"Fix it before it breaks." That's preventive maintenance in five words. Simple, obvious, and — if you've ever tried to run a PM program in a real factory — surprisingly hard to get right.
Preventive maintenance (PM) is maintenance performed on a regular schedule to reduce the likelihood of equipment failure. Unlike reactive maintenance (fix it when it breaks) or predictive maintenance (fix it when data says it's about to break), PM is calendar-based: you service the machine at fixed intervals regardless of its current condition.
Sounds reasonable. So why do so many PM programs fail?
The Three Types of Preventive Maintenance
Not all PM is the same. Understanding the differences is where good programs separate from bad ones.
1. Time-Based PM
The simplest form. You perform maintenance at fixed calendar intervals:
- Inspect the boiler every Monday
- Grease the conveyor bearings every 2 weeks
- Replace hydraulic filters every 3 months
- Overhaul the compressor annually
Time-based PM is easy to schedule and easy to understand. The downside? It ignores actual usage. A machine running 24/7 needs service more often than one running 8 hours a day. If your entire PM program is time-based, you're either over-maintaining some equipment or under-maintaining others.
2. Usage-Based (Meter-Based) PM
Maintenance triggered by actual operating metrics rather than calendar time:
- Change oil after every 5,000 operating hours
- Replace belts after 100,000 cycles
- Service the motor after processing 50,000 units
Usage-based PM is more accurate than time-based because it tracks actual wear. The challenge: you need reliable meters, counters, or integration with your production system to track usage. Many plants start with time-based PM and evolve toward usage-based as they mature.
3. Condition-Based PM
The bridge between preventive and predictive maintenance. You inspect equipment on a schedule, but the inspection determines whether maintenance actually happens:
- Check vibration levels monthly; replace bearings only if vibration exceeds threshold X
- Inspect belt tension quarterly; adjust or replace only if tension is out of spec
- Test insulation resistance every 6 months; rewind motor only if resistance drops below Y
Condition-based PM is more efficient than pure time-based because you're not replacing parts that are still good. But it requires trained technicians who can perform inspections correctly and a system to track readings over time.
The PM Trap: Why Most Programs Fail
Here's the uncomfortable truth: most preventive maintenance programs are bloated, outdated, and waste technician time on tasks that don't prevent failures.
The "copy-paste" problem. Many PM programs are inherited — copied from a previous plant, a manufacturer's manual that hasn't been updated since 2008, or an Excel sheet maintained by someone who left three years ago. Nobody questions whether the tasks are still relevant because "we've always done it this way."
The checkbox mentality. Technicians walk through PM routes ticking boxes: "inspected." "checked." "OK." Are they actually inspecting? Or are they going through the motions because they have 40 more PMs to complete before lunch?
The PM backlog spiral. PMs generate work orders. When the work orders pile up, the backlog grows. When the backlog grows, PMs get deferred. When PMs get deferred, unplanned failures increase, which generates more reactive work orders, which pushes PMs further behind. This is the death spiral of a PM program.
Building a PM Program That Actually Works
Step 1: Identify What's Critical
Not every machine deserves the same level of attention. Use a criticality matrix:
- Critical (A) — failure stops production, creates safety risk, or costs >$10,000 per hour of downtime. These get the most frequent, most thorough PMs.
- Important (B) — failure slows production or costs money but doesn't stop everything. Less frequent PMs, condition-based where possible.
- Support (C) — failure is a minor inconvenience. Minimal PM. Maybe just an annual check.
Too many plants give every machine the same PM frequency. That's how you end up with 400 PMs per week and a backlog of 1,200.
Step 2: Define Tasks That Actually Prevent Failure
For each critical asset, ask: "What specific failure modes are we trying to prevent?" Then design tasks that address those failure modes.
Bad PM task: "Inspect pump." Good PM task: "Check pump seal for leakage (>2 drops/min = replace seal). Measure bearing temperature with IR gun (>75°C at housing = investigate). Record readings."
The first tells a technician to look at something. The second tells them what to look for, what's acceptable, and what to do about it.
Step 3: Set Realistic Frequencies
PM frequency should be based on:
- Manufacturer recommendations (but with skepticism — they're often conservative)
- Your own failure data — how often does this thing actually break?
- Operational context — a dusty environment wears equipment faster than a clean room
Don't default to "monthly" because it's easy. A pump that fails every 8 months doesn't need monthly PMs. A conveyor in a 24/7 operation might need weekly inspections even though the manual says quarterly.
Step 4: Track Compliance and Adjust
Run the PM program for 90 days, then review:
- PM compliance rate — are PMs being completed on time?
- Failures detected during PMs — are the inspections actually finding problems?
- Failures between PMs — what's still breaking that the PMs aren't catching?
Use this data to adjust frequencies, add tasks, remove unnecessary tasks, and rebalance technician workloads. A PM program is never "done." It's a continuous improvement cycle.
Step 5: Use a CMMS
This is where the technology piece becomes essential. A CMMS (Computerized Maintenance Management System) automates the scheduling, generates work orders automatically, tracks compliance, and stores inspection data over time.
Without a CMMS, you're managing PMs in Excel or on a whiteboard — which works for about 20 machines and one shift. Beyond that, it falls apart. A CMMS ensures that PMs are generated, assigned, tracked, and recorded without anyone having to remember anything.
The Payoff
A well-run PM program delivers measurable results:
- 30-50% reduction in unplanned downtime — the most cited benefit across industry studies
- 15-25% longer equipment life — machines that are maintained on schedule last longer than machines that are run to failure
- Lower repair costs — a scheduled bearing replacement costs a fraction of replacing the bearing plus the shaft it destroyed when it seized
- Better production planning — maintenance happens in planned windows, not during production runs
- Audit-ready compliance — documented PM history satisfies regulators, insurers, and customers
Preventive maintenance isn't exciting. It won't win awards or make headlines. But it's the foundation everything else is built on. Predictive maintenance, condition monitoring, reliability-centered maintenance — all of them assume you're already doing the basics.
Start with PM. Do it well. The rest follows.
See how OpexMX handles PM scheduling and compliance tracking. Built for factories that need PM to actually work.