Schedule compliance is the share of your frozen weekly maintenance schedule that got completed inside that week. You can count it in work orders or in planned hours. On its own it tells you little, because 3 plants in very different shape can post the same score.
Loading your schedule to 100% and achieving 100% schedule compliance are 2 completely different things.
In my experience people confuse those 2 all the time.
Below are the ways the number gets calculated, why loading comes before targets, where the famous 90% comes from, the 2 metrics you need beside it, how it gets gamed, and how it differs from PM compliance.
How do you calculate schedule compliance?
There are 2 basic ways, and a third that many people assume is the same as the second.
You can count jobs: the jobs in the frozen weekly schedule completed within the week, divided by all the jobs in that schedule, times 100.
The downside is that a job is a job. A 20-hour job and a 3-hour job carry the same weight.
So I prefer hours: the planned hours of the scheduled jobs completed, divided by the planned hours of every job in the frozen schedule. Then a 20-hour job counts for more than a 3-hour job.
Doc Palmer prefers the work order count, because supervisors find it easier to read. That's a fair point.
| Variant | Numerator | Denominator | Who uses it |
|---|---|---|---|
| Work orders | Scheduled jobs completed within the week | All jobs in the frozen schedule | SMRP's work order metric, Palmer, most CMMS vendors |
| Hours vs scheduled | Planned hours of the scheduled jobs completed | Planned hours of all jobs in the frozen schedule | My preferred formula |
| Hours vs available | Scheduled work performed, in hours | Total hours available to schedule | SMRP's hours metric |
The last 2 rows only give the same answer when the week is loaded to 100% of capacity.
Every time you quote a compliance percentage, say which denominator produced it.
Use the same basis on both sides
In the hours version, use planned hours on top and planned hours on the bottom.
If you put actual hours over planned hours, your estimating accuracy ends up inside a scheduling metric. Track estimating accuracy as a separate number. Our planning accuracy article covers it.
The one exception is emergency work. It bypasses planning, so those jobs have no planned hours and you'll have to use actuals.
What the standards say
SMRP publishes 2 schedule compliance metrics in the 7th Edition of its Best Practices, and they give different answers. Its hours-based metric (5.4.3) divides by the total time available to schedule. Its work order metric (5.4.4) counts jobs, and it's weekly only.
Was the week actually full?
A compliance score is a fraction, and the denominator is whatever you chose to put into the week. Under-load the week and the score climbs while the plant does less work.
Load the schedule to 100% of your available hours. That's my third scheduling principle, and Palmer's fourth says the same.
Schedule 80%, and you've already told the crew that finishing 80% of the available work is good enough.
In a perfect week, when nothing goes wrong, Parkinson's Law takes over: the crew finishes what's in the schedule and stops.
Scheduling 80% is also a poor way to allow for emergencies, because it builds in the belief that break-ins are normal.
Overloading to 120% is worse. It builds in the acceptance that a good chunk of the schedule won't get done, and it wastes coordination effort on isolations, access and materials for jobs you already know aren't happening.
There's also a measurement argument. Load a crew with 400 available hours to 400 hours and every missing hour shows up. Give the same crew a 320-hour schedule and let them finish it, and you'll never work out where the other 80 hours went.
What "available" means
In my course, available hours are wrench time hours. I estimate task durations on wrench time too.
Start with a low wrench time assumption, say 30%. Once compliance holds for a few weeks, lift it to 40%, then 45%. Smaller steps after that.
Palmer works on gross clock hours. Whichever you use, put your capacity and your estimates on the same basis. Mix them and you'll be out by a factor of 2 to 3.
Suspect you're under-loading? Bump your wrench time assumption so you put 5% or 10% more hours into the schedule for a few weeks. Then see what happens.
Still drowning in break-ins after you've fixed prioritisation? Lower your wrench time assumption. You still load to 100% of that smaller number.
Is 90% schedule compliance world class?
You'll see 90% quoted almost everywhere, usually with no source. Whether it means anything depends on how full the week was.
SMRP does print it. Both its schedule compliance metrics give best in class as greater than 90%.
Read the cautions that come with it, because they usually get left off. The hours metric is only accurate if 100% of available maintenance hours are scheduled. The work order metric says the 90% target should be balanced against percent planned time.
Palmer treats the same number as a warning light. Below 40%, he says, the supervisor ignored the schedule. Above 90% usually means management didn't give the supervisor enough work to beat Parkinson's Law.
