Run-to-failure is a strategy — just an expensive one

Most asset fleets are still maintained on one of two schedules: fixed intervals that service healthy equipment on a calendar, or run-to-failure that waits for the breakdown and pays for it in overtime and unplanned outage. Both spend a finite maintenance budget on the wrong question. The right one isn’t is this asset healthy? — it’s which asset’s failure would cost me the most, and can I see it coming?

From gut feel to a health index

The first move is to replace tribal knowledge with a health index by asset and asset type — a scored, current read on condition that ranks the fleet rather than describing one machine at a time. On top of that sits the predictive question: which assets are likely to fail or underperform, and when. That’s what turns maintenance from a calendar exercise into a triage list.

A failure has a blast radius

The cost of a failure is rarely just the asset. It’s the liability exposure if the asset suffers an accident, and the upstream and downstream assets taken out with it. A cheap component in the wrong position can strand an expensive chain; an expensive asset with plenty of redundancy may barely matter. Ranking assets by condition alone misses this — you have to weight each by what its failure actually puts at risk.

Optimizing the maintenance dollar

Once you can see condition, failure probability, and blast radius together, the budget question becomes answerable:

• How should maintenance dollars be allocated to buy down the most liability and exposure per dollar spent?
• What happens if we defer or accelerate a given job — the what-if scenario, not the fixed schedule?
• What do the historic maintenance trends by asset and asset type tell us about where spend is quietly recurring with no improvement?

Answer those and maintenance stops being a cost center run on habit and becomes a risk-weighted allocation: the next dollar goes to the asset whose failure would hurt most, not the one that happened to be next on the calendar or loudest in the control room.

Predictive maintenance was never about predicting every failure. It’s about ranking the fleet by the cost of failure and spending where it buys down the most risk — which is the difference between a maintenance budget that shrinks the problem and one that just keeps pace with it.