PM Interval Fundamentals

When Run-to-Failure Is the Right Call: Choosing a Strategy per Failure Mode

By Rovaryn Digital· November 20, 2026· 8 min read

Why a run to failure maintenance strategy is a decision, not a default

The budget meeting is two weeks away, and the PM list has grown again. Your lead technician has just handed you the hours estimate, and it exceeds what the crew can deliver. Reading down the list, you notice tasks on a shop-floor exhaust fan, a set of break-room appliances, and a bank of indicator lamps. Each one costs a few minutes of labor and a few dollars of parts. None of them would stop production if it failed. Yet they sit on the same calendar as the compressor and the main line drive, competing for the same technician hours.

The uncomfortable question is whether some of those tasks should exist at all. Many managers hesitate because "run to failure" sounds like neglect. It is not, provided it is chosen on purpose, for a specific failure mode, with the numbers written down. A deliberate choice has a stated reason and a review date, while neglect has neither.

This article sets out the criteria that make run-to-failure the right call, how to compare it with PM cost one failure mode at a time, and how to record the decision so it survives a change of staff.

Choose per failure mode, not per machine

Strategy selection fails when it is made at the asset level. A single pump has several ways to fail: a seal leaks, a bearing seizes, a coupling wears, a motor winding shorts. Each has its own consequence and its own detectability. A sensible plan may inspect the bearing, replace the seal on a calendar, and let a redundant indicator lamp burn out.

Treating the whole machine as "PM" or "run-to-failure" forces you to over-maintain the cheap parts or under-protect the critical ones. The unit of analysis is therefore a pair: a component and the way it fails. Three questions decide the strategy for each pair.

  1. What happens when it fails? Consider safety, environmental release, product quality, damage to neighboring parts, and lost production.
  2. Can you see it coming? If a cheap inspection or a condition reading gives enough warning, you can act before failure without a fixed calendar task.
  3. What does prevention cost compared with the failure? This is the arithmetic, covered below.

For the wider context on how PM frequencies are built in the first place, the preventive maintenance interval and cost guide is the place to start.

The criteria that justify planned run-to-failure

A failure mode is a candidate for deliberate run-to-failure when most of the following hold. No single test settles it, and the safety test overrides all the others.

  • No safety or regulatory consequence. If failure could injure someone, release a hazardous material, or breach a permit condition, it is not a candidate. Confirm any statutory inspection duties with the relevant regulator rather than relying on a plant-level judgment.
  • Low consequence for production and quality. The failure does not stop the line, scrap product, or damage other equipment.
  • Fast, cheap, planned recovery. A spare is on the shelf or available from a local supplier the same day, and one technician can swap it quickly. If the repair needs a long-lead part or a specialist, the failure is no longer cheap.
  • Redundancy or a workaround. A standby unit, a bypass, or a manual fallback keeps the process running during the repair.
  • No useful wear pattern. Some components fail at random, with no wear-out trend that a PM could catch. Replacing them early on a calendar throws away remaining life and gives no protection, because the next failure is as likely to come tomorrow as at the end of the interval.
  • Low cost of the item itself. When the part costs less than the labor to inspect it, inspection alone can exceed the cost of waiting.

Failure data sharpens each of these tests. If your records show a component failing at irregular times with no pattern, that supports run-to-failure. If they show failures clustering at a consistent age, a calendar or usage-based task may pay for itself. The article on using failure history to set PM intervals shows how to read that pattern, and the comparison of calendar-based vs usage-based maintenance helps once you decide a task is warranted.

Comparing PM cost with run-to-failure cost for one failure mode

Once a failure mode passes the screening questions, compare the two regimes on annual cost. Run-to-failure has no PM line, but it carries failures. The comparison for each failure mode is:

Annual PM cost = (PM labor hours × labor rate + PM parts cost) × PM events per year

Annual run-to-failure cost = expected failures per year × cost per failure

Cost per failure should include repair labor, the replacement part, any damage to neighboring parts, and the cost of lost output during the repair. PM events per year is 365 ÷ the interval in days. Expected failures per year should come from your own failure history for that specific failure mode, not from a general estimate.

