Almost every technical organisation wrestles with the same question: do you maintain an installation only once it breaks, or intervene beforehand to prevent failures? The answer is rarely "always one or the other". Preventive and corrective maintenance are not opposites you pick between, but building blocks of a single maintenance strategy. Per asset you determine which approach yields the lowest total cost and the lowest risk. This article explains the definitions, pros and cons, and a practical method for choosing.
What is corrective maintenance?
Corrective maintenance (also called reactive maintenance) is carried out after a failure or defect has occurred. You repair or replace a component only when it no longer functions properly. Within corrective maintenance there is a distinction between planned and unplanned variants: a broken lamp you replace on your next round is planned-corrective, whereas a pump that fails unexpectedly and halts production is unplanned-corrective.
The approach is popular because it seems simple: you do nothing until something goes wrong and you use the full service life of a part. For non-critical assets with low replacement costs, that is often a sensible choice.
What is preventive maintenance?
Preventive maintenance, by contrast, is carried out in advance, on a fixed schedule, to prevent failures before they occur. Think of periodic inspections, lubrication, replacing wear parts and calibrating at set intervals. The goal is to raise reliability and reduce unplanned downtime.
Preventive maintenance is usually planned on time (for example, every quarter) or on usage. With usage-based maintenance you base the interval on actual use, such as running hours or number of cycles, rather than on the calendar. That aligns better with how intensively an asset is actually loaded. Scheduling these tasks efficiently is done through work orders and maintenance planning, typically supported by software.
Weighing the pros and cons
Corrective maintenance has low preparation costs and no unnecessary interventions: you only replace what has genuinely worn out. The drawback is unpredictability. Failures occur at inconvenient moments, often with consequential damage, rush parts and longer downtime. The cost per incident is higher and the safety and production risks greater.
Preventive maintenance offers predictability, less unplanned downtime and a longer asset life. The flip side is that you sometimes perform maintenance that was not yet needed, leading to unnecessary labour, material consumption and even introduced faults (opening up a well-running machine carries its own risks). Too frequent maintenance is waste; too little undermines the whole idea.
The maintenance spectrum
In practice these strategies form a spectrum from reactive to increasingly smart intervention:
- Reactive (corrective): repair after failure. Suited to cheap, non-critical parts.
- Preventive: intervene at fixed intervals (time or usage) to get ahead of failure.
- Condition-based maintenance: intervene based on actual condition, measured via inspections or sensors (vibration, temperature, oil analysis). Usage-based/running-hours maintenance is a simple, practical variant of this: usage is an indirect indicator of wear.
- Predictive maintenance: use data and models to predict when failure is likely, so you intervene at exactly the right moment. This requires measurement data, history and analysis, and is not cost-effective for every asset.
Important: usage-based maintenance is not "AI". You simply plan on actual use, which is a level-headed, robust form of condition-based thinking.
How do you choose per asset?
There is no universally correct strategy; the choice depends on the asset. For each installation or component, weigh the following factors:
- Criticality: what happens if this asset fails? Does production stop, does a safety or environmental risk arise, or is it merely a nuisance?
- Failure behaviour: does the part wear predictably over time or usage, or does it fail randomly? Wear-sensitive parts lend themselves to preventive or usage-based maintenance; random failures often do not.
- Cost of downtime: what does an hour of unplanned downtime cost, including consequential damage, rush logistics and lost production?
- Maintenance costs: how expensive and labour-intensive is preventive intervention compared with repairing after failure?
A structured way to analyse failure behaviour and criticality is reliability and FMECA. This lets you determine, with sound reasoning, which failure modes matter and which strategy fits.
A practical approach to setting your mix
A workable maintenance strategy is almost always a mix. Here is how to reach a well-founded balance:
- Inventory your assets and map their criticality. Separate critical from non-critical installations.
- Deliberately choose corrective for non-critical, cheap parts with limited consequential damage. Being deliberately reactive is legitimate; it saves unnecessary costs.
- Choose preventive or usage-based for wear-sensitive, critical assets. Base the interval on time or, where it makes sense, on running-hours-based maintenance.
- Consider condition-based or predictive maintenance only where the cost of downtime is high and measurement data is available or feasible.
- Measure and adjust. Record failures, lead times and costs, and adapt intervals based on what the data shows.
Recording and analysing this data typically happens in a maintenance system. Read more about what a CMMS is and how such a system helps with planning and record-keeping.
Conclusion
Preventive and corrective maintenance are not a matter of choosing one against the other. They are two points on a spectrum running from reactive repair to predictive intervention. The best maintenance strategy is a considered mix, determined per asset on the basis of criticality, failure behaviour and the cost of downtime. Start with insight into your assets, choose deliberately, and adjust based on data. That way you keep your installations reliable at the lowest possible total cost.
Frequently asked questions
What is the difference between preventive and corrective maintenance?
Corrective maintenance is carried out after a failure has occurred: you repair or replace only once something breaks. Preventive maintenance is carried out in advance, on a fixed schedule by time or usage, to prevent failures. Corrective uses the full service life but is unpredictable; preventive gives predictability but can lead to unnecessary work.
Is preventive maintenance always better than corrective?
No. For non-critical, cheap parts with limited consequential damage, deliberately choosing corrective maintenance is often the cheapest option. Preventive pays off mainly for critical, wear-sensitive assets where unplanned downtime is expensive or dangerous. The right choice differs per asset.
What is usage-based (running-hours) maintenance?
With usage-based maintenance you schedule tasks based on actual use, such as running hours or the number of cycles, rather than on the calendar. It is a practical variant of condition-based thinking: usage serves as an indirect measure of wear. It is emphatically not AI, but level-headed planning on measurable use.
What is the difference between condition-based and predictive maintenance?
With condition-based maintenance you intervene based on the measured current state, for example vibration or temperature. Predictive maintenance goes a step further: using data and models you predict when failure is likely, so you intervene at exactly the right time. Predictive requires more measurement data and analysis and is not cost-effective for every asset.