How to build a Preventive Maintenance schedule that works

Start with failure modes

The first question for every preventive maintenance task should be simple: What failure are we trying to prevent or detect? This is the foundation of Reliability-Centered Maintenance (RCM), the approach popularized by John Moubray. Rather than starting with a calendar interval and assigning a task to an asset, RCM starts with the relationship between the asset, its failure modes, the consequences of those failures, and the maintenance task that can address them.

A single asset can have multiple failure modes and various criticality rankings, and those failure modes may require completely different maintenance strategies.

For example, a pump might experience bearing degradation, seal failure, or electrical problems. A vibration inspection may be appropriate for detecting bearing degradation, while lubrication may be better managed according to operating hours or another appropriate interval. Some failures may call for scheduled restoration or scheduled replacement. Others may require failure-finding tasks, particularly where a hidden protective function is involved. And for some low-consequence failures, run-to-failure may be the most rational strategy.

The goal isn’t to prevent every failure at any cost. It’s to choose a maintenance strategy that makes sense for the specific failure and its consequences.

“The best PM task is one that has a clear reason for existing. If we can’t explain which failure mode it addresses and how it changes the risk, we should question whether it belongs on the schedule,” says Jason Afara, a former maintenance manager, now a Director at Fiix by Rockwell Automation.

That failure-mode-first mindset also makes PM reviews more productive. Instead of asking, “Are we doing enough maintenance?” the team can ask, “Are we managing the important failure modes effectively?”

Set the right preventive maintenance frequency

“Determine your interval” is often where generic maintenance guidance stops. In practice, deciding how frequently a task should occur is one of the most important and most difficult parts of PM optimization.

Start with three inputs:

1. Understand the failure behavior. Start from a complete asset inventory and ask: is the failure age-related, condition-related, or essentially unpredictable? Age-related failures may support scheduled replacement or restoration. Condition-related failures may be better suited to inspections or condition monitoring. If a failure is genuinely random and no effective proactive task exists, adding more calendar-based maintenance may not reduce risk.
2. Understand the P-F interval. For condition-based maintenance, the P-F interval is the time between the point when a potential failure can first be detected and the point when functional failure occurs. A useful inspection interval needs to provide enough opportunity to detect the developing problem and act before the equipment fails.
3. Balance cost and consequences. Frequency should reflect the cost of performing the PM, the probability and consequence of failure, the cost of downtime and repair, and the risk of over-maintenance.

Moubray’s RCM framework reinforces the principle that proactive maintenance should be justified by the consequences and economics of failure, not simply because a task can be performed. Your computerized maintenance management system (CMMS) history can add evidence to that decision. Failure records can reveal recurring patterns, while Weibull analysis can help teams understand failure behavior and support interval decisions where the data is sufficient.

Another useful safeguard is to avoid treating manufacturer recommendations as automatic PM intervals. Manufacturer guidance can be an important input, but the operating environment, duty cycle, criticality, and failure history of the specific asset may justify a different approach. The schedule should reflect the asset as it actually operates, not simply the assumptions behind a generic interval. “We don’t want the shortest possible interval. We want the interval that gives us enough protection without spending maintenance resources where they don’t reduce risk,” says Jason.

Use Weibull analysis to set preventive maintenance intervals

Weibull analysis is a statistical method that fits your failure history (time between failures) to a curve. It tells you whether a failure mode gets more likely as a part ages. Many don’t. The key output of the analysis is the shape value, beta (β).

In short: only a β value of above 1 supports age-based replacement. Below 1 or around 1, replacing parts on a calendar won’t reduce failures. The table below helps illustrate this in a clear way:

Beta (β) answers the question: should we replace this part on a schedule? The answer to how often? comes from beta (β), the part’s usual lifespan (η), and the cost of breakdown versus a planned replacement.

An example from Jardine and Tsang shows how this works. In a Foundry, a critical bearing had failed 5 times. Each failure costs about twice as much as a planned replacement. The analysis showed clear wear-and-tear, with the lowest-cost of replacement age being 14 weeks.

The more useful finding was that anywhere from 10-14 weeks cost about the same. Earlier than 10 weeks was over-maintenance, and later than 14 pushed the costs up quickly. The flat range matters to planners. A PM can move a week or two to fit crew capacity without adding cost.

