Preventive Maintenance: From Fixing Breakdowns to a Schedule That Prevents Them

Preventive Maintenance: From Fixing Breakdowns to a Schedule That Prevents Them

Part three sorted out the visits: who goes, when, in what order, and with which part in the van. But everything built so far starts at the same point: something broke. The request arrives, we log it, dispatch it, execute it and close it. That is sound operations, but it is a race that never ends, because the entire source of your workload sits outside your control.

Preventive maintenance is the move from waiting for the failure to scheduling work ahead of it. In practice it is the difference between a company that sells its technicians' hours and one that sells the continuity of its client's assets. The gap between the two is not intent. It is three things: an asset register, plans anchored to that register, and checklists that produce data you can go back to.

The asset register: the foundation nothing works without

There is no preventive maintenance without a precise answer to a simple question: what exactly are you maintaining, and where is it? The asset register is that answer, and it is not a spreadsheet refreshed once a year. It is the record every plan, every visit, every spare part and every cost hangs on.

The minimum for each asset:

• A permanent reference number with a physical barcode or QR tag on the asset itself, so a technician opens the right record with one scan instead of searching by name and picking the wrong unit.
• A location precise enough to reach it: building, floor, room or roof — not just the client's name.
• Type, model, serial number and capacity, installation and warranty expiry dates, and the supplier who installed it.
• The contract the asset falls under, because the same equipment type may sit under a different commitment at another client.
• Current status, plus a full history of visits, failures and replaced parts.

The payoff shows up months later, not on day one. Once an asset has a history, you can answer questions that used to be settled on instinct. Do we repair this unit for the fourth time or replace it? Which model costs us more in maintenance despite the lower purchase price? Which site consumes more crew time than it pays for?

One implementation warning: do not try to capture everything at once. That is the most common way these projects die during data entry. Start with critical assets only — the ones whose failure stops the client's operation, breaches a contractual commitment, or costs a lot to replace — then grow the register visit by visit as technicians work.

Plans and checklists: what the technician actually does

A preventive plan is not a sentence like routine maintenance every three months. It is a defined set of tasks, at a defined frequency, anchored to a defined class of assets, generating a work order automatically on its due date with no human intervention and no reliance on anyone's memory.

The checklist is the content of the visit. The difference between a useful checklist and a worthless one is that the useful one produces data: instead of unit checked, it records the pressure reading, the temperature differential, the current draw, the filter condition — as numbers in fields, not free text. Numbers accumulate and show a trend; free text stays paper.

The single most important line in the whole checklist is the out-of-range one. When a technician records a reading outside accepted limits, the system must open a corrective work order linked to that same visit and that same asset. That link is precisely what turns checklists from a formality into a system: the observation does not sleep on a sheet in the technician's van, it becomes assigned work with an owner and a date.

Likewise, parts consumed on a preventive visit are issued against that same work order, so the cost lands on the right contract instead of disappearing into general expenses. Part six covers that side in detail.

When to trigger the visit: calendar versus runtime and meter readings

There are two ways to trigger a plan, and confusing them is a common reason maintenance costs money without preventing failures.

Calendar-based triggering: monthly, quarterly, annually. Simple, easy to plan, what contracts usually specify, and the only acceptable basis for some statutory inspections. Its weakness is that it does not know how hard the asset actually worked: an air-conditioning unit running at full load in an exposed site and an identical one running two hours a day get the same schedule, so you maintain what does not need it and neglect what has burned through its expected life.

Runtime or meter-based triggering: generator hour meters, vehicle mileage, cycle counts, differential pressure across a filter. This tracks the physical reality of the asset, cuts unnecessary visits and pulls forward ones that would have come too late.

The practical answer is a mix, not a choice: calendar for statutory and contractual inspections that must happen regardless of usage, readings for assets whose consumption varies sharply between sites. And runtime triggering only works if you have a disciplined reading source — either a meter read automatically, or a documented rule that the technician records the reading on every visit. Without that you have built a schedule on data that never arrives.