Here is a worked example with round, illustrative inputs. Take an auxiliary ventilation fan with a redundant unit alongside it.

  • PM task: 0.5 labor hours at $40 per hour, plus $5 in parts, every 90 days.
  • PM events per year: 365 ÷ 90 ≈ 4.06.
  • Annual PM cost: (0.5 × $40 + $5) × 4.06 = $25 × 4.06 ≈ $101.
  • Failure history: one failure in four years, so 0.25 failures per year.
  • Cost per failure: 2 labor hours at $40, plus a $150 replacement motor, plus no lost output because the standby unit carries the load. That totals $230.
  • Annual run-to-failure cost: 0.25 × $230 = $57.50.

On these inputs, run-to-failure costs less than PM for this failure mode. Change one input and the verdict can flip. If the standby unit were absent and each failure stopped a line, the cost per failure would include lost output, and the PM task would likely win. The value of the method is that you can see which input drives the answer.

Be careful with the cost of lost output. Industry estimates of downtime cost vary widely, and a figure borrowed from another plant says little about yours. Calculate your own by multiplying the hours of lost production by the fully loaded cost of the capacity you lose. The downtime cost calculation guide walks through the method.

Reactive work generally costs more per event than planned work, because of premium labor, expedited parts, and secondary damage. We do not quote a published multiplier here. The right ratio for a failure mode is the one in your own records, and the worked example above uses your inputs, not a borrowed figure. For the broader argument, see reactive vs preventive maintenance cost. This article stays on the narrower question of selection by failure mode.

What planned run-to-failure still requires from you

A planned run to failure maintenance strategy is not zero work. It shifts effort from calendar tasks to preparation, and the preparation is what separates it from neglect.

  • Stock the spare. The cost advantage assumes quick recovery. If the replacement takes weeks to arrive, the failure is expensive after all.
  • Define the response. Say who is notified, who fixes it, and what the interim workaround is.
  • Keep failure records. Log each failure with its date and repair cost, so the decision can be revisited with evidence.
  • Set a review trigger. Revisit the choice if failures per year rise, if the process changes so that a failure becomes more consequential, or after a fixed period such as annually.
  • Watch for hidden consequences. A component that "just fails" may take a neighboring part with it. If the log shows collateral damage, the cost per failure has changed.

A note on scope: MaintenanceCalculator is a planning tool, not a CMMS. It does not issue work orders, dispatch technicians, or track spare-parts inventory, and that is deliberate. The stocking and response steps above belong in whatever system you already use for those jobs. The planning question, which regime each failure mode sits in and what it costs, is what the calculators address.

Documenting the choice so it survives staff turnover

Any run to failure maintenance strategy that lives only in a technician's head will be reversed the first time the component fails and someone asks why it was not on the PM list. For each failure mode kept off the PM schedule, write down:

  1. The asset, the component, and the failure mode.
  2. The consequence of failure and why it is acceptable.
  3. The cost comparison, with the inputs and the date they were taken.
  4. The spare and response arrangements.
  5. The review date, and the condition that triggers an earlier review.
  6. The name of the person who approved it.

Keep this beside the PM schedule so a reader sees both lists together. Anyone auditing your program can then see that every asset was considered and every exclusion has a reason. If an insurer, auditor, or regulator sets requirements for specific equipment, check those directly; a cost comparison does not override them.

The PM vs Run-to-Failure Cost Comparison Calculator takes your PM cost and your own failure history and returns a per-failure-mode regime verdict, so the comparison above does not have to be rebuilt by hand for every row. It is an Excel workbook. To get future articles on interval economics, subscribe to the newsletter on this site.

Get maintenance guides in your inbox

Related guides