If you manage a factory long enough, you'll hear three words thrown around constantly: reactive, preventive, predictive. Each is a philosophy, a budget line item, and — if you get it wrong — a source of chronic downtime.
The problem isn't choosing one. The problem is most plants don't know where they stand, what each strategy actually costs, or which mix makes sense for their specific equipment.
Here's a practical breakdown: what each approach looks like on the factory floor, what the data says about cost, and how to build a strategy that fits your plant — not a consultant's slide deck.
Reactive Maintenance: Fix It When It Breaks
Also called "run-to-failure." You wait until a machine stops working, then you fix it. No planning, no scheduling, no inspection between failures.
What it looks like in practice:
- A conveyor belt snaps at 2 PM on a Tuesday. Production stops. The maintenance team drops everything, finds the belt in the store, replaces it, and the line runs again by 4 PM. No record is kept of why it snapped.
- A pump seal leaks overnight. By morning, the motor is flooded. Now you're replacing both the seal and the motor. Total cost: 3x what the seal replacement would have cost.
The cost data: Industry studies consistently show that reactive maintenance costs 3 to 5 times more than planned preventive maintenance. This isn't just theory — the U.S. Department of Energy and multiple reliability studies confirm the same number.
Why the multiplier is so high:
| Cost Factor | Planned Repair | Emergency Repair |
|---|---|---|
| Labor cost | Normal rates | Overtime + call-out fees |
| Parts cost | Ordered at standard price | Expedited shipping, premium pricing |
| Downtime impact | Scheduled during off-hours | Unplanned production loss |
| Secondary damage | None | Often damages other components |
| Technician efficiency | Prepared with right tools | Multiple trips, missing parts |
When reactive makes sense: Not every machine deserves a PM program. For low-cost, non-critical equipment with redundancy — a spare pump, a backup fan, a lighting circuit — run-to-failure can be the most economical choice. A $200 fan doesn't need a $500 annual PM program.
Preventive Maintenance: Schedule It Before It Breaks
Preventive maintenance (PM) is maintenance performed on a fixed schedule — time-based or usage-based — to reduce the probability of failure.
What it looks like in practice:
- Every 30 days, a technician inspects the air compressor: checks oil level, belts, filters, and drainage. Takes 20 minutes. The work order is generated automatically by the CMMS.
- After every 5,000 operating hours, the gearbox oil is changed. The meter reading triggers the work order. The oil change costs $150 in materials and one hour of labor. Skipping it would eventually cost $8,000 for a gearbox rebuild.
The data: A well-executed PM program delivers 30-50% reduction in unplanned downtime and 15-25% longer equipment life, according to data compiled across multiple industries.
The trap: PM programs easily become bloated. "We've always greased it monthly" — whether it needs it or not. Tasks get copied from one plant to another without question. Technicians develop "checkbox mentality" — walking through inspections without actually inspecting. The result is wasted labor and equipment that still breaks.
When PM makes sense: For critical equipment with known failure patterns. If a bearing typically fails after 8,000 hours of operation, schedule replacement at 7,500 hours. If a filter clogs after 3 months, replace it at 2.5 months. PM works when the failure mode is predictable and the task directly addresses it.
Predictive Maintenance: Fix It When Data Says It Will Break
Predictive maintenance (PdM) uses condition monitoring data — vibration, temperature, oil analysis, thermal imaging — to determine the actual condition of equipment and predict when failure will occur. Instead of maintaining on a schedule, you maintain on demand.
What it looks like in practice:
- Vibration sensors on a motor transmit data weekly. Over three months, the vibration trend increases gradually. At month four, the system flags the bearing as "approaching failure." The bearing is replaced during a planned shutdown — before it fails. The motor never stops unexpectedly.
- Oil analysis on a hydraulic system shows increasing particle count. The filter is changed and the oil is polished. What would have been a $12,000 pump failure becomes a $400 maintenance task.
