Preventive maintenance performs tasks at fixed time or usage intervals regardless of actual equipment condition, while predictive maintenance uses real-time condition data, vibration, temperature, oil analysis, to determine maintenance timing based on how the equipment is actually behaving. Both are proactive strategies aimed at avoiding breakdown, but they trigger maintenance work in fundamentally different ways, on a calendar versus on evidence.
The distinction matters because choosing the wrong approach for a given asset wastes money in one of two directions. Preventive maintenance applied to equipment with unpredictable failure patterns risks over-maintaining some machines and under-maintaining others, since a fixed interval assumes a failure pattern that may not match reality. Predictive maintenance applied to low-criticality equipment risks spending on sensors and analysis capability that the asset's actual risk profile never justified.
This guide covers how each approach actually works, the real cost and complexity tradeoffs between them, how to decide which strategy fits a given asset, and where the two approaches genuinely work best combined rather than chosen exclusively.
How Preventive Maintenance Actually Works
Preventive maintenance schedules tasks based on elapsed time, usage cycles, or a fixed interval defined by manufacturer recommendation or internal experience, independent of the equipment's actual current condition.
The Core Mechanism
A preventive maintenance task fires when its trigger, a calendar date, an hour meter, a production cycle count, is reached, regardless of whether the specific machine actually needs attention at that moment. This is what makes preventive maintenance simple to implement and predictable to schedule, since the entire program can be planned months in advance without any ongoing condition monitoring infrastructure.
The Structural Tradeoff
The same fixed-interval logic that makes preventive maintenance simple also makes it structurally imprecise. A component serviced on schedule might have had significant remaining useful life, representing wasted maintenance spend, while a different unit of supposedly identical equipment operating under harsher conditions might fail well before its next scheduled interval arrives.
Key Insight: Preventive maintenance trades precision for simplicity, scheduling by calendar rather than condition, which produces both wasted early maintenance and occasional missed failures.
How Predictive Maintenance Actually Works
Predictive maintenance monitors actual equipment condition in real time and triggers maintenance only when data indicates a genuine developing problem, rather than on a fixed schedule.
The Core Mechanism
Sensors and monitoring technologies, vibration analysis, thermal imaging, oil analysis, and acoustic monitoring among the most common, continuously or periodically measure specific equipment health indicators against an established normal baseline. When a measurement drifts meaningfully from that baseline, the deviation triggers a maintenance action targeted at the specific developing issue the data has identified.
The Structural Tradeoff
This precision comes at a real cost. Predictive maintenance requires upfront investment in sensors, monitoring infrastructure, and the analytical capability to interpret the resulting data correctly, along with an established baseline period before the system can reliably distinguish a genuine developing problem from normal variation.
Key Insight: Predictive maintenance trades simplicity for precision, requiring real investment in monitoring infrastructure to schedule maintenance only when equipment condition data genuinely warrants it.
Reliability-centered maintenance, which formally weighs failure modes and consequences to select the right strategy per asset, including this preventive-versus-predictive decision, is covered in [Reliability-Centered Maintenance: A Practical Guide].
Choosing Between Preventive and Predictive for a Given Asset
Neither approach is universally correct. The right choice depends on specific characteristics of the asset in question, not organizational preference for one methodology over the other.
When Preventive Maintenance Fits Better
Preventive maintenance fits equipment with well-understood, predictable wear patterns, where a component's typical service life is genuinely consistent across units and operating conditions, and where the cost of the fixed-interval task itself is low relative to the cost of monitoring infrastructure that predictive maintenance would require.
When Predictive Maintenance Fits Better
Predictive maintenance fits high-criticality equipment where unplanned failure is expensive, and where failure modes are genuinely detectable through condition monitoring before they cause a breakdown. McKinsey's analysis of industrial maintenance programs found that predictive maintenance approaches can reduce maintenance costs by 18 to 25 percent when applied to the right assets, a gain that comes specifically from avoiding both over-maintenance and unplanned failure on equipment where the investment is justified.
- Preventive maintenance fits: predictable wear, low monitoring cost justification, lower-criticality equipment
- Predictive maintenance fits: high-criticality equipment, detectable failure modes, cost of unplanned downtime that justifies monitoring investment
Key Insight: The right strategy depends on asset criticality and failure-pattern predictability, not a blanket preference for either approach across the entire plant.
A Worked Example: Choosing Between the Two Strategies
The decision framework is easier to apply against two specific, contrasting assets rather than in the abstract: a standard HVAC exhaust fan and a critical production line gearbox in the same facility.
The Exhaust Fan
The exhaust fan runs continuously but at low criticality, since a failure causes discomfort rather than a production stoppage, and a replacement unit can typically be installed within a day. Its wear pattern, bearing degradation over a fairly predictable service life, is well understood and consistent across similar units. Preventive maintenance, a scheduled bearing inspection and lubrication every few months based on manufacturer guidance, fits this asset well, since the cost of occasionally servicing it slightly early is trivial compared to any monitoring infrastructure investment.
