Take any commercial operation’s maintenance ledger and sort it into two piles: work that was planned, and work that was a reaction to something breaking. The ratio between those piles is the most honest sentence the operation can say about itself. And in most operations, the reactive pile dominates — not because the team is careless, but because the scheduling model guarantees it.

Watch the model run for a quarter and the mechanism becomes obvious.

Deferred prevention manufactures the very emergencies that crowd out prevention. The loop is closed and self-feeding: reactive load defers PMs, deferred PMs generate reactive load.

The current model, step by step

  • Week one: the PM calendar says the rooftop units get serviced. But two corrective work orders came in Monday — a tenant complaint and a leak — and correctives are loud while PMs are silent. The technician hours go to the loud work. The PM moves to next week.
  • Week three: next week never came, because more correctives arrived. They always do. The PM is now a month behind, and here’s the mechanism most operations never name: every deferred PM slightly raises the probability of the next corrective. The filter not changed becomes the coil that freezes. The belt not inspected becomes the bearing failure. Deferred prevention manufactures the very emergencies that crowd out prevention.
  • Quarter’s end: the ledger shows the result. Reactive work consumed the budget at emergency pricing — after-hours rates, expedited parts, the coordination costs nobody tracks — while PM completion sits somewhere south of 70%, a number nobody reports because nobody is measuring it. The loop is closed and self-feeding: reactive load defers PMs, deferred PMs generate reactive load. This is why the ratio is structural, not moral. The team didn’t fail. The scheduling model did exactly what its design implies.

McKinsey’s research quantifies the exit price: operations that move to predictive, data-driven maintenance see cost reductions of 30–45% and downtime cut substantially. Read that as the size of the tax the reactive loop charges.

The same quarter, with the scheduling model replaced

Now rerun the quarter with two structural changes — not more discipline, different mechanics.

First: PM work orders create, schedule, and escalate themselves. The rooftop unit service doesn’t wait for someone to remember it amid the correctives; it exists in the queue with a deadline, it books vendor capacity in advance, and if it slips, it escalates to a human as an exception requiring a decision — visible, owned, documented — rather than dissolving silently into “next week.” The PM stops competing for attention, because attention is no longer what schedules it. That’s the foundation any serious commercial building maintenance program rests on: prevention that doesn’t depend on a quiet week.

Second: the calendar stops being the only trigger. Sensors on critical assets — vibration, temperature, current draw, runtime hours — generate condition-based work orders when readings drift, which catches the failures the calendar can’t see and skips the service the asset doesn’t yet need. Calendar PM is a guess averaged across all assets; condition data is the asset telling you itself.

STAGE 1 Automated Escalation

PMs stop competing for human memory. They self-schedule and escalate as exceptions when deadlines slip.

STAGE 2 Condition Triggers

Asset sensors generate work orders based on actual performance drift, skipping unnecessary calendar PMs and catching blind failures.

STAGE 3 The Reverse Loop

Completed prevention reduces reactive load, which in turn protects the team’s capacity for further prevention.

By the new quarter’s end, the ledger reads differently in a specific, mechanical way: the reactive pile shrinks because fewer failures occur (PMs actually happened) and because drifting assets got caught at the cheap stage (sensors flagged them). The remaining reactive work is genuine surprise — and it’s affordable, because it’s no longer the whole budget. The loop now runs in reverse: completed prevention reduces reactive load, which protects capacity for prevention.

The ratio as a leading indicator

Here’s what changes for the operator personally: the reactive-to-preventive ratio becomes a number you manage instead of a verdict you receive. Watch it monthly per site and it predicts cost trajectory two quarters out — a site drifting reactive is a site about to get expensive, visible while there’s still time to act. Operators told us in interview after interview that their budget “didn’t exist” — that it was a record of emergencies. The semi-autonomous scheduling layer in Sweven FM was built to attack exactly that loop at its mechanism: the silent deferral.

The Ledger Test

Pull your last quarter’s ledger and sort it into the two piles. Whatever the ratio is — was it chosen, or did it just happen?

Sources: