PM Plan Optimisation Techniques: How to "Trim the Fat" and Stop Drowning in Work Orders

It is the year 2026. We have AI-driven analytics, IoT sensors on almost every asset, and mobile CMMS capabilities that fit in our pockets. Yet, when I walk into manufacturing plants or processing facilities, I still hear the same exhausted complaint from Maintenance Managers: "We are drowning in PMs, but our equipment keeps failing."

If you are reading this, you are likely facing the "PM Paradox": You have a schedule packed with Preventive Maintenance (PM) tasks, your team’s wrench time is maxed out, and yet your unscheduled downtime remains stubbornly high. You don't need more maintenance. You need the right maintenance.

This guide is not about adding to your workload. It is about the "PM Diet"—a rigorous approach to trimming the fat from your maintenance strategy. We will explore PM plan optimisation techniques that focus on quality over quantity, leveraging data to do less work that yields higher reliability.

The Core Question: Why Isn't My Current PM Schedule Preventing Failures?

Before we fix the plan, we must understand why the traditional approach fails. The core question most Reliability Engineers ask is: "I am following the manufacturer's recommendations; why is the machine still breaking?"

The Answer: Because time-based maintenance often ignores the reality of failure patterns.

For decades, we operated under the assumption that assets degrade linearly with time—that the older a part gets, the more likely it is to fail. However, reliability studies (dating back to the airline industry and confirmed repeatedly in manufacturing) show that only 11% to 18% of failures are age-related.

The vast majority (82% to 89%) of failure modes are random. They are caused by process variability, operator error, or—crucially—intrusive maintenance.

The Trap of Intrusive Maintenance

Every time a technician opens a gearbox to inspect it, introduces a new seal, or re-torques a bolt that was fine, they introduce a risk of human error. This is known as the "Worn-Out Infant Mortality" curve. By over-scheduling intrusive PMs, you aren't just wasting time; you are actively introducing defects into your system.

PM Plan Optimisation is not just about efficiency; it is about risk reduction. The goal is to transition from "Scheduled Restoration" to "Condition-Based Intervention."

Follow-Up: How Do I Identify Which Tasks to Cut? (The "Kill Your Darlings" Audit)

Once you accept that "more PMs" does not equal "more reliability," the natural next question is: "How do I know which tasks are useless and which are critical?"

You need to perform a PM Audit. This is the "PM Diet" in action. You are looking to cut the caloric intake of your maintenance department (labor hours) while increasing muscle mass (asset uptime).

The 6-Question Filter

To optimize your plan, take your top 10 worst-performing assets or your most labor-intensive PM routes. Review every single line item on the work order against this filter. If a task cannot survive these questions, kill it.

1. Does this task address a specific, likely failure mode?
   • Bad: "Inspect Conveyor." (Too vague).
   • Good: "Inspect belt tension and check for fraying edges to prevent tracking loss."
2. Is the task feasible?
   • Can the technician actually see/measure what is asked without a shutdown? If it requires a shutdown but is scheduled as a running inspection, it’s getting pencil-whipped.
3. Does the task duplicate another defense?
   • Are you manually checking pressure gauges on a system that has a SCADA alarm and auto-shutoff? If so, verify the sensor, don't read the gauge.
4. Is the frequency scientifically justified?
   • Why is this weekly? Is the P-F interval (the time between a potential failure being detectable and functional failure) actually less than a week? If the P-F interval is 3 months, a weekly inspection is waste.
5. Is it intrusive?
   • Does the task require disassembly? If yes, can it be replaced with a non-intrusive technology (like ultrasound or thermography)?
6. Does the cost of the PM exceed the cost of the failure?
   • For non-critical assets (Run-to-Failure candidates), spending $500 in labor annually to save a $50 part is poor economics.

Case Study: The Over-Maintained Conveyor

To illustrate this filter in action, consider a real-world scenario from a mid-sized bottling plant. The maintenance team was struggling with a palletizer conveyor that had a weekly PM checklist containing 18 separate tasks, taking two technicians two hours to complete (4 man-hours/week). Despite this, the conveyor suffered chain failures quarterly.

Upon applying the 6-Question Filter, the team discovered:

• Task: "Remove guard and inspect chain slack." Verdict: Intrusive. Removing the guard often led to misalignment during reassembly.
• Task: "Lubricate bearings." Verdict: No scientific frequency. They were over-greasing, blowing out seals, and attracting dust which ground down the chain.
• Task: "Check motor temperature." Verdict: Duplication. The motor had a thermal overload trip installed.

