Preventive vs Predictive vs Reactive Maintenance: Which Strategy Is Right for Your Operation?

Reactive maintenance is the practice of repairing or restoring an asset only after it has failed or degraded to a point where it cannot perform its function. Also called run-to-failure or corrective maintenance, it is not always wrong. For non-critical, low-cost assets with easy replacement and no production dependency it is often the most economical choice. The problem is that most Indian enterprises apply it universally — including to critical production assets, utility systems and safety-critical equipment where it is catastrophically expensive.

The true cost of reactive maintenance extends well beyond the repair bill:

• Direct repair cost is always higher under emergency conditions
• Unplanned downtime halts production and triggers penalty clauses in supply contracts
• Emergency spare parts procurement carries premium pricing and express logistics cost
• Primary failures cause secondary damage to connected systems
• The maintenance team stays permanently in fire-fighting mode with no capacity for planned work

Reactive maintenance costs 3 to 5 times more per event than an equivalent planned preventive job. For a facility spending 80 percent of its maintenance budget reactively, this is not a statistic — it is a financial crisis in slow motion. That said, reactive maintenance has a legitimate place in every maintenance strategy. The question is whether it is your choice or your default.

Preventive maintenance is maintenance performed on a fixed schedule based on time intervals or usage thresholds, regardless of the current condition of the asset. Oil change every 500 hours. Bearing inspection every 3 months. Filter replacement every 6 weeks. PM is schedule-driven, not condition-driven. It replaces the uncertainty of failure timing with the certainty of planned intervention — allowing parts to be ordered in advance, labour to be scheduled during low-production windows, and the maintenance team to work at full effectiveness rather than under emergency pressure.

Two scheduling methods exist:

• Time-based PM — simpler to implement, suitable for most asset classes, triggers on calendar intervals
• Usage-based PM — ties maintenance triggers to runtime hours, production cycles or kilometre readings, more precise for heavily utilised assets

A well-executed PM program typically reduces total maintenance cost by 25 to 35 percent within the first 12 months simply by reducing emergency callouts and secondary damage events. One important caution: scheduling PM too frequently adds cost without adding protection and erodes technician time. The right PM interval is determined by manufacturer recommendations, historical failure data and operational intensity — not by anxiety. In Sapphire, PM schedules are automated. The system raises the work order, assigns the technician and tracks completion without a manager having to remember a single date.

Predictive maintenance is condition-based maintenance that uses real-time or continuously collected asset performance data to identify degradation patterns and predict failure before it occurs. Where PM asks "is it time?" predictive maintenance asks "what is the asset actually telling us?" Data inputs include vibration analysis to detect bearing wear and imbalance, thermal imaging to identify overheating electrical components, oil analysis to detect contamination and metal particle accumulation, ultrasonic testing to identify pressure leaks and electrical arcing, and motor current signature analysis to detect winding faults.

The business case is strong: predictive maintenance when properly implemented reduces unplanned downtime by 30 to 50 percent and maintenance costs by a further 10 to 25 percent compared to a pure PM program. However, it requires investment in sensors and monitoring infrastructure, data science capability or an intelligent EAM platform to interpret the data, and an asset population large and critical enough to justify the investment.

For most Indian MSMEs, predictive maintenance is not the starting point — it is the destination. The starting point is getting PM right. Once PM is running cleanly and the asset data history is building in your EAM, the platform can begin identifying patterns and moving toward predictive capability. Predictive maintenance is not a tool for large corporations only. It is a capability that scales down to any organisation willing to build the data foundation first. For the full IoT and sensor integration story, see Article 41.

The matrix is not a ranking. It is a map. Every facility needs all three strategies — in the right proportions, for the right assets.

Not every asset deserves the same maintenance approach. The strategy should be determined by two axes: asset criticality — the impact of its failure on safety, production and compliance — and failure consequence — the cost and recovery time of a failure event.

• Low criticality, low failure consequence: reactive maintenance is acceptable and often optimal. Examples include office furniture, non-production lighting fixtures, backup storage equipment
• Medium criticality, moderate failure consequence: preventive maintenance is the right default. Examples include production support equipment, HVAC systems, secondary vehicles, warehouse handling equipment
• High criticality, high failure consequence: preventive maintenance as the baseline with predictive monitoring added where data infrastructure exists. Examples include primary production machinery, power generation and distribution systems, safety-critical equipment, medical devices

Most Indian facility managers intuitively know which assets are critical. The problem is not the knowledge — it is the absence of a system to execute the right strategy consistently for each asset class without relying on individual memory and experience.

A practical three-step framework for defining the right maintenance strategy mix for your facility:

• Step 1 — Audit your asset register. List every asset and assign it a criticality rating of high, medium or low based on its impact on production, safety and compliance. If you do not have an asset register, building one is the first action before any maintenance strategy discussion. See Article 11: Asset Register — Building Your Single Source of Truth
• Step 2 — Map current maintenance behaviour to each asset. Honestly assess whether each asset is currently being maintained reactively, preventively or predictively. Most facilities will find that reactive is the dominant mode across all criticality levels
• Step 3 — Shift the ratio deliberately. For every high criticality asset currently on reactive maintenance, create a PM schedule immediately. For medium criticality assets without PM schedules, add them in order of failure history. Do not try to implement predictive maintenance before PM is running cleanly — build the data foundation first

This is not a one-time exercise. The maintenance strategy mix should be reviewed every 12 months as assets age, as the portfolio changes and as the organisation's data maturity grows. In Sapphire this review is built into the platform through lifecycle reporting and maintenance cost analysis that surfaces which assets are consuming disproportionate reactive maintenance spend.