Preventive vs Predictive Maintenance: When to Use Each

Preventive maintenance follows a calendar. Predictive maintenance follows the data. Both are valid strategies, and most well-run plants use a combination of the two. The question is not which one is better. The question is which one fits each piece of equipment in your facility.

A time-based oil change every 3,000 hours works perfectly well on a standard gearbox. But putting a $2,000 vibration sensor on a $500 utility pump makes no financial sense, no matter how good the data is. On the other hand, a critical compressor with a $50,000 replacement cost and a history of random bearing failures is a textbook case for predictive monitoring.

This article walks through the differences between the two strategies, the cost and complexity tradeoffs, and a practical framework for deciding which approach fits which equipment. If you are looking for a comparison against reactive (run-to-failure) maintenance instead, see preventive vs reactive maintenance.

Preventive Maintenance: The Calendar Approach

Preventive maintenance (PM) means servicing equipment on a fixed schedule: every 30 days, every 500 run hours, every 10,000 cycles. The work happens regardless of equipment condition. When the calendar says it is time, you inspect, clean, lubricate, adjust, or replace the component.

The schedule comes from three sources:

• Manufacturer recommendations. The OEM specifies service intervals based on design life and testing. These are a good starting point but often conservative.
• Failure history. Your own data on when components typically fail. If bearings on a particular conveyor last 14 months on average, scheduling replacement at 12 months gives you a safety margin.
• Technician experience. Your experienced maintenance team knows which components need more attention and which can go longer between services. This tribal knowledge is valuable input for setting intervals.

PM is the backbone of most industrial maintenance programs. It is straightforward to plan, easy to schedule, and does not require sensor infrastructure or analytical tools. A well-run PM program typically reduces emergency breakdowns by 25-40% compared to a purely reactive approach. For a full guide, see what is preventive maintenance.

Strengths of Preventive Maintenance

• Simple to implement and manage with a basic CMMS or even a spreadsheet
• Works on any equipment regardless of age, type, or sensor availability
• Low technology overhead: checklists, hand tools, basic inspection skills
• Predictable workload for maintenance planning and scheduling
• Reduces breakdowns significantly compared to reactive maintenance

Weaknesses of Preventive Maintenance

• Over-maintenance. You replace parts that still have useful life remaining. That bearing you changed at 12 months might have run perfectly to 20 months. PM trades some useful component life for the safety of a scheduled replacement.
• Under-maintenance. Time-based schedules miss failures that happen between intervals. A bearing that starts degrading one week after its annual inspection will not get caught until the next inspection, 11 months later.
• Wasted labor on healthy equipment. Technicians spend time inspecting machines that show no signs of problems. Across hundreds of assets, this adds up to significant labor hours that could be used elsewhere.
• Does not catch random failures. PM works best on failure modes with a predictable wear pattern. Random, sudden failures (electrical faults, seal blowouts from foreign object damage) are not prevented by time-based schedules.

Predictive Maintenance: The Data Approach

Predictive maintenance (PdM) means monitoring equipment condition in real time and performing maintenance when the data shows that a failure is developing. Instead of replacing a bearing every 12 months, you install a vibration sensor that detects early-stage bearing degradation and replace it only when the data says it needs it.

PdM relies on condition monitoring technologies:

• Vibration analysis for rotating equipment: bearings, shafts, gearboxes, motors. Detects imbalance, misalignment, and bearing wear weeks to months before failure. Read our vibration analysis basics guide.
• Oil analysis for hydraulic systems, gearboxes, and engines. Detects contamination, metal particles, and chemical degradation.
• Thermal imaging for electrical systems, connections, and heat-generating equipment. Detects hot spots from loose connections, overloaded circuits, or failing insulation.
• Ultrasonic testing for compressed air leaks, steam traps, and bearing lubrication. See our compressed air leak detection guide.
• Motor current analysis for electric motors. Detects rotor bar defects, stator winding issues, and mechanical load problems.

