Getting Started with Predictive Maintenance Analytics

Most facilities management organisations operate reactively — fixing things when they break. Predictive maintenance uses historical data to anticipate failures before they happen, reducing downtime, emergency callouts, and total maintenance cost.

The maintenance maturity ladder

1. Reactive — Fix it when it breaks
2. Preventive — Scheduled maintenance regardless of condition
3. Condition-based — Maintain based on measured condition
4. Predictive — Anticipate failures using data and models

What data do you need?

The minimum dataset for predictive maintenance analytics includes:

• Asset register with installation dates and asset types
• Work order history with completion dates and failure descriptions
• Ideally 2-3 years of history for statistically meaningful patterns

Key techniques

• MTBF analysis — Mean Time Between Failures shows average reliability per asset type
• Failure probability curves — Probability of failure based on age and usage
• Pattern detection — Seasonal, environmental, and usage-based failure patterns
• Risk scoring — Combine failure probability with consequence to prioritise action

Start with these tools

AssetArc provides two complementary tools:

• Failure Prediction — MTBF analysis and probability-of-failure modelling
• Failure Patterns — Detect seasonal, cyclical, and environmental patterns in failures

Import your work order history as a CSV and you'll have failure probabilities within minutes.