The cost reality: PdM delivers the highest ROI per maintenance dollar spent — studies show a 5:1 to 10:1 return on PdM investment. But it requires sensors, software, and trained personnel. A full vibration monitoring program on 100 motors costs more to set up than a PM program for the entire plant.
When PdM makes sense: For critical assets where failure is catastrophic — in terms of safety, production loss, or repair cost. And only when you already have a solid PM program in place. Implementing PdM on a plant that can't even do basic PM is like installing a GPS on a car with no engine.
| Approach | Best For | Cost Profile | Complexity |
|---|---|---|---|
| Reactive | Low-criticality, low-cost equipment | Low upfront, high per-failure | None |
| Preventive | Equipment with predictable failure patterns | Moderate upfront, predictable | Low to moderate |
| Predictive | Critical assets with catastrophic failure risk | High upfront, very low per-failure | High |
The Hybrid Approach: What Most Plants Actually Need
Here's the truth that doesn't make it into the brochures: very few plants run 100% preventive or 100% predictive. The most reliable, cost-effective plants run a hybrid strategy.
A typical breakdown in a well-managed plant:
- Reactive: 10-15% of maintenance activity — for truly low-criticality items
- Preventive: 60-70% — the backbone of the program for most equipment
- Predictive: 15-25% — for the most critical, highest-impact assets
The mix shifts as the plant matures and as technology costs decrease. But 100% predictive is not the goal. The goal is the right level of maintenance for each asset.
Decision Framework: What Strategy Goes Where
Use a simple criticality approach:
Step 1: Classify every asset
- A (Critical): Failure stops production, creates safety risk, or costs >$10,000/hour downtime
- B (Important): Failure slows production or costs <$10,000/hour
- C (Support): Failure is a minor inconvenience
Step 2: Match strategy to criticality
- A assets: Predictive (vibration, oil analysis, thermography) + robust PM backup
- B assets: Preventive (time-based or usage-based PM) + condition-based where practical
- C assets: Reactive (run-to-failure) or minimal PM
Step 3: Review and adjust quarterly
- Are A-asset failures decreasing? Is the PdM program detecting issues before failure?
- Are B-asset PMs still relevant? Any tasks that can be removed or frequencies extended?
- Are C-asset failures causing unexpected problems? Maybe they need to be reclassified.
How to Start Transitioning from Reactive to Planned
Most plants in Southeast Asia operate at 50-70% reactive maintenance. Getting from there to a balanced hybrid strategy takes time — but the first steps are straightforward.
Month 1-2: Get visibility. Start tracking what's breaking, how often, and how much it costs. Without data, every decision is a guess. A CMMS like OpexMX gives you this in days, not months.
Month 3-4: Build PMs for your top 20 failure modes. Don't try to cover everything. Identify the failures that cost you the most downtime or the most money, and build PM tasks that directly prevent them.
Month 6-8: Add condition monitoring for your most critical assets. Start simple — vibration analysis on your top 10 motors, oil analysis on your top 5 hydraulic systems. Prove the ROI before scaling.
Month 9-12: Review and refine. Look at your failure data. Are the PMs working? Is the PdM catching issues? Adjust frequencies, add tasks, remove what's not working.
The worst thing you can do is try to do everything at once and burn out your team. Slow, systematic improvement beats a failed "big bang" implementation every time.
The Bottom Line
Reactive maintenance is expensive. Preventive maintenance is the minimum viable strategy. Predictive maintenance is a powerful tool for the right assets. Most plants need all three, in the right proportions.
The question isn't which strategy is "best." It's which strategy is best for each specific machine in your plant.
Start with data. Build your PM foundation. Add PdM where it matters. Let reactive be the exception, not the default.
See how OpexMX helps you track reactive vs planned maintenance, build PM schedules, and identify PdM opportunities — built for factories in Southeast Asia, not boardrooms in Silicon Valley.