The Optimization Result: The team replaced the 18-step weekly PM with a monthly "dynamic inspection" (strobe light inspection of the chain while running) and installed an automatic oiler. The labor dropped from 208 hours/year to 24 hours/year, and chain failures dropped to zero because the intrusive "guard removal" was eliminated.

The "Pencil-Whip" Indicator

Look at your PM procedures data. If a technician consistently completes a 4-hour PM route in 45 minutes, your plan isn't optimized; it's being ignored. This is a massive red flag that the tasks are either impossible to perform or perceived as useless by the floor staff. Optimisation starts with honesty.

Follow-Up: I Have Thousands of Assets. Where Do I Start?

You cannot optimize everything at once. Attempting to overhaul the PM plan for an entire plant simultaneously is a recipe for burnout. The follow-up question is strictly about prioritization: "How do I rank my equipment to know where to apply these techniques first?"

The industry standard for this is Asset Criticality Analysis (ACA).

The Criticality Matrix

You must rank every asset on two axes: Severity of Failure and Probability of Failure.

1. Safety/Environmental Risk (The Trump Card): If a failure hurts someone or causes a spill, it is Criticality A.
2. Operational Impact: Does this machine stop the whole line (bottleneck), or is there a standby unit?
3. Maintenance Cost: Is it expensive to fix?

Calculating the Risk Priority Number (RPN)

To remove emotion from this ranking and make it objective, use a standard RPN formula: Severity (1-10) x Occurrence (1-10) x Detection (1-10).

• Severity: 10 is a safety hazard; 1 is a minor cosmetic issue.
• Occurrence: 10 is daily failure; 1 is once every 5 years.
• Detection: 10 means you won't know it failed until the line stops; 1 means you have instant sensor alerts.

Thresholds for Action:

• RPN > 100: Immediate PM Optimisation required. These are your "bad actors."
• RPN 40-100: Review annually.
• RPN < 40: Run-to-Failure candidate.

The 80/20 Rule of Optimisation:

• Criticality A (Top 20%): These assets get Reliability Centered Maintenance (RCM). You analyze every failure mode. You use predictive maintenance sensors. You spare no expense.
• Criticality B (Middle 50%): These get standard OEM recommendations, optimized by experience.
• Criticality C (Bottom 30%): These are Run-to-Failure. Delete the PMs. Yes, let the bathroom exhaust fan break. Let the lightbulb burn out. Redirect those labor hours to the Criticality A assets.

By formally declaring 30% of your assets as "Run-to-Failure," you immediately free up hundreds of technician hours. This is the "fat" you are trimming.

Follow-Up: How Do I Move from Time-Based to Condition-Based Maintenance (CBM)?

You’ve cut the waste. Now, how do you modernize the remaining tasks? The question here is: "If I stop checking it every Monday, how do I ensure it doesn't break on Tuesday?"

This requires shifting from a calendar mindset to a condition mindset, utilizing the P-F Curve.

Understanding the P-F Interval

The P-F Interval is the time between the point where a potential failure is detectable (P) and the point of functional failure (F).

• Time-Based PM: Guesses where you are on the curve.
• Condition-Based PM: Measures exactly where you are.

Practical Optimisation Techniques for CBM

Instead of "Replace Bearings every 10,000 hours" (which wastes good bearings), you implement:

1. Vibration Analysis: For rotating equipment like motors and pumps, vibration sensors detect pitting in bearings months before failure.

   • Optimisation: Replace monthly manual greasing with greasing only when decibel levels indicate friction.

2. Oil Analysis: Don't change hydraulic fluid based on the calendar. Change it when the particle count or viscosity demands it.

   • Optimisation: This saves thousands in consumables and waste disposal fees.

3. Ultrasonic Testing: Use ultrasound to detect air leaks or electrical arcing.

   • Optimisation: Instead of tightening every connection in a panel (dangerous and intrusive), scan the panel. If it's silent, it's tight. Move on.

By integrating these technologies into your equipment maintenance software, you trigger work orders based on data, not dates.

Follow-Up: What Are the Common Mistakes to Avoid?

Even with the best intentions, PM optimisation can backfire if executed poorly. The question here is: "What pitfalls should I watch out for?"