He also puts a ceiling on it. A plant where half the work can't wait until next week has a best possible compliance of 50% on a fully loaded schedule. And even a pretty good plant carries around 20% reactive work.
Your realistic ceiling is roughly 1 minus your reactive fraction.
When you're starting out, don't be shocked by 40% or 50%. Aim to get above 70% and hold it there for a run of weeks. Then push for 80% and 85%.
Above 90% is achievable on a fully loaded week, with emergency and fill-in work already low.
In my course, 90% to 92% compliance with 2% emergency work and 6% fill-in work, held week after week, is top performance.
What counts as completed "as scheduled"?
Write these rules down before you calculate anything, because your CMMS won't decide them for you.
Fix the week. Monday to Sunday for most plants, or aligned to your roster. The schedule locks at Thursday's review meeting and goes out on Friday morning, which leaves the weekend for staging and permits.
The day of the week doesn't matter. A job scheduled for Tuesday and finished on Thursday still counts. I don't look at which day, or at how long it took, only at whether the job got done inside the week.
Break-in work stays out of both the numerator and the denominator. Palmer gives an example. A crew is scheduled 100 work orders and completes 200, of which 50 were on the original schedule.
Compliance is 50%. Scoring it at 200% would reward the crew for abandoning the plan. Scoring it at 25% would punish them for doing extra work.
Take your snapshots on the right evenings
Your CMMS almost certainly won't produce these metrics out of the box. You'll either customise the system or calculate outside it, in a spreadsheet or something like Power BI.
- Sunday evening, before the week starts: snapshot every job in next week's schedule. That's your baseline and your denominator. Take it too early and you'll count fill-in work as emergency work.
- Sunday evening, at the same time: list every other open job in the CMMS. If your CMMS records the date a job was created, you can skip this one.
- Monday evening, after the week closes: snapshot the status of every job in the frozen schedule. Monday gives you a day to fix time writing first.
- Tuesday morning: calculate.
Without that baseline, anyone who can change the date on a late work order can change your compliance score.
Because each site builds its own calculation, comparing schedule compliance between 2 plants usually means comparing 2 home-made formulas.
The 3 numbers you need
Schedule compliance is probably the single most important measure of your planning and scheduling process. You still can't read it on its own.
I teach it as 1 of 3 metrics: schedule compliance, emergency work and fill-in work.
Emergency maintenance is work identified during the frozen week that was so urgent it had to be done right then and there. It didn't exist on Sunday evening.
Fill-in work already existed in your CMMS before the week started. It wasn't in the frozen schedule, and it got done anyway. It isn't urgent.
Fill-in work is the metric most plants don't track, and it's the one that makes the other 2 readable.
| Metric | Numerator | Denominator | My target |
|---|---|---|---|
| Schedule compliance | Scheduled work completed within the frozen week | The frozen weekly schedule, in jobs or planned hours | Above 70% to start, then 80% to 85%, then 90%+ on a fully loaded week |
| Emergency maintenance | Work completed in the week that didn't exist before it | The frozen weekly schedule (actual hours for the emergency jobs) | Under 15% to start, then under 10%. The best plants run under 2%, measured in hours |
| Fill-in work | Work completed in the week that existed before it but wasn't scheduled | The frozen weekly schedule | Under 15%, and 5% or less at the top end |
SMRP calls under 10% emergency work best in class. I think that's too soft.
The fill-in targets come from plants I've run and comparable plants in the same company around the world. Fill-in work isn't a commonly used metric, so there's no published benchmark for it.
Don't chase fill-in work to 0. If a scheduled job can't go ahead and the crew picks up a ready job instead, that's fine.
If 20% emergency work sounds high to you, it should. That's a very reactive maintenance environment.
Reading fill-in work
High fill-in work with low schedule compliance tells you people aren't sticking to the schedule. They're working to their own priorities.
High fill-in work with high schedule compliance usually means you aren't loading to your true capacity, or your estimates are well off. It can also mean a tick and flick culture, where the boxes get ticked and the work doesn't get done.
A typical improvement program looks like this. These are illustrative teaching figures from my course.
| Scenario | Schedule compliance | Emergency work | Fill-in work | What it tells you |
|---|---|---|---|---|
| 1. Reactive plant | 60% | 25% | 30% | Typical of a reactive organisation, or one that has only just started planning and scheduling. Fix prioritisation and work request quality first to cut the break-ins |
| 2. After about 3 months | 75% | 10% | 12% | More productive, and a good place to be. Not a good place to stay |
| 3. After another 3 to 6 months | 85% | 3% | 22%, up | See below |
| 4. Top performance | 90% to 92% | 2% | 6% | Sustain that week after week and you've earned a pat on the back. Then check that it's real |
In scenario 3, compliance is climbing, emergency work has collapsed, and fill-in work has gone up.