PdM catches developing problems that time-based PM misses. It also avoids unnecessary maintenance on equipment that is running fine. Plants with mature PdM programs report 8-12% lower maintenance costs than PM-only plants, primarily from avoiding over-maintenance and catching failures earlier. For a detailed overview, see what is predictive maintenance.

Strengths of Predictive Maintenance

• Maintains equipment based on actual condition, not arbitrary calendar dates
• Catches developing failures weeks or months in advance
• Eliminates unnecessary maintenance on healthy equipment
• Extends component life by using the full useful life instead of replacing early
• Provides detailed diagnostic data that helps technicians plan better repairs

Weaknesses of Predictive Maintenance

• Higher upfront cost. Sensors, data collection hardware, software, and the initial setup cost significantly more than a clipboard and a PM checklist.
• Requires analytical skills. Someone on your team (or a contracted specialist) needs to interpret vibration spectra, oil analysis reports, and thermal images. The data does not interpret itself.
• Not suitable for all equipment. Simple, low-cost assets do not justify the sensor and analysis investment. Equipment without clear measurable degradation indicators is hard to monitor.
• Takes time to build baseline data. PdM models need 3-6 months of normal operation data before they can reliably detect anomalies. You are running on faith during that ramp-up period.
• Technology dependency. Sensors fail. Networks go down. Software needs updates. You need a plan for what happens when the monitoring system is offline.

Head-to-Head Comparison

Strategy Selection: Which Equipment Gets Which Approach?

The right strategy depends on three factors: how critical the asset is, how expensive the consequences of failure are, and whether the failure mode can be detected by sensors before it causes a breakdown.

Equipment That Suits Preventive Maintenance

PM is the right choice for equipment where:

• Failure follows a predictable time or usage pattern (seals, filters, belts, lubricants)
• The cost of condition monitoring exceeds the cost of periodic replacement
• The equipment is simple enough that visual inspection and basic tools catch most problems
• Manufacturer-recommended intervals are well-established and reliable

Typical PM-first equipment: conveyor belts, basic pumps, air filters, lubrication points, drive belts, simple valves, lighting systems, safety equipment (fire extinguishers, emergency showers).

Equipment That Suits Predictive Maintenance

PdM is the right choice for equipment where:

• Failure consequences are severe (production loss, safety risk, environmental impact)
• The failure mode produces a detectable symptom before breakdown (vibration, heat, noise, oil contamination)
• The lead time between detectable symptom and failure is long enough to plan a repair (days to weeks, not minutes)
• The cost of the monitoring equipment is justified by the asset value and failure cost

Typical PdM-first equipment: large electric motors, critical pumps and compressors, gearboxes, turbines, transformers, CNC spindles, high-speed bearings, cooling systems on critical processes.

The Hybrid Approach: Best of Both

Most plants should not choose PM or PdM exclusively. The practical answer is a hybrid approach where you use the right strategy for each asset based on the matrix above.

Here is what a hybrid program looks like in practice:

• 80% of assets on PM. Standard time-based or usage-based PM covers the majority of your equipment. This is your foundation.
• 10-15% of assets on PdM. Your most critical, highest-consequence assets get continuous or periodic condition monitoring in addition to basic PM.
• 5-10% of assets on run-to-failure. Non-critical, low-cost, no-collateral-damage items are deliberately run to failure.

The hybrid approach gives you the broad coverage of PM with the precision of PdM on the assets where it matters most. As your program matures and your team develops condition monitoring skills, you can gradually shift more assets from PM to PdM.

Building a Hybrid Program: Step by Step

If you already have a PM program and want to add predictive capabilities, here is the practical path.

Step 1: Identify your PdM candidates

Pull your asset criticality rankings and your breakdown history. Look for assets that are: high criticality, high repair cost, frequent or costly breakdowns despite existing PM, and equipped with rotating components or measurable degradation indicators. Your top 5-10 assets on this list are your PdM pilot candidates.

