Engineering Myth #1: Does Preventative Maintenance Always Save Money?

Preventative maintenance has been gospel in industrial engineering for decades. The logic seems unassailable: fix things before they break, avoid costly downtime, save money. But does the maths actually support blanket PM programmes, or have we been following doctrine without checking the sums?

The assumption nobody questions

Walk into any plant and you'll find PM schedules pinned to notice boards, contractors booked months in advance, and maintenance windows blocking production time like clockwork. It's what responsible facilities do.

The assumption runs like this: equipment that receives regular attention lasts longer and fails less catastrophically. Therefore, scheduled maintenance must be cheaper than reactive repairs. The second part doesn't automatically follow from the first, but we've treated it as if it does.

When PM makes perfect sense

Some assets genuinely benefit from calendared intervention. Safety-critical systems top the list. If failure could injure someone or trigger a regulatory breach, you inspect and service on schedule regardless of cost-benefit analysis.

Beyond safety, certain equipment types respond well to preventative work:

• Components with predictable wear patterns (filters, belts, seals)
• Assets where minor degradation cascades into expensive secondary damage
• Systems where failure creates costs far exceeding the PM investment
• Equipment under warranty or insurance stipulating regular service

For these cases, the engineering and the economics align. You're not just avoiding breakdowns; you're avoiding situations where a £200 bearing failure destroys a £40,000 motor.

The hidden costs nobody totals

Here's where the myth starts to crack. Preventative maintenance isn't free, and it's not just the direct costs of labour and parts.

Every PM window takes equipment offline. For critical production assets, that means lost output. Even scheduled downtime costs money — sometimes more than the component you're servicing would cost to replace.

Then there's induced failure. Strip down a gearbox that's running fine and you introduce opportunities for error. Cross-threaded bolts, contamination during reassembly, seals damaged during installation. We've attended more than one "failed bearing" callout where the bearing had been replaced preventatively three weeks earlier.

Parts replaced on schedule rather than condition often have useful life remaining. You might change a contactor at 10,000 operations because the manual says so, even though inspection shows it's good for another 5,000. That's money in the skip.

The compliance complication

Here's where it intersects with regulatory requirements. Certain PM activities aren't optional. Your electrical installation must be inspected within defined periods. Safety systems need functional testing. Lifting equipment requires thorough examination.

But compliance-driven PM and cost-effective PM aren't the same thing. The legal minimum might be annual inspection; that doesn't mean quarterly strip-downs save you money. We see clients conflate the two, layering commercial PM programmes onto regulatory requirements until they're maintaining everything constantly.

The regulations are generally clearer than people think. BS 7671 doesn't mandate replacing contactors every year. It requires periodic inspection and testing at intervals appropriate to the installation. Condition matters more than calendar.

What condition monitoring actually tells you

The counterpoint to time-based PM is condition-based monitoring. Instead of changing components on a schedule, you monitor their actual state and intervene when data suggests it's necessary.

Vibration analysis on motors. Thermography on electrical panels. Oil analysis on hydraulic systems. Current signature analysis on drives. These techniques tell you what's actually happening rather than what the maintenance schedule assumes might be happening.

The technology has become affordable enough that monitoring costs less than premature replacement for many asset classes. A thermal camera pays for itself quickly if it prevents one unnecessary switchgear shutdown.

This isn't about abandoning all preventative work. It's about making PM decisions based on engineering evidence rather than inherited assumptions.

The reliability-centred middle ground

Reliability-centred maintenance takes a harder look at which assets actually benefit from scheduled intervention. It asks uncomfortable questions: what's the failure mode? What's the probability? What are the consequences?

Some components fail randomly regardless of PM. Electronics, for instance. A capacitor in a drive might last three years or fifteen. Stripping the drive annually won't change that, but it will cost you.

Other assets degrade predictably and benefit from intervention at specific intervals. These are your PM candidates.

The trick is telling the difference. That requires data, not assumptions. Failure histories, manufacturer guidance, condition monitoring trends. Engineering analysis rather than comfort blankets.

Running the numbers honestly

Before committing to a PM programme, calculate what you're actually spending:

• Direct costs: labour, parts, consumables, contractor fees
• Indirect costs: production loss, coordination overhead, logistics
• Induced costs: failures introduced during maintenance, parts replaced prematurely
• Opportunity costs: could the budget achieve more elsewhere

Then compare against your actual reactive maintenance costs for similar assets. Include failure frequency, repair costs, and yes, downtime.

Sometimes PM wins convincingly. Sometimes it's marginal. Sometimes you're spending three pounds to save one.

What to do next

If your PM programme has grown through accretion rather than analysis, it might be costing more than it saves. Lutrom Engineering works with clients to separate compliance requirements from commercial decisions, and to build maintenance strategies based on asset condition rather than calendar dates. We can review your electrical systems, identify what actually needs scheduled intervention, and show you where monitoring makes more sense than assumption.