With only a handful of failures, treat (β) and the interval as estimates. Plan around a range, not a single date.

Its important to keep three things in mind when performing Weibull Analysis:

• Analyze one failure mode at a time. Lumping failures together makes them look random, and PM looks useless.
• Count the assets that haven’t failed yet. Leaving them out makes part look shorter lived and pushes the interval too short.
• Check that repairs restore the part to as good as new. Weibull assumes they do. If failures keep coming closer together, the asset as a whole is wearing out, and Weibull will give a misleading interval. Plot the times between failures in order and look for a trend before you run the analysis.

Make the schedule fit your maintenance capacity

A PM schedule can be technically correct and still fail because more work is scheduled than the maintenance team can realistically complete. This is one of the most practical problems in maintenance planning and one that is easy to overlook when schedules are built without considering execution.

Before adding another recurring task, consider the capacity required to perform it:

• Technician availability
• Estimated job duration
• Required skills and resources
• Parts and materials
• Production access
• Planned downtime
• Emergency and break-in work

If 200 PMs become due during the same week, the issue isn’t necessarily that technicians aren’t working hard enough. The schedule itself may be poorly distributed.

Consider a five-person crew working 40 hours each. That gives you 200 total labor hours on paper. If 200 PMs average an hour each, you’ve already consumed the crew’s entire theoretical capacity before accounting for reactive work, planning, delays, or other demands. Actual wrench time further reduces the hours available to physically execute maintenance.

Instead of scheduling against 100% of theoretical capacity, use historical wrench time and reactive maintenance hours to establish a realistic weekly capacity. If PM demand exceeds it, stagger the work across maintenance windows before it turns into backlog.

Instead of allowing large numbers of tasks to become due simultaneously, distribute recurring work across the available maintenance windows. Accurate job estimates, thoughtful planning, and staging of parts and materials can make that workload more predictable.

“A PM schedule has to fit the plant, not just the equipment. If production access, parts, or technician availability aren’t considered, a technically perfect schedule can still become a backlog,” says Mike Cooper, Solutions Consultant at Fiix by Rockwell Automation.

Capacity planning also makes it easier to distinguish planned maintenance from reactive work. Emergency jobs will happen. The goal is not to create a schedule that assumes zero interruptions; it’s to build a schedule that remains achievable when real-world interruptions occur.

Capacity should also be viewed over the planning horizon, not just week by week. A task that is easy to complete individually can create a substantial workload when multiplied across hundreds of assets. Grouping compatible work, planning around production windows, and reviewing upcoming PM demand can help prevent predictable capacity spikes before they become overdue work.

Measure what actually gets done

A PM that exists in the CMMS but is repeatedly deferred isn’t functioning as part of a reliable maintenance strategy. That’s why schedule compliance, or schedule attainment, matters.

At a minimum, teams should monitor:

• Completed versus scheduled work
• PM backlog
• Deferred PMs
• Reasons for missed work
• Rules for rescheduling

The reasons behind missed work are especially important. A PM may be deferred because production access wasn’t available, a critical part was missing, an emergency breakdown took priority, or the schedule was simply overloaded.

“A missed PM is data. We need to know why it was missed before we decide what to do about it. Sometimes the task needs to move; sometimes the schedule needs to change; sometimes the PM itself needs to be challenged,” says Jason.

Not every deferral is a problem. Some are legitimate decisions based on changing operational priorities or risk. The problem is when deferral becomes the default mechanism for dealing with an unrealistic schedule. Establish clear rules for rescheduling, so deferred work doesn’t disappear into an ever-growing backlog.

How to calculate compliance and PM yield

A PM that exists in your CMMS but is repeatedly deferred isn’t doing much to prevent failure. That’s why it’s important to measure not only whether PMs are completed, but whether they’re completed on time.

One useful measure is PM compliance:

PM compliance (%) = (PMs completed on time ÷ PMs scheduled) × 100

Using the 10% rule, “on time” means completing the PM within 10% of its scheduled interval. For example, a PM with a 30-day interval has a window of roughly three days on either side of its due date. A weekly PM has a much tighter window of around half a day.

As a general benchmark, teams can aim for 90% PM compliance or higher, with 95% often used as a world-class target. The important thing is consistency. If PMs regularly fall outside their expected window, look at why they’re being deferred and whether the schedule reflects the maintenance team’s actual capacity.