1. The "Copy-Paste" Catastrophe

Many organizations buy a new CMMS and simply copy-paste their old, bad data into the new system. Or, they copy PM plans from one site to another without accounting for operating context.

• Context Matters: A pump running clean water indoors needs a different plan than the exact same pump moving slurry outdoors in freezing temperatures.

2. Ignoring the Operators (Total Productive Maintenance - TPM)

Your operators know the machines better than anyone. A common mistake is optimizing PMs in a conference room without consulting the floor.

• The Fix: Move simple tasks (cleaning, visual checks, tightening) to operators (Autonomous Maintenance). This frees up your skilled technicians for complex prescriptive maintenance tasks.

3. Starving the Backlog

When you delete PMs, you might feel nervous that you aren't "doing enough." Do not fill the void with busy work. If you free up 20 hours a week, put that time into Root Cause Analysis (RCA) on the last major breakdown. Shift from reactive fixing to proactive solving.

4. The "Other" Trap (Data Hygiene)

Optimisation relies on historical data. A critical mistake is allowing technicians to close work orders with generic failure codes like "Broken," "Fixed," or "Other." If 40% of your failure history is labeled "Other," you cannot analyze which PMs are effective.

• The Fix: Enforce a standardized failure code hierarchy (e.g., Component > Sub-Component > Failure Mode) in your CMMS. You cannot optimize what you cannot define.

Follow-Up: How Do I Measure if the Optimisation is Working?

You’ve implemented the changes. Your boss asks, "Is it worth it?" You need metrics that prove value, not just activity.

Stop Measuring "PM Compliance"

High PM compliance (99%) often means your technicians are pencil-whipping tasks or the tasks are too easy. It is a vanity metric.

The Real KPIs of Optimisation

1. PM Effectiveness (Yield): Formula: (Number of Defects Found during PM / Total PMs Completed) * 100.

   • If you inspect a machine 52 times a year and find 0 problems, your inspection frequency is too high. You want a yield of roughly 10-15%. This means you are looking at the right time.

2. Corrective vs. Preventive Ratio: Target a 20:80 ratio. If 50% of your work is still emergency repairs, your PMs are not effective yet.

3. MTBF (Mean Time Between Failures): This is the ultimate lagging indicator. If your PM optimisation is working, MTBF should trend upward over 6-12 months.

4. Wrench Time: Optimized plans reduce travel time and part hunting. Use mobile CMMS features to track how much time is actually spent fixing vs. administrative tasks.

For a deeper dive into reliability metrics, organizations like the Society for Maintenance & Reliability Professionals (SMRP) offer excellent benchmarking guides.

Follow-Up: What About the Future? AI and the 2026 Landscape

The final question looks forward: "How does AI change PM optimisation?"

In 2026, we are moving beyond simple "Predictive" (telling you it will break) to "Prescriptive" (telling you how to fix it and when).

AI-Driven Optimisation

Modern platforms utilize AI predictive maintenance to dynamically adjust PM schedules.

• Scenario: Your static plan says "Change filter every 30 days."
• AI Reality: The AI analyzes pressure differential trends, production load, and intake air quality. It notices that during low-production months, the filter lasts 60 days. It automatically pushes the PM due date, saving labor and parts. Conversely, during high dust loads, it pulls the date forward to prevent a clog.

This is Dynamic PM Scheduling. It is the pinnacle of optimisation because it aligns maintenance strictly with asset needs in real-time.

The Role of Integration

Optimisation requires data silos to break down. Your CMMS must talk to your ERP and your SCADA system. Integrations allow your maintenance plan to see the bigger picture—inventory levels, production schedules, and financial constraints—to recommend the optimal time for downtime.

Conclusion: Start Small, but Start Today

PM plan optimisation is not a one-time project; it is a culture shift. It requires the courage to stop doing work that doesn't add value.

Your Action Plan:

1. Select 5 critical assets.
2. Audit their PMs using the 6-Question Filter.
3. Implement one condition-based trigger (e.g., vibration or amp draw).
4. Measure the results for 90 days.

By trimming the fat, you don't just get a leaner schedule; you get a more reliable, safer, and profitable facility. The goal of maintenance is not to be good at fixing things—it is to ensure they don't break in the first place.

For more on how to structure your preventive strategies, explore our guide on preventive maintenance workflows.