The equipment that fails on the job site was going to fail.

The excavator goes down at 7am on a Thursday. The operator calls it in. The parts are three days out. The crew stands around for two hours until something else can be arranged. The subcontractor who was supposed to arrive Friday is now pushed to the following week. The GC is on the phone. The job is behind schedule before the week is halfway done.

The maintenance record, when someone looks at it, shows the machine was 340 hours past its hydraulic service interval. The failure was not unpredictable. It was predicted - by a service schedule that nobody was tracking against actual usage hours. The machine had no way to surface a flag, and nobody was looking.

Why unplanned downtime costs so much more than maintenance

A scheduled hydraulic service on an excavator costs roughly $400 to $800 in parts and labor, plus four to six hours of planned downtime during a low-utilization window. An unplanned hydraulic failure on an active job site costs the same parts plus emergency labor rates, plus crane rental if the machine needs to be extracted, plus crew idle time while work stops, plus schedule disruption that ripples through everything downstream. The equipment repair bill might be similar. The total operational cost of the failure is three to four times higher - and that's before any penalty clauses with the GC.

The pattern repeats across every type of field equipment: vehicles, lifts, generators, compressors, specialized tools. The difference between planned and unplanned failure isn't usually the repair cost. It's the context: planned maintenance happens on your schedule, unplanned failure happens on the equipment's schedule, which is invariably the worst possible time.

"Maintenance deferred is downtime chosen - you just don't get to pick when."

What preventive maintenance actually requires to work

The failure mode of most preventive maintenance programs isn't the intent - it's the tracking. Every operation that runs equipment has some version of a maintenance schedule. The challenge is keeping that schedule current against actual usage, not calendar assumptions. A generator that's used 20 hours a week and a generator used 80 hours a week have the same oil life on the manufacturer's calendar schedule, but completely different real-world service windows.

Usage-based maintenance tracking requires knowing how much each piece of equipment is actually running - which means capturing usage hours in real time, not reconstructing them from invoices or operator memory. When usage data flows automatically from equipment telematics or daily job logs, the maintenance schedule stays calibrated to reality. When it doesn't, you're running a calendar-based system that drifts further from accuracy every day of use.

The three maintenance categories that matter in the field

Preventive maintenance covers scheduled intervals - oil changes, filter replacements, fluid checks, belt inspections. These are the highest-leverage maintenance activities because they cost almost nothing relative to the failures they prevent. Condition-based maintenance covers items that are inspected and serviced based on observed condition rather than fixed intervals - brake wear, tire condition, hydraulic seals. Predictive maintenance covers items where performance data can flag impending failure before it happens - unusual operating temperatures, pressure anomalies, increased fuel consumption that indicates engine stress.

Most field operations handle preventive maintenance erratically and condition-based maintenance reactively. Predictive maintenance is largely unavailable without telematics systems that most field equipment operations haven't deployed. The highest-impact starting point for most operations is simply making preventive maintenance intervals visible, assigning accountability, and building alerts that surface when a piece of equipment is approaching a service threshold - before it crosses it.

Building a maintenance system that field crews actually use

Maintenance programs fail in the field for the same reason apps fail in the field: they require the person who is least invested in the system to do the most work to maintain it. A maintenance log that lives in a spreadsheet in the office is not a maintenance system - it's a record that gets updated after something breaks.

Field-usable maintenance systems have three characteristics. They surface information proactively rather than requiring someone to go looking. They are updated by the people closest to the equipment - operators and techs - using tools that take less than 60 seconds. And they create accountability at the equipment level rather than the job level, so a machine's service history travels with the machine, not with whatever crew last used it.