Overview: There is no universal ratio. The right maintenance mix is set asset by asset, using criticality (what happens if it fails) and predictability (whether failure gives warning). Critical, predictable assets suit predictive maintenance; stable, wear-based ones suit preventive; and low-impact, cheap-to-replace items are often left to run to failure.
Ask a good TPM consultancy how much preventive maintenance a plant should do, and the honest answer is rarely a single number. The right mix of preventive, predictive and breakdown maintenance depends on the asset in front of you, not on a company-wide rule applied to every machine. A conveyor motor, a safety valve and a spare hand pump each deserve a different level of attention. Getting that balance wrong is expensive in both directions: over-maintain and you burn labour and spares on machines that do not need it, under-maintain and you pay in unplanned downtime.
What do preventive, predictive and breakdown maintenance actually mean?
Breakdown maintenance, also called reactive or run-to-failure, means you fix the equipment after it stops. No inspection schedule, no sensors, just repair on failure. It sounds careless, but for the right asset it is a deliberate and cost-effective choice.
Preventive maintenance (PM) is time or usage based. You service the machine at set intervals, say every 500 running hours or once a quarter, whether or not it shows symptoms. It works best where wear follows a predictable pattern.
Predictive maintenance (PdM) uses condition data, things like vibration, temperature, oil analysis or current draw, to spot a developing fault and act just before failure. It needs sensors, some analytics and people who can read the signals, so it carries a higher setup cost.
Why not just do preventive maintenance on everything?
It is because the blanket PM quietly wastes money. Servicing a low-value pump on a fixed calendar when it rarely fails ties up technicians who are needed elsewhere. Worse, every intervention introduces a small risk of induced failure, a gasket seated wrong or a bolt over-torqued. Studies of failure behaviour have long shown that many components do not wear out on a neat schedule, so a fixed interval does not reduce their risk. This is exactly why a maintenance strategy has to be chosen per asset rather than imposed as one policy for the whole plant.
How does a TPM consultancy decide the right strategy for each asset?
Most reviews start with a criticality assessment. The team scores each asset on the consequence of failure across a few dimensions, then on how likely and how predictable that failure is.
Consequence usually looks at:
- Safety and environmental risk
- Production loss if the asset goes down
- Quality impact on the product
- Repair cost and spare availability
Failure behaviour asks a second question: does this asset give warning before it fails, or does it go from healthy to broken with no notice? An asset that degrades slowly and measurably is a good candidate for predictive monitoring. One that fails randomly and cheaply is often better left to run to failure.
Put the two together and a clear pattern emerges:
| Asset profile | Sensible default strategy |
|---|---|
| High criticality, gives measurable warning | Predictive (condition monitoring) |
| High criticality, but failure is sudden | Preventive, often with redundancy built in |
| Moderate criticality, wear is predictable | Preventive on interval |
| Low criticality, cheap and quick to replace | Breakdown / run-to-failure |
This is where TPM and reliability-centred maintenance (RCM) work well together. RCM is usually reserved for the critical and safety-related equipment, where a detailed failure-mode analysis justifies each task. TPM then spreads basic care across the whole plant through autonomous maintenance, handing routine cleaning, lubrication and inspection to the operators who run the machines every day.
What does a healthy maintenance mix look like?
There is no single correct split, but many high-performing plants aim to shift work away from firefighting and towards planned activity. As a rough direction of travel, a mature operation keeps reactive work to a small minority of total maintenance, leans on preventive work for the bulk of its stable assets, and applies predictive monitoring to its most critical machines. The exact numbers matter less than the trend. If most of your week is an unplanned breakdown response, the mix is off, whatever the asset list says.
How do you measure whether the mix is working?
You do not judge a strategy by how busy the maintenance team looks. You judge it by the numbers that reflect equipment health and stability:
- OEE (Overall Equipment Effectiveness): the headline TPM metric, combining availability, performance and quality. Rising OEE usually means the mix is holding.
- MTBF (mean time between failures): longer is better and signals fewer surprises.
- MTTR (mean time to repair): shorter means faster recovery when something does fail.
- Unplanned downtime and the ratio of planned to unplanned work.
TPM frames these losses through the six big losses: breakdowns, setup and adjustment, minor stops, reduced speed, defects and start-up losses. They map neatly onto the three parts of OEE. Tracking them tells you not just that a machine failed, but which type of loss is eating your capacity.
Common pitfalls to avoid
- Treating the mix as fixed. Assets age, production volumes change and failure data matures, so strategies should be reassigned over time.
- Buying predictive sensors before fixing basics like lubrication and cleaning. Autonomous maintenance comes first.
- Ignoring operators. The people running the equipment often spot the early symptoms before any dashboard does.
Conclusion
The right mix of preventive, predictive and breakdown maintenance is not a formula you copy from another plant. It is a set of deliberate, asset-by-asset decisions built on criticality, failure predictability and cost, then checked against OEE and reliability data and adjusted as things change. Done well, it moves a plant from constant firefighting to planned, calm and predictable maintenance, which is the whole point of a TPM programme.
Frequently asked questions
Is breakdown maintenance ever the right choice?
Yes. For low-critical assets that are cheap and quick to replace, and where failure does not affect safety, quality or production, running to failure is often the most economical option. Planning heavy maintenance for such items usually costs more than the failures do.
What is the difference between preventive and predictive maintenance?
Preventive maintenance is scheduled by time or usage, regardless of the machine’s actual condition. Predictive maintenance uses condition data such as vibration or temperature to act only when a fault is developing, so you avoid both surprise failures and unnecessary servicing.
How do I know which assets need predictive maintenance?
Look for assets that are both critical and give measurable warning before failure. If a machine is expensive to lose and its condition can be monitored with sensors, it is a strong candidate for a predictive approach.
Where does TPM fit alongside RCM?
RCM is typically applied to critical and safety-related equipment through detailed failure-mode analysis. TPM works across the whole plant, using autonomous maintenance to give operators day-to-day care of their machines. Most mature programmes use both together.
Which metrics show the mix is working?
OEE, MTBF, MTTR, unplanned downtime and the ratio of planned to unplanned work. Improving OEE and MTBF alongside falling unplanned downtime is a good sign the strategy is holding.
How often should the maintenance mix be reviewed?
Treat it as a living plan rather than a one-time decision. Review it as assets age, production changes and your failure data grows, then move assets between strategies as the evidence suggests.