Predictive Maintenance vs. Preventive Maintenance (PPM) — The Difference and When It Pays

"Predictive maintenance" has become a magic phrase — sensors, IoT, artificial intelligence that knows when the compressor is about to break. And from this a feeling sometimes arises that old-fashioned preventive maintenance, the kind that runs by calendar, is already outdated. That is a mistake. Predictive maintenance and preventive maintenance do not compete with one another — they are two different places on the same continuum, and in most buildings in Israel the right combination between them, rather than a choice between the two, is what yields the best result. This article explains exactly how they differ, how the predictive world works in practice, and when it is truly worth the investment — alongside one principle that must never be forgotten for a moment: statutory mandatory inspections stay on the calendar, always, regardless of any sensor.

The maintenance continuum — three approaches, not two

To understand the place of predictive maintenance, you first have to lay out the whole continuum. Maintenance in a building moves along a single axis, from the most passive approach to the most proactive:

• Breakdown maintenance (reactive): you deal with something only when it breaks. The detector didn't beep, the pump went silent, the panel caught fire — and then you call a technician. This is the most expensive, most dangerous and most legally exposed approach. We expanded on its failure in preventive maintenance vs. breakdown maintenance in an office building.
• Time-based preventive maintenance (preventive / PPM): inspections and treatments planned according to a fixed schedule — every month, every six months, every year — regardless of the equipment's momentary physical condition. You replace a filter because thirty days have passed, not because you measured that it is clogged.
• Condition-based predictive maintenance (predictive / condition-based): you monitor the equipment's real condition in real time, and perform treatment only when the data indicates a deterioration that forecasts an approaching failure. You replace a bearing not by the calendar and not after it breaks — but when its vibration signature has begun to change.

The central distinction: preventive maintenance asks "how much time has passed?", while predictive maintenance asks "what is the equipment's condition now?". Both are proactive — both try to catch the fault before it happens — but they get there in different ways, and each way has its own cost, accuracy and appropriate place.

What preventive maintenance (PPM) is — and its strength and weakness

Preventive maintenance is the backbone of sound building management. It is simple to understand, easy to implement, and requires no special technology: a calendar, a list of systems, and a defined frequency for each one. It is also the approach on which almost all regulation in Israel is built — the legal inspections are defined at a fixed cadence, not by condition measurement. We built a whole guide on this concept in the complete guide to preventive maintenance in an office building.

The strength of time-based is its simplicity and certainty: you know in advance what will happen and when, it is easy to budget, easy to document, and easy to verify against regulatory requirements. The weakness stems from exactly the same source — the calendar knows nothing about the real condition. This has a twofold meaning:

• Over-maintenance: sometimes you replace a perfectly sound component just because the date arrived — a waste of a part, of labor and of unnecessary downtime.
• Under-maintenance: and sometimes a component starts to deteriorate two days after the periodic inspection — and breaks well before the next date, because the calendar did not see it coming.

Despite these limitations, time-based remains the right basis for the overwhelming majority of systems in a building — simply because in most cases the cost of continuous monitoring is not justified against the risk. The real question is where, and when, it is worth adding a predictive layer on top of it.

What predictive maintenance is — and how it "sees" a failure in advance

The idea at the base of predictive maintenance is simple: mechanical and electrical equipment almost never breaks all at once. Before the final failure it almost always "signals" — the vibration changes, the temperature rises, the electrical current distorts, the oil becomes contaminated. Predictive maintenance is the ability to read these subtle signs, identify a trend of deterioration, and act while it is still possible to plan an orderly treatment instead of absorbing an emergency shutdown. Instead of asking "how much time has passed since the inspection?", it directly measures the equipment's health.

In practice it relies on four main families of condition monitoring:

• Vibration analysis: the most prominent tool for rotating equipment — motors, pumps, blowers, chiller compressors. A bearing that begins to wear, a shaft that has gone out of balance or a loose coupling change the vibration signature long before the ear hears anything. A fixed vibration sensor can identify the change at a latent stage.
• Thermography: thermal imaging that locates hot spots — a loose electrical connection in a panel, a bearing heating up, an unbalanced load. This is one of the most mature and widespread monitoring techniques, and it is already today part of an annual inspection of electrical panels.
• Oil analysis: testing an oil sample from a motor, a compressor or a hydraulic system reveals metal particles, contaminants or chemical breakdown — direct evidence of internal wear that cannot be seen from the outside.
• Electrical monitoring (current / motor): analyzing the current and voltage signature of a motor exposes developing failures — a faulty winding, an alignment problem, an abnormal load — before they bring the motor down.

These are the "senses". What turns them from a spot check into a true predictive system is the layer above them: IoT sensors that measure continuously, transmit the data to a platform, and analytics that identify trends and anomalies over time. We expanded on the sensor layer itself in IoT sensors in buildings — what is monitored and why. When a system knows not only what the vibration is right now but how it has changed over the past three months, it can warn that "this bearing will reach failure in a few weeks" — and that is exactly the promise of predictive maintenance.

Advantages and disadvantages — the decision table

No approach "wins" in every situation. Here is an honest comparison of the three, on the axes that matter to whoever actually manages a building:

Breakdown maintenance

Advantage: zero up-front cost, zero planning. Disadvantage: the most expensive in practice, the most dangerous, the most legally exposed — the fault is discovered when it is already an incident. Suitable only for cheap, non-critical components whose failure creates no risk or secondary damage.

Preventive maintenance (time-based)

Advantage: simple, predictable, easy to budget and document, and suited to regulation. Disadvantage: "blind" to the real condition — it may treat too early (waste) or too late (failure between inspections). This is the right default for most systems in a building.