Step 2: Start with route-based monitoring

You do not need online sensors to start PdM. A portable vibration meter, a thermal camera, and an ultrasonic detector let a trained technician walk a route and collect condition data on 20-30 assets in a few hours. Monthly routes on your critical assets give you 80% of the PdM benefit at 20% of the cost of permanent online monitoring.

Step 3: Add online sensors where justified

For your top 3-5 most critical assets, consider permanent vibration sensors with wireless data transmission. These provide continuous monitoring and early alerts without requiring a technician to walk a route. The cost per sensor has dropped significantly in recent years, making online monitoring practical for more equipment than it was five years ago.

Step 4: Develop internal analytical skills

Send one or two technicians to vibration analysis certification (ISO 18436-2). Train your team on basic thermal imaging interpretation. Build internal capability so you are not dependent on external consultants for every data review. Machine learning tools can supplement this by flagging anomalies for human review.

Step 5: Adjust PM schedules based on PdM data

As you collect condition data, use it to tune your PM intervals. If vibration monitoring shows bearings consistently lasting 18 months on a machine where you replace them every 12 months, extend the PM interval. If oil analysis shows contamination building faster than expected, shorten the oil change interval. This is where PM and PdM work together instead of against each other.

Common Mistakes to Avoid

• Jumping to PdM before PM is solid. If your PM compliance is below 80%, you have a bigger problem than sensor technology. Fix your PM program first. PdM is a layer on top of PM, not a replacement for it.
• Putting sensors on everything. Monitoring a $200 motor with a $1,500 sensor setup does not make financial sense. Use the criticality matrix to focus PdM investment where the payback is clear.
• Collecting data without acting on it. If nobody reviews the vibration data or responds to alerts, the sensors are just expensive decorations. Assign clear ownership: who reviews the data, how often, and what happens when a threshold is exceeded.
• Expecting instant ROI. PdM needs 6-12 months to build baselines and start catching real problems. Budget for a ramp-up period where you are investing without visible returns. The payback comes when you avoid the first major unplanned breakdown that would have cost $50,000+.
• Ignoring the P-F interval. The P-F interval is the time between when a failure becomes detectable (P) and when it causes a functional failure (F). If the P-F interval is 2 days, monthly monitoring will miss it. Match your monitoring frequency to the P-F interval of the failure modes you are trying to catch.

Connecting to OEE and Other KPIs

Both PM and PdM improve OEE by reducing the Availability losses from unplanned breakdowns. PdM adds an additional benefit by extending component life, which reduces your overall parts spend. Track these KPIs to measure the impact of your maintenance strategy choices:

• OEE and specifically the Availability factor. This is your top-level measure.
• MTBF (Mean Time Between Failures). Should increase as PM and PdM reduce unplanned failures.
• MTTR (Mean Time to Repair). Should decrease as PdM gives you advance warning to pre-stage parts and plan repairs.
• Maintenance cost per unit produced. Should decrease as you shift from reactive to planned and from over-maintenance to condition-based.
• PM Compliance. Tracks whether your PM program is being executed consistently.

For a complete list of metrics, see 15 maintenance KPIs every plant manager should track.

Where Dovient Fits

Dovient helps maintenance teams get more value from both PM and PdM programs by making repair knowledge accessible and actionable.

• PM that improves over time. When technicians document what they find during PM tasks, including photos, notes, and measurements, Dovient builds a condition history for each asset. Over time, this data helps you set smarter PM intervals.
• Faster response when PdM finds a problem. When a vibration alert says "bearing degradation on Pump 7," the next question is: how do we fix it? Dovient matches the symptom to past repairs on that asset and gives the technician a clear path to the fix.
• Knowledge that stays in the plant. Your best vibration analyst is retiring next year. The insights they have about your equipment need to survive their departure. Dovient captures that expertise so the next person can access it at the machine.

Want to talk about your maintenance strategy? Schedule a conversation with our team to discuss which approach fits your plant.