Reactive vs Preventive vs Predictive Maintenance

Every maintenance team runs some mix of three strategies: fix things when they break, service them on a schedule, or watch their condition and act before they fail. The debate over preventive vs predictive maintenance is often framed as old versus new, but in practice the right answer is a deliberate blend. This guide explains how each approach works, where each one pays off, and how to decide which assets deserve which strategy in a factory, warehouse, utility plant or large commercial building.

What is the difference between reactive, preventive and predictive maintenance?

Reactive maintenance (also called breakdown or run-to-failure maintenance) means no work is done until an asset stops working, or performs badly enough that someone complains. Preventive maintenance (PM) means servicing an asset at fixed intervals, either by calendar time or by usage such as running hours or cycles, regardless of its current condition. Predictive maintenance (PdM) means measuring the asset’s actual condition through vibration, temperature, oil quality, current draw and similar signals, and intervening only when the data shows degradation.

The three are not rungs on a ladder that every site must climb in order. They are tools, and each has its own cost profile and failure risk. The table below summarises the trade-offs.

When does reactive maintenance make sense?

Maintenance managers tend to treat the word “reactive” as an insult, yet waiting for failure is the cheapest sensible option for a certain class of equipment. Think of something that costs little, can be swapped in minutes, poses no hazard when it stops and leaves every other system running. Planning and logging inspections for it would cost more than the occasional replacement. A desk lamp, a store-room exhaust fan that nothing depends on, or a standard light fitting with spares already on the shelf all fit this description.

Trouble starts when nobody actually chose run-to-failure; it simply became how the site works. Signs that a team has slipped into unplanned firefighting include:

• Most work orders are raised by users rather than generated by the maintenance plan
• Technicians spend the shift moving from one urgent call to the next, with no time left for scheduled work
• The same assets keep appearing in the breakdown log with similar symptoms
• Spares are bought in a hurry, often from whichever vendor can deliver today
• Nobody can say with confidence when a critical asset was last serviced

If those patterns sound familiar, the issue is not that reactive maintenance exists. It is that critical equipment is being treated as if its failure did not matter.

What does a good preventive maintenance programme look like?

Preventive maintenance is the backbone of most industrial and commercial sites because it is simple to plan and easy to audit. A sound programme rests on three things: a complete asset register, sensible task frequencies and reliable proof that the work was actually done.

Start with the asset register

You cannot schedule what you have not listed. Each asset needs a unique ID, location, make and model, installation date, criticality rating and the responsible team or vendor. Physical tags, ideally QR codes, let technicians pull up the right history at the asset itself instead of searching through files in the maintenance office.

Set frequencies from evidence, not habit

Manufacturer manuals are the starting point, but operating conditions matter. A compressor in a dusty plant needs filter attention far more often than the manual’s default; a lightly used standby pump may need less. Review frequencies once you have a year or so of history: if a PM task never finds anything, consider stretching it; if failures occur between PMs, tighten the interval or change what the task checks.

Typical PM tasks on an industrial site include:

• Air handling units and other HVAC plant: filters washed or changed, belts re-tensioned, coils cleaned, following an HVAC preventive maintenance checklist
• Diesel generators: fuel, oil, coolant and battery verified, with a load test at set intervals
• Switchboards and panels: terminations scanned for heat and loose connections re-torqued
• Pumps, motors and compressors: greasing, shaft alignment and a look at the seals
• Fire pumps and hydrants: test runs at the intervals the local fire authority expects

The weak point of many PM programmes is not planning but execution. A schedule that lives in a spreadsheet, with technicians ticking paper sheets, makes it very hard to know whether a task was done properly, done partly, or done only on paper.

Is predictive maintenance worth the investment?

Predictive maintenance aims to service equipment at the right moment: late enough to avoid unnecessary work, early enough to avoid failure. Common condition-monitoring techniques include:

• Vibration analysis on motors, fans, pumps and gearboxes to detect imbalance, misalignment and bearing wear
• Infrared thermography on electrical panels, switchgear and connections to find hot spots
• Oil analysis on gearboxes, hydraulic systems and engines to track wear metals and contamination
• Ultrasonic testing for compressed air leaks, steam traps and early bearing faults
• Electrical signature monitoring of current and voltage to spot motor and load problems

These techniques work, but they bring costs: sensors or instruments, data platforms and, above all, people who can interpret readings and turn them into work orders. Predictive maintenance makes clear sense for assets where an unplanned failure is very expensive, dangerous or slow to recover from, such as a main production line motor, a critical chiller or a transformer. For a modest building with a handful of standard assets, a disciplined preventive programme with periodic thermography and vibration rounds by a specialist contractor often delivers most of the benefit at far lower cost.

How do you choose the right strategy for each asset?

The practical answer to preventive vs predictive maintenance is to decide asset by asset, using criticality as the guide. A simple approach that works for most sites:

1. List every maintainable asset in the register and group similar items together.
2. Rate each group on the consequence of failure: safety, production or service impact, repair cost and recovery time.
3. Rate the likelihood of failure using breakdown history, age and operating conditions.
4. Assign reactive maintenance to low-consequence groups where replacement is quick and cheap.
5. Assign preventive maintenance to groups whose wear is driven by time or usage and whose failures matter.
6. Assign condition monitoring to the small set of high-consequence assets where early warning saves real money or reduces real risk.
7. Review the assignment every year using actual breakdown and cost data.

Most sites end up with the bulk of their assets on preventive schedules, a small critical core under condition monitoring and a tail of minor items run to failure by design. That mix is healthy as long as it is deliberate, documented and revisited.

Which metrics show whether your maintenance strategy is working?

Whatever blend you choose, a short set of metrics will tell you whether it is working. Plain-language definitions of these terms are in the glossary.

To put numbers on the cost of downtime and see how MTTR, MTBF and uptime interact, try the maintenance downtime calculator. It is useful when building the case for moving an asset from reactive to preventive or predictive care.

Digitising the maintenance mix

All three strategies depend on the same foundation: a record of every asset, every work order and every intervention, with timestamps and evidence. Paper logs and spreadsheets break down quickly at scale, especially when work is spread across shifts, contractors and buildings.

A modern CMMS should let you schedule preventive work orders by calendar or meter, let users raise breakdown tickets by scanning a QR code on the asset, track response and repair times against SLAs and keep a full service history per asset. humblx Facility Maintenance covers this for the plant that keeps a building running, from chillers and generators to lifts, switchgear, pipework and the sewage treatment plant, and every closed job carries a photo as proof. For sites exploring sensor-led monitoring, condition data is most useful when it flows into that same work order trail rather than sitting in a separate dashboard.

Whichever strategy fits each asset, the goal is the same: fewer surprises, shorter outages and a maintenance team that spends most of its day on planned work. More approaches for plants and warehouses are on the industrial facilities page.