Why Most PM Programs Underdeliver
A manufacturing operation launches a PM program.
The maintenance team executes it diligently.
PM compliance reaches 85%.
Unplanned downtime does not decline meaningfully.
The operations manager concludes that PM does not work.
The actual conclusion should be different.
The PM program was not calibrated to the actual failure behavior of the assets it was designed to protect.
The intervals were set from manufacturer recommendations — written for average utilization conditions that do not match this facility.
The tasks were copied from a generic maintenance template that was never validated against this facility's actual failure history.
The trigger types — calendar intervals for every asset regardless of utilization variability — systematically over-maintain low-utilization assets and under-maintain high-utilization ones simultaneously.
A PM program that is structurally miscalibrated produces high compliance and poor outcomes.
This guide builds a PM program from the evidence rather than from assumptions — producing a program that is calibrated to actual failure behavior from the start and improves as the maintenance history builds.
Step 1: Build the Asset Register
A PM program cannot exist without an accurate, complete list of the assets it is designed to protect.
This step is straightforward in principle and frequently underestimated in practice.
What the asset register must contain:
Every production asset that warrants scheduled maintenance — identified by location, asset type, and manufacturer and model information.
A parent-child hierarchy that distinguishes major assemblies from their sub-components — so that PM tasks can be assigned at the appropriate level rather than at a generic machine level that obscures which specific component is being maintained.
The production criticality classification for each asset — Tier 1 mission-critical, Tier 2 significant, Tier 3 supportive — which determines the maintenance investment level each asset warrants.
What the asset register does not need at this stage:
Perfect data on every component of every machine.
Complete maintenance history from previous systems.
Photographs of every asset.
Starting with a functional asset register that captures the information needed to build the PM schedule is more valuable than delaying the program start while pursuing perfect data completeness.
The register improves as the program runs.
Step 2: Identify Failure Modes for Tier 1 and Tier 2 Assets
For Tier 3 assets — those whose failure does not significantly affect production, quality, or safety — a PM program may not be warranted at all.
Run-to-failure with a documented emergency response plan is often the most cost-effective strategy for low-criticality assets.
For Tier 1 and Tier 2 assets, the next step is failure mode identification — an honest assessment of how each asset fails and what warning signs precede the failure.
This does not require a formal Failure Mode and Effects Analysis for every asset.
It requires answering three practical questions for each significant asset class.
Question 1: What are the most common failure modes for this asset type in our operation?
Pull the last 24 months of maintenance records for each asset class.
What fault codes appear most frequently?
What components are replaced most often in corrective repairs?
What is the average time between significant failures on these assets?
Question 2: Does each failure mode produce detectable precursor signals before functional failure occurs?
Bearing wear produces increasing vibration and temperature before functional failure.
Hydraulic seal degradation produces gradually increasing oil consumption before complete seal failure.
Cutting tool wear produces increasing cycle time deviation before dimensional failure.
Some failure modes — electrical relay failures, sudden mechanical fractures — produce no reliable precursor signal.
Question 3: How much time typically elapses between the appearance of precursor signals and functional failure?
This is the P-F interval — the window available for condition-based intervention.
A long P-F interval means there is time to plan and execute a scheduled repair after detecting the precursor signal.
A short P-F interval means rapid response is required when the signal appears — or the window closes before intervention is possible.
The answers to these three questions determine which trigger type is appropriate for each asset's PM schedule.
Step 3: Select the Trigger Type for Each Asset Class
Three trigger types exist for PM scheduling.
The correct trigger type for each asset depends on the failure mode analysis from Step 2 — not on scheduling convenience.
Calendar-based triggers
Applied to: Assets with relatively uniform utilization where the failure mode does not produce reliable precursor signals — or where condition monitoring data is not yet available.
How it works: PM tasks are scheduled at fixed time intervals — weekly, monthly, quarterly — regardless of how much the asset has actually run since the last PM.
Strength: Simple to schedule and easy to communicate to the maintenance team.
Limitation: Systematically wrong for assets with variable utilization. A calendar interval set for average utilization over-maintains low-utilization periods and under-maintains high-utilization periods simultaneously.
Usage-based triggers
Applied to: Assets where failure correlation is stronger with usage than with time — cycle counts, run hours, units produced, meters processed.