The Production Line Gearbox
The gearbox sits directly in the critical path of the main production line, where an unplanned failure stops output entirely and a replacement or major repair takes days to complete. Its failure modes, gear tooth wear and bearing degradation, are both genuinely detectable through vibration analysis well before they cause a breakdown. Predictive maintenance is justified here specifically because the cost of a single unplanned failure would exceed years of monitoring investment, and the failure signature is detectable early enough to act on.
Key Insight: The same decision framework applied to a low-criticality fan and a critical gearbox produces two entirely different, both correct, strategy choices for the same facility.
Cost and Complexity Compared Directly
Beyond the criticality-driven decision logic, the two approaches differ concretely across four practical dimensions worth weighing before committing budget to either one.
- Upfront investment: Preventive maintenance requires minimal upfront cost beyond a documented schedule and standard parts inventory. Predictive maintenance requires sensors, monitoring software, and a baseline data collection period before the system becomes reliable
- Ongoing labor: Preventive maintenance labor is predictable and schedulable well in advance. Predictive maintenance requires analytical capability to interpret condition data correctly, either in-house expertise or a monitoring service, which is an ongoing cost beyond the sensors themselves
- Failure coverage: Preventive maintenance addresses wear-related failure modes well but does not catch failures outside the expected pattern. Predictive maintenance catches a wider range of developing issues, provided the specific failure mode is one the chosen monitoring technology can actually detect
- Time to value: Preventive maintenance can be implemented and functioning within weeks. Predictive maintenance typically requires a baseline period of weeks to months before the system can reliably distinguish genuine deviation from normal variation
Key Insight: Preventive maintenance wins on upfront simplicity and speed to implement, while predictive maintenance wins on failure coverage and precision, once its baseline period is complete.
Combining Both Approaches in Practice
Most manufacturing plants do not choose exclusively between preventive and predictive maintenance. They apply each strategy to the assets it fits best, within the same broader planned maintenance program.
A Tiered Approach Across the Plant
Lower-critical equipment typically stays on preventive maintenance, since the cost of occasional early servicing is modest and monitoring infrastructure would not pay for itself. Higher-critical equipment, where unplanned failure carries a genuine production or safety cost, is where predictive investment concentrates, since that is where avoiding a single unplanned failure can offset the monitoring cost many times over.
Predictive Maintenance as an Evolution, Not a Replacement
Predictive maintenance is often described as a form of preventive maintenance in the broader sense, since both aim to prevent failure before it happens. The practical distinction that matters for implementation is not philosophical but operational: predictive maintenance is the more complex and more precise version of that same underlying goal, applied selectively where the investment is justified.
Key Insight: Most plants run a tiered program, preventive maintenance for lower-criticality equipment and predictive maintenance concentrated where unplanned failure cost justifies the monitoring investment.
Within the Lean System
Connection to Lean Principles
Choosing preventive versus predictive maintenance deliberately, rather than defaulting to one approach across every asset, reflects the lean principle of eliminating waste, in this case the waste of over-maintaining low-risk equipment or under-monitoring high-risk equipment through a one-size-fits-all policy.
Connection to Lean Tools
Both strategies are inputs into the broader [Planned Maintenance: Optimization Strategies for Reliability] program, which is where the scheduling and criticality-ranking discipline that determines which asset gets which strategy actually happens. The formal methodology for making that determination is [Reliability-Centered Maintenance: A Practical Guide].
Connection to Continuous Improvement
Tracking which strategy assignment turns out to be wrong for a given asset, a preventive-maintenance asset that keeps failing between intervals, or a predictive-maintenance asset whose monitoring investment never pays off, is a natural input into the [PDCA Cycle: The Foundation of Continuous Improvement], adjusting the strategy assignment based on real operating evidence rather than the original assumption.
Frequently Asked Questions
Q: What is the main difference between preventive and predictive maintenance?
Preventive maintenance performs tasks at fixed time or usage intervals regardless of actual equipment condition. Predictive maintenance uses real-time condition data, such as vibration or temperature readings, to trigger maintenance only when the data indicates a genuine developing problem rather than an arbitrary calendar date.
Q: Is predictive maintenance always better than preventive maintenance?
No. Predictive maintenance requires real investment in sensors and monitoring capability that only pays off on higher-criticality equipment where unplanned failure is genuinely costly. Lower-criticality equipment with predictable wear patterns is often better and more economically served by simpler preventive maintenance instead.
Q: How much can predictive maintenance actually save?
McKinsey's analysis of industrial maintenance programs found predictive maintenance can reduce maintenance costs by 18 to 25 percent when applied to the right assets, though the actual gain depends heavily on selecting equipment where failure modes are genuinely detectable through condition monitoring in the first place.
Q: Can preventive and predictive maintenance be used together?
Yes, and most manufacturing plants do exactly this in practice. Lower-critical equipment typically stays on preventive schedules, while predictive monitoring concentrates specifically on higher-criticality assets where the cost of avoiding unplanned failure clearly justifies the monitoring investment required.
Q: What technologies are commonly used for predictive maintenance?
Vibration analysis, thermal imaging, oil analysis, and acoustic monitoring are the most common condition-monitoring technologies used in practice, each suited to detecting different failure modes depending on the type of equipment and the specific wear pattern actually being monitored.
LeanSuite: A complete lean manufacturing software
Schedule Demo
