2026-08-25
"Maintenance" is one word hiding very different strategies with very different costs and outcomes. A facility that runs everything reactively pays more and breaks down more; one that understands the maintenance types and assigns its assets to them intelligently saves money and runs stably. Here are the four main types and how to choose between them.
Repair after failure: the AC stops → a request → a technician → a fix. It is usually the most expensive path (downtime + emergency repair + rushed parts), but it is acceptable for cheap, non-critical assets whose routine inspection isn't worth its cost — ordinary lighting, for example. The key is making it organized corrective: a numbered, documented, measured request — not a phone call that gets forgotten.
Scheduled inspection and service by time or usage before failure: monthly filter cleaning, monthly elevator checks, quarterly calibration. It is the backbone of any O&M contract, and most contracts require it explicitly. Its one weakness is inspecting things that didn't need it — which is why frequencies get tuned with data.
Intervening when a real indicator appears — abnormal vibration, high temperature, an odd pressure reading — instead of a blind schedule. It needs periodic measurements or sensors, and suits medium-to-critical equipment where routine inspection alone isn't enough.
Using operating data, history and analytics (sometimes AI) to predict failure before symptoms appear, and scheduling intervention at the optimal moment. It is the most mature and costly to set up, and makes sense for highly critical assets — but it is meaningless before you own the foundation: a clean digital record of every failure and every service, which a maintenance system builds day by day.
Everything above assumes one thing: you know your assets' history. Without a digital record — how often this AC failed, when it was last serviced, what it cost you — the choice remains guesswork. So the first step is always a system that documents every request and every service automatically, even if your team starts in the simplest form: WhatsApp requests turning into documented work orders.
MRFQ builds that foundation for you: photo-documented work orders, automatic preventive schedules, and a complete history per asset — with your team working over WhatsApp in any language, no training needed. Arabic-first and PDPL-aligned. Try it free for 14 days at mrfq.sa.
Four: corrective (after failure), preventive (scheduled before failure), condition-based (when an indicator appears), and predictive (data analysis forecasting failure).
Preventive runs on a fixed schedule regardless of equipment condition; predictive analyzes operating data to intervene only at the optimal time. Predictive is more precise but needs mature data and records.
No — it is the rational choice for cheap, non-critical assets. The mistake is relying on it alone for critical assets, or leaving it undocumented and unmeasured.
Two steps: a system that documents every request automatically (in MRFQ, WhatsApp requests become numbered work orders), then preventive schedules for critical assets — and within months you have a record to guide the rest.
A common benchmark is a scheduled-preventive majority (many facilities target 70–80%); a rising share of emergency corrective work signals assets being managed reactively.