How it works: PM tasks are triggered when the asset reaches a defined usage threshold since the last maintenance event — regardless of how many calendar days have elapsed.
Strength: Accurately reflects actual wear accumulation rather than time elapsed. A press that ran 200,000 strokes receives its maintenance at 200,000 strokes whether that took 30 days or 90 days.
Limitation: Requires accurate usage data — either from automatic machine signal capture or from disciplined manual meter entry.
Condition-based triggers
Applied to: Tier 1 assets with detectable failure mode precursors and data connectivity to capture those precursors automatically.
How it works: PM tasks are triggered when a monitored condition parameter — OEE performance trend, vibration amplitude, temperature deviation, current draw — crosses a configured threshold that indicates the P point of the failure mode.
Strength: The most accurate trigger type available — maintenance happens when the asset actually needs it, not when the calendar or usage counter assumes it does. Eliminates both over-maintenance and under-maintenance simultaneously.
Limitation: Requires machine connectivity and a monitoring system capable of detecting the relevant condition parameters. Requires sufficient P-F interval to allow scheduled response after threshold crossing.
Most manufacturing PM programs use all three trigger types simultaneously — calendar triggers for general facility and support equipment, usage triggers for production-critical assets with measurable utilization, and condition triggers for the highest-criticality assets with accessible machine connectivity.
Step 4: Set the Initial PM Intervals
With trigger types selected, the initial intervals for each PM task must be set.
The critical principle here is that the initial interval should be set conservatively — shorter than the expected failure interval — and then extended as maintenance history accumulates evidence that the interval is appropriate.
Starting with an interval that is too long and experiencing failures is more expensive and more disruptive than starting with an interval that is slightly too short and extending it when findings consistently show no wear approaching the intervention threshold.
For calendar-based PMs:
Start with an interval that is 70 to 80% of the shortest failure interval observed in the last 24 months of maintenance history for that asset class.
If the asset class has no failure history — because it has never been maintained systematically — use the manufacturer's recommendation as the starting point and adjust based on the first 6 months of PM findings.
For usage-based PMs:
Start with a usage threshold that is 70 to 80% of the average usage accumulated between significant corrective events in the last 24 months.
If usage data is not available, use the manufacturer's recommendation in terms of operating hours and convert to the relevant usage metric for the asset.
For condition-based PMs:
The trigger threshold is set at the P point of the failure mode — the condition parameter value that reliably indicates the onset of detectable degradation.
Establishing this threshold accurately requires either historical failure data showing the condition parameter values preceding each failure event, or a conservative initial threshold that is tightened as the monitoring dataset builds.
Step 5: Define the PM Task Content
The PM interval determines when the maintenance team arrives at the asset.
The PM task content determines what they do when they get there.
Both must be right for the PM to be effective.
Minimum required content for every PM task:
The specific inspection or intervention steps — in enough detail that a technician who has not previously performed this PM can complete it correctly.
The pass and fail criteria for each inspection point — what is acceptable condition and what triggers a corrective action.
The consumables and spare parts required — with specific part numbers to eliminate storeroom confusion.
The estimated time to complete — so that PM windows can be scheduled realistically against production availability.
The safety requirements specific to this asset — isolation procedures, PPE requirements, and any operational constraints.
What poor PM task content looks like:
"Inspect and lubricate machine" — no specificity on which components, what lubricant, how much, or what condition warrants replacement.
"Check all wear parts" — no definition of which parts, what wear limit triggers replacement, or what constitutes acceptable wear.
PM tasks written at this level of generality produce high compliance rates and low effectiveness — the technician completes the task in the time allocated and signs off, but the specific inspection points that would detect the developing failure are not explicitly checked.
Step 6: Build the PM Schedule Into the Maintenance System
With asset register, failure mode analysis, trigger types, intervals, and task content defined, the PM schedule is ready to be built into the maintenance management system.
The maintenance system — whether a CMMS or a simpler scheduling tool — must support three capabilities to execute the PM program reliably.
Automatic work order generation — PM work orders should be generated automatically when the trigger condition is met, rather than requiring a planner to manually create each work order from a schedule spreadsheet. Manual PM scheduling systems produce PMs that are missed during busy periods and duplicated during slow periods.
Mobile execution — PM work orders should be accessible to technicians on mobile devices at the machine, with the task checklist, pass-fail criteria, and parts list available without returning to a desktop system. PM data captured at the machine during the inspection is more accurate and more complete than data entered retrospectively.