Predictive maintenance (condition-based)

Advantage: the most accurate — you treat exactly when needed, not before and not after; it catches failures a schedule would miss; it maximizes equipment life. Disadvantage: high setup cost (sensors, platform, integration), requires expertise to interpret the data, and is not economically justified for every piece of equipment. Suitable for systems where a combination of criticality, cost and access difficulty justifies the investment.

When predictive maintenance really pays — and when time-based is enough

This is the only practical question that matters. Predictive maintenance is not a goal in itself — it is a tool that justifies its cost only when directed at the right asset. Three conditions, each of which increases the worthwhileness, and in combination make the return clear:

• Criticality: the more a component's failure disables or endangers, the more continuous monitoring is worth. The chiller compressor that cools the entire building, the backup generator that is supposed to work at the moment of truth, the fire pump — all are cases where an unforeseen shutdown is far more expensive than the cost of the sensor.
• Cost of the component or the shutdown: equipment that is expensive to replace, or whose shutdown drags large secondary damage (business interruption, harm to tenants, cascading failure), justifies monitoring that prevents reaching catastrophic failure.
• Access difficulty: equipment located somewhere hard or expensive to reach for frequent inspection — on the roof, in a shaft, underground — benefits from a sensor that reports remotely instead of a repeated physical visit.

The systems where the three conditions most often meet in a building are clear: compressors and chillers in central air conditioning, generators and UPS, elevators, and main pumps (pressure boosting, suppression, drainage). These are natural candidates for a predictive layer.

And conversely — where is time-based simply enough? For the vast majority of systems. Replacing filters, visual inspections, lighting service, seal checks, cleaning — all are cheap, non-critical in a single failure, and convenient to inspect. Attaching an IoT sensor to an air filter is over-engineering: the cost of the monitoring exceeds the cost of simply replacing it by the calendar. The guiding rule: always start from time-based as the base, and add a predictive layer surgically — only on the few assets where it truly returns the investment. If you want to build that base correctly, the preventive-maintenance program generator gives an orderly starting point for the inspections and frequencies for each system.

Maintenance by property type — context changes the equation

The worthwhileness of predictive maintenance is not identical in every building. A data center, a hospital or an office tower with large central systems justifies a far broader predictive layer than a small residential building, where the equipment is simple, distributed and relatively cheap to replace. The more central, expensive and critical the systems, the higher condition-based climbs up the order of priority. We expanded on this fit in preventive maintenance by property type, and it is worth reading before deciding how deep to go into the predictive world.

The role of the BMS and IoT — the infrastructure that enables prediction

Predictive maintenance is not a collection of isolated sensors — it is a system. The vibration of the pump, the panel temperature, the motor current and the water flow rate become meaningful only when someone collects them continuously, keeps history, and identifies when the current value deviates from the normal trend. This is where the building management system (BMS) comes in: it is the platform that centralizes the data from the building's systems, presents it on one dashboard, and alerts on anomalies. We expanded on this infrastructure in the guide to building management systems (BMS).

It is important to temper expectations: a BMS or IoT sensors are a tool, not a magic solution. They provide data — but data without interpretation is noise. A system that generates a hundred alerts a day that no one reads is worse than a simple calendar that is actually carried out. The real value of the predictive layer is born only when there is a party that interprets the data, decides what requires action, and translates an alert into an actual treatment. Without this link, the investment in technology remains a pretty dashboard that changes nothing.

The critical point — mandatory inspections stay on the calendar, always

This is the most important point in the article, and it is easy to miss it in the enthusiasm for the technology: predictive maintenance adds a layer, it does not replace anything — and in particular it does not replace the statutory inspections. The certified elevator inspector's test, the annual functional test of fire detection and suppression submitted to the firefighting authority, the periodic electrical inspections, the disinfection of water reservoirs — all of these are defined by law and regulation on a fixed schedule, and they must be performed on time regardless of any sensor.

Even if a sophisticated vibration sensor reports that the elevator is "perfectly healthy", that does not exempt it from the certified inspector's test on its due date — neither legally nor from a safety standpoint. Predictive monitoring can warn early of a developing fault and save an emergency shutdown, but the legal approval rests on a defined periodic inspection, not on sensor data. Whoever imagines that sensors "make mandatory inspections redundant" creates for themselves a severe legal and insurance exposure. The predictive layer improves safety and planning — it is never a substitute for the statutory obligation.

The winning model — combination, not choice

If we distill the whole article into one sentence: do not choose between preventive and predictive maintenance — combine them in layers. The right structure for almost every building is three-layered:

1. Mandatory layer: all the statutory inspections, on a fixed calendar, without compromise — the legal and safety foundation.
2. Preventive layer (time-based): an orderly PPM program for all the other systems — the operational foundation that covers the vast majority of the building efficiently and at a reasonable cost.
3. Predictive layer (condition-based): condition monitoring focused only on the critical, expensive or hard-to-access assets — a surgical upgrade that returns the investment there, and only there, where it is justified.

The key is not the technology itself but the party that holds the three layers together: schedules the mandatory inspections, runs the preventive program, interprets the predictive data, and decides when an alert becomes an action. Without one central party, even the most sophisticated sensors turn into noise, and the most important inspections slip between the vendors.

At Domera Hub we build and operate this layered model for every property — a mandatory-inspections calendar, a complete preventive-maintenance program, and focused condition monitoring where it truly pays. See more in the knowledge hub — building systems, read preventive maintenance vs. breakdown maintenance, use the maintenance-program generator — or get a quote for a review of the maintenance status in your building.