There’s also schedule compliance, which looks at execution from a different angle:

Schedule compliance (%) = (Scheduled work hours completed ÷ Total work hours on the weekly schedule) × 100

This formula originates from Palmer, and instead of counting individual PMs, schedule compliance measures the work hours completed from the schedule established at the beginning of the week. For example, if a crew starts the week with 400 scheduled work hours but emergencies and other interruptions mean they complete only 320 of those hours:

320 ÷ 400 × 100 = 80% schedule compliance

Looking at both metrics can tell you more than either one alone. PM compliance shows whether preventive maintenance is happening when it should, while schedule compliance shows whether the weekly workload is realistic and executable. If either consistently falls short, the answer may not be to push technicians harder. It may be time to adjust PM frequencies, stagger recurring work, or leave more capacity for the reactive work that inevitably breaks into the schedule.

Audit and optimize your PM schedule

Optimization doesn’t mean simply doing fewer PMs. It means doing the right PMs. A useful PM audit can examine every recurring task and ask:

• What failure mode does it address?
• How often does it find a problem? (i.e., PM yield)
• Does it actually reduce risk?
• Is the frequency still appropriate?
• Is there evidence of over-maintenance?
• Could the task be replaced by condition monitoring?
• Should the task be eliminated?

Download your free PM schedule audit checklist (599KB PDF opens in new tab)

Note: Any legally required safety critical or warranty linked PMs are off limits without a formal review.

The last question can be uncomfortable, particularly when a PM has existed for years. But historical existence is not evidence of effectiveness. Your CMMS can provide useful evidence for the review. Look for PMs with findings, PMs repeatedly completed with no findings, failures occurring between PMs, PMs frequently deferred, and corrective work generated by inspections.

None of these metrics should be interpreted in isolation. A PM with no findings might be highly effective at preventing a failure, or it might be unnecessary. A PM that generates many corrective work orders may be valuable because it catches problems early, or it may indicate that the inspection is being performed too late.

The goal is to determine PM effectiveness: whether the task is actually managing the failure mode it was designed to address.

PM optimization is not a race to get the number of PMs down. It’s about understanding which tasks protect the business, which don’t, and where the data tells us to change course.

A PM audit can also reveal where maintenance effort is being spent without generating useful information. If a recurring inspection almost never identifies degradation, while failures repeatedly occur through a different pathway, the team has an opportunity to redirect effort. That might mean changing the task, changing the interval, adding a better measurement, or accepting run-to-failure when the risk is low.

Move from calendar-based to condition-based scheduling

Calendar-based maintenance makes sense when time or operating hours are a useful predictor of failure or when an effective scheduled task can be justified. But when equipment condition can be measured reliably, condition-based maintenance can provide a more targeted approach.

Common condition indicators include:

• Vibration
• Temperature
• Pressure
• Oil condition
• Sensor thresholds

Instead of automatically generating a work order every three months, for example, a condition-based strategy can trigger maintenance when a monitored parameter indicates that equipment health is deteriorating. The key is reliable measurement and a clear response strategy. A sensor value is only useful when the organization knows what threshold matters, what failure mode it indicates, and what action should follow.

What a good PM schedule looks like

A strong PM schedule connects the asset and failure mode to the appropriate task, trigger, and expected effort.

This table is only the starting point. A real schedule also needs to account for technician capacity, required materials and skills, production constraints, planned downtime, and the consequences of delaying the task.

Build a preventive maintenance schedule that evolves with your operation

A preventive maintenance schedule shouldn’t be treated as a static list of recurring work. It should be a living maintenance strategy that connects failure modes to the right tasks, intervals, resources, and evidence.

The goal isn’t to schedule more maintenance. It’s to schedule the maintenance that matters, at the frequency that makes sense, in a way the team can actually execute and then use the results to keep improving the schedule.

Sources and influences

Moubray, J. (1997). Reliability-centered maintenance (RCM II). Industrial Press.

Jardine, A.K.S., & Tsang, A.H.C. (2013). Maintenance, replacement, and reliability: Theory and applications (2nd ed.). CRC Press.

Palmer, D. (2019). Maintenance planning and scheduling handbook (4th ed.). McGraw-Hill Education.

Frequently asked questions