PM finding capture — the work order should capture what was found during the PM — not just that the PM was completed. Whether the inspection found acceptable condition, wear approaching the intervention threshold, or a defect requiring corrective action is the data that calibrates the PM interval over time.
Without PM finding capture, the program cannot improve.
With it, every PM cycle generates the evidence that confirms whether the interval is appropriate, needs shortening, or can be extended.
Step 7: Review and Calibrate at 90 Days
A PM program launched without a 90-day review is a program that has been set and forgotten rather than managed.
The 90-day review asks five specific questions.
Are PMs being completed within their scheduled windows?
Completion rates below 85% indicate either that the PM schedule is conflicting with production requirements — in which case the scheduling approach needs adjustment — or that the maintenance team is capacity-constrained.
What are PM findings showing?
PMs that consistently find no wear approaching the intervention threshold are candidates for interval extension.
PMs that consistently find wear at or beyond the intervention threshold are candidates for interval shortening.
Have any unplanned failures occurred on assets with active PM programs?
A failure on an asset with an active PM indicates either that the PM interval is too long, the PM task content is not addressing the failure mode that occurred, or the trigger type is not appropriate for the failure mode.
Each failure on a PM-covered asset is an opportunity to improve the program rather than a sign that PM does not work.
Is the maintenance team capturing PM findings at the level of detail needed for interval calibration?
Generic completion entries — "PM complete, no issues" — do not support interval calibration.
Specific finding entries — "bearing shows early wear, estimated 3 to 4 months to intervention threshold at current utilization" — build the evidence base that makes the program more accurate over time.
Are PM windows aligned with production scheduling?
PMs that are consistently deferred because production cannot release the asset indicate a scheduling alignment problem that needs to be resolved at the planning level rather than on the production floor.
The PM Program Maturity Path
A PM program is not a project with a completion date.
It is a continuous improvement program that becomes more accurate and more effective as the maintenance history it generates builds the evidence base for interval calibration.
Month 1 to 3: Asset register built. Initial PM schedule launched. Calendar triggers for most assets. Conservative intervals based on maintenance history and manufacturer recommendations.
Month 3 to 6: First 90-day review. Interval adjustments based on PM findings. Usage-based triggers implemented for high-utilization Tier 1 assets where meter data is available. First-pass Bad Actor identification from corrective work order data.
Month 6 to 12: Condition-based triggers implemented for the highest-criticality assets with machine connectivity. PM task content refined based on technician feedback and finding patterns. Emergency PM additions for failure modes identified in the first 6 months that were not in the initial program.
Month 12 onwards: Annual program review. Interval optimization based on 12 months of accumulated PM finding data. Transition from calendar to usage-based triggers for additional asset classes as utilization data becomes available. Continuous calibration cycle.
Frequently Asked Questions
How long does it take to build and launch a PM program for a mid-sized manufacturing facility?
The asset register, failure mode analysis, and initial PM schedule can be built in four to six weeks with appropriate stakeholder involvement.
Loading the schedule into a CMMS and completing technician training takes an additional two to three weeks.
A functional PM program — generating and completing work orders — is achievable within eight weeks of project start for most mid-sized facilities.
What PM compliance rate should we target?
85% is the minimum acceptable PM compliance rate for a well-designed program.
Below 85% indicates either scheduling conflicts with production, maintenance team capacity constraints, or PM task content that is taking significantly longer to complete than estimated.
Above 95% compliance combined with flat or rising unplanned downtime indicates a different problem — the PM program is being completed but is not calibrated to prevent the failures that are actually occurring.
How do we handle PM compliance for assets that production cannot release for maintenance?
Build PM windows into the production schedule rather than negotiating them reactively.
PM requirements should be visible to production planners when they are scheduling production orders — not discovered when the maintenance team arrives to perform the PM and finds the asset in production.
This requires either a shared production-maintenance planning process or a planning tool that shows both production orders and maintenance requirements in the same scheduling view.
A PM program built from evidence — actual failure history, usage data, and condition monitoring signals — is not just a compliance exercise. It is the architecture that converts maintenance effort into failure prevention. The difference between a program that produces compliance and one that produces reliability is the quality of the calibration beneath the schedule.
Login
Blog
