMS records, I found that a PM task was marked "completed" when the maintenance technician signed the work order. I then compared the sign-off timestamps to the maintenance team's actual working hours. Forty-three percent of the PM sign-offs were timestamped during periods when the maintenance technician was documented as working on a breakdown elsewhere in the plant. They could not have been in two places at once. The PM was signed off without being performed.

This is not a rare occurrence. It is the natural consequence of a system that prioritises compliance metrics over maintenance effectiveness. The maintenance technician is under pressure to show high PM compliance. They are also under pressure to respond to breakdowns immediately. When both priorities conflict — and they always conflict, because breakdowns do not respect PM schedules — the breakdown wins. The PM is signed off to protect the metric. The equipment runs without maintenance. And the next breakdown comes sooner.

PM compliance is the most manipulated metric in maintenance management. Not because maintenance technicians are dishonest, but because the system demands a number that the system's own priorities make impossible to achieve.

The frequency fiction

The second problem with most PM schedules is that the frequencies are wrong. They were set by the equipment manufacturer — a number in the installation manual, based on the manufacturer's testing under conditions that do not match your operating environment, your material, your duty cycle, or your operator behaviour. Or they were set by a maintenance manager ten years ago who picked "monthly" because it sounded reasonable.

I have never audited a PM schedule where the frequencies were based on actual failure data for the specific equipment in the specific operating context. And yet, this is the only basis that makes sense. A bearing in a high-vibration, high-temperature, high-contamination environment will fail faster than the same bearing in a clean, temperature-controlled environment. If your PM frequency is based on the bearing manufacturer's generic recommendation, you are either maintaining too often — wasting time and money on unnecessary interventions — or not often enough, and waiting for the failure to tell you.

The plants that get maintenance right use reliability-centred maintenance (RCM) principles: they analyse actual failure modes, determine the failure patterns for each mode (age-related, random, or infant mortality), and set PM frequencies based on the actual failure distribution. This is harder than copying numbers from a manual. It is also the difference between a PM schedule that prevents failures and one that creates a false sense of security.

The checklist charade

The third problem is the content of the PM task. Most PM checklists I have audited are a list of generic activities: "inspect," "check," "clean," "lubricate." These words appear on PM checklists because they are easy to write, easy to sign, and impossible to audit for quality. What does "inspect hydraulic system" mean? What is the technician looking for? What acceptance criteria apply? What happens if the inspection finds something?

In one plant, the PM checklist for a CNC machining centre included the task "check spindle for abnormal noise." I asked the maintenance technician what constituted abnormal noise. "You know," they said. "Something that sounds wrong." I asked how they knew what wrong sounded like. "Experience," they said. I asked how a new technician would know. They shrugged.

A PM task that depends on the subjective judgement of an experienced technician is not a system. It is a person. When that person retires, transfers, or is reassigned, the PM task loses its effectiveness, and nobody knows because the compliance metric still shows 95%.

Effective PM tasks are specific, measurable, and actionable. Not "check spindle noise" but "measure spindle vibration at 3000 RPM using accelerometer at position X. Record value in mm/s. If value exceeds 4.5 mm/s, generate work order for spindle inspection." This is a task that any trained technician can perform, that produces a recordable data point, and that has a defined response. It is also a task that costs more to develop and deploy than "check spindle noise." Which is why most plants do not do it.

The parts problem

Even when the PM is correctly scheduled, correctly scoped, and correctly executed, there is one more failure mode: spare parts. I have seen PM tasks completed perfectly — inspection done, issue identified, replacement part ordered — and then the part sits on backorder for six weeks while the equipment runs to failure. The PM did its job. The follow-up did not happen.

In one aerospace plant, a PM inspection identified wear on a critical hydraulic valve. The maintenance team ordered the replacement part immediately. The part was on a twelve-week lead time. Four weeks before the part arrived, the valve failed during production, causing an uncontrolled hydraulic fluid release that contaminated a batch of 180 parts. The PM had prevented nothing. The failure mode it identified was allowed to progress to failure because the supply chain could not deliver the replacement in time.

Effective preventive maintenance requires effective spare parts management. Critical spare parts — parts whose failure would stop production or affect safety — must be identified, stocked, and managed. The cost of stocking critical spares is trivial compared to the cost of a production stoppage. And yet, I have audited plants where the spare parts inventory for a 20-million-euro production line consisted of bearings, seals, and filters valued at less than 15,000 euros. The maintenance manager knew the risk. The finance director had cut the inventory budget.

What world-class maintenance looks like

World-class maintenance is not about PM compliance percentages or CMMS dashboards. It is about a simple outcome: equipment runs when it is supposed to run, and does not fail unexpectedly. This requires four things:

1. Condition-based monitoring. Instead of time-based PM tasks, monitor the actual condition of the equipment. Vibration analysis, oil analysis, thermography, ultrasound — these technologies tell you when a component is degrading, before it fails. They replace calendar-based PM with condition-based PM, which is more effective and more efficient. A plant with good condition monitoring will intervene on a bearing when the vibration trend shows degradation, not when the calendar says "monthly inspection."

2. Operator-led maintenance. The operator who runs the equipment eight hours a day knows more about its behaviour than any maintenance technician who visits once a month. Simple daily checks — visual inspection, listening for changes, checking temperatures, cleaning — conducted by the operator catch problems early. This is not about reducing the maintenance team's workload. It is about adding the operator's sensory data to the maintenance system.

3. Failure mode-driven PM. Every PM task should map to a specific failure mode that has been identified through FMEA or failure data analysis. If a PM task does not prevent a specific, identified failure mode, it is not preventive. It is ritual.

4. Mean time between failure as the primary metric. Not PM compliance. Not maintenance cost. MTBF. If MTBF is improving, the maintenance system is working. If it is flat or declining, the PM schedule is not preventing failures, regardless of what the compliance metric says.

The uncomfortable truth

Most maintenance systems I have audited are designed to satisfy auditors, not to prevent failures. The PM schedule is a document that demonstrates "we have a maintenance system." The compliance metric demonstrates "we execute the system." The CMMS report demonstrates "we track the system." And the equipment fails on a schedule that is entirely independent of the PM schedule, because the PM tasks were never designed to address the actual failure modes of the equipment in its actual operating context.

If your preventive maintenance schedule is not reducing your unplanned downtime year over year, it is not preventing anything. It is creating the illusion of control while the equipment degrades according to its own schedule, not yours. The fix is not more PM tasks, higher compliance targets, or better CMMS software. The fix is understanding your equipment, understanding your failures, and maintaining based on reality rather than on a calendar.

Your equipment does not care about your PM schedule. It cares about condition. And condition does not have a calendar. It has a trend. Learn to read the trend, and you will prevent failures. Ignore the trend, and you will explain them. That is the choice. Everything else is marketing.