The times the contract promised, not the ones in the manager's head

The most repeated dispute in maintenance contracts is rarely about quality of work. It is about time. Settling it requires separating two numbers people routinely blur: response time, the gap between the request being logged and the technician arriving or work starting, and resolution time, the gap until the asset is back in service. A serious contract separates them and ties each to the severity of the fault.

Operationally, that means the contract's commitments must live in the system as rules that run, not as a file in a folder nobody opens. Each contract has fault categories, each category has a response time and a resolution time, and the system warns before the breach rather than a week after it.

The part that always gets forgotten is exceptions. If the delay was caused by waiting on a part the client supplies, by site access being denied, or by the client asking to postpone, the clock must stop with a documented, logged reason. Without that documentation you will walk into a renewal meeting holding numbers that disagree with the client's, and nothing to prove your version. And the monthly compliance report that falls out of this discipline is the same document that accompanies your payment application — which is the next part.

Condition-based and predictive maintenance: what is realistic and what is vendor talk

Both terms are sold heavily to contracting and maintenance companies today, and they deserve a clear separation.

Condition-based maintenance means the decision to intervene comes from an actual measurement of the asset's state: thermal imaging of electrical panels, vibration readings on pumps and motors, oil analysis, differential pressure. This is realistic today for a Saudi SME, because it does not require a full sensor infrastructure. It requires handheld instruments, a trained technician, reading fields in the checklist, and acceptable limits defined in advance. On high-value critical assets the effect shows quickly.

Predictive maintenance means a model that forecasts failure from an asset's history and continuous readings. It is not impossible, but it needs inputs most companies do not have yet: failure history classified with discipline over years, enough continuous readings, and a large enough population of similar assets to form a pattern. Running a prediction model on incomplete data produces false alarms; two months later everyone ignores the alerts, and you have paid for a system that cost your crews their trust in warnings.

The correct order is clear: a clean asset register, then disciplined logging of visits and failures for at least a year, then condition measurements on critical assets — and only then does talk of prediction mean anything. Anyone selling you step four before step one is selling a dashboard, not a system.

The Origami view

When we start with a maintenance company, we do not open with what system do you want. We open with one question: how many preventive visits were planned last month, and how many were actually completed? In most cases there is no numeric answer, and that itself is the answer. The gap between planned and completed is the first indicator we build, because it exposes whether the preventive plan is real or just a contract clause remembered at renewal.

As a technology company serving the Saudi contracting and maintenance sector, the rules we apply in delivery are these: registration starts with critical assets, not with everything; every out-of-range checklist line must automatically generate a corrective work order, otherwise the checklist is a formality; and contract commitments are written into the system as rules that warn before the breach. Prediction we do not switch on before there is data worth basing a decision on, because an alert technicians do not trust is worse than no alert at all.

Where this leaves you

Preventive maintenance is not a schedule pinned to a wall. It is three connected layers: defined assets, plans that generate their own work orders, and checklists that produce accumulating numbers. Once those layers exist, a growing share of your crews' work becomes planned rather than sudden, and that alone cuts emergency visits and frees up time that used to burn in firefighting.

That leaves the question that decides survival rather than efficiency: is any of this profitable? Part five covers cost control and payment applications — charging hours, parts and subcontractors against a job number, comparing them against the contract value, why change orders decide the argument later, and how to manage the gap between doing the work and being paid for it.

Sources

• Saudi Building Code — requirements for operating and maintaining mechanical, electrical and protection systems in buildings.
• General Directorate of Civil Defence — safety requirements and periodic maintenance and documentation of alarm and firefighting systems.
• Zakat, Tax and Customs Authority — e-invoicing requirements connected to maintenance contracts and completed visits.
• Saudi Vision 2030 — the national direction on raising operations and maintenance efficiency and extending asset life.