Usually it means you've built an efficient crew and they're getting through more work than the schedule holds. In an improving plant, rising fill-in work tells you to load more work into the schedule.
Check your estimating accuracy first, satisfy yourself there's no tick and flick, then turn the loading up.
How schedule compliance gets gamed
Most of these show up in one of the 3 metrics or in the loading figure. One doesn't show up anywhere.
| The move | The dashboard | What's happening | What exposes it |
|---|---|---|---|
| Schedule only what will definitely finish | High | The reactive work still happens, off-schedule | Fill-in work rises |
| Pad the estimates so fewer jobs fit | High | Fewer hours assigned, wrench time falls | Fill-in work rises, hours vs available falls |
| Add break-in work after the lock | High | The lock means nothing | The Sunday evening baseline catches it |
| Count cancelled work as complete | High | The work was never done | Nothing, which is why you audit completions |
| Tick and flick | High | The work was never done | Fill-in work high and compliance high |
| Load the week to half the available hours | High | Most of the week was never scheduled | Hours vs available collapses |
| Pay the supervisor on the number | High, then a fire | The supervisor protects their bonus | Emergency work stays flat while compliance climbs |
Who should own schedule compliance?
Palmer took 25 typical reasons a weekly schedule falls apart and assigned each one to either management or the supervisor. The tally came out at 22 management causes and 3 supervisor causes.
Tie supervisor pay to schedule compliance, he says, and you force supervisors to choose between helping the plant and helping their family.
The person who owns the weekly schedule must also own schedule compliance, and must also approve break-ins. In most plants that's the plant manager or the production manager. Every emergency break-in gets at least a verbal OK from that person.
That person has to sit across both maintenance and operations, or at least be able to speak for both.
Please don't make one person accountable for schedule compliance and give someone else the authority to approve break-ins. You'll end up with a holy mess. I've seen it many times.
There's a related trap. Your crew opens a pump for a scheduled service and finds far more work than anyone expected. That's emergent work, and your supervisor has to make a judgement call.
If the extra scope is best done now, do it now, even if it costs you the metric this week. If it's big enough to bust the week, your supervisor takes it to whoever approves the schedule.
Schedule compliance vs PM compliance
These 2 get mixed up all the time.
Schedule compliance tells you how much of the frozen week got done. PM compliance tells you whether your preventive calendar was done on time. They use different denominators and run on different cadences.
| Schedule compliance | PM compliance | |
|---|---|---|
| Question it answers | How much of the frozen week got done? | Was the preventive calendar done inside its tolerance? |
| Denominator | The jobs in the frozen weekly schedule | PM and PdM work orders due in the period |
| Unit | Work orders or planned hours | Work orders |
| Cadence | Weekly | Monthly |
| Timing test | Done inside the week. The day doesn't matter | Done inside the tolerance window around the due date |
| Is early a problem? | No | Yes. Early is as bad as late |
| What breaks it | Emergency work, under-loading, poor estimates | Resource shortages, poor-quality PMs, floating due dates |
| Is 100% good? | No. Sustained 100% means the week was never full | Close to it is the target, provided the PM list is worth doing |
The 10% rule you'll see quoted belongs to the PM due window. A frozen week has no tolerance.
A worked example
The numbers below are illustrative.
After leave and training, and at your current wrench time assumption, your crew has 400 hours of capacity next week. Your job estimates are on wrench time too, so both sides are on the same basis.
The schedule locked at Thursday's review meeting holds 28 work orders and 280 planned hours. That's 70% loaded, and it's already the first finding.
By the end of the week, 22 of the 28 jobs are finished, worth 180 planned hours. 6 slipped. And 4 emergency jobs that didn't exist on Sunday broke in and were completed, taking 60 actual hours.
- Work orders: 22 ÷ 28 = 79%
- Hours vs scheduled: 180 ÷ 280 = 64%
- Hours vs available: 180 ÷ 400 = 45%
- Emergency work: 60 ÷ 280 = 21%
- Fill-in work: 0%
The 79% looks fine on a slide. The 45% says less than half your crew's available hours went on scheduled work.
The 21% tells you emergencies took a big bite. But 180 hours of scheduled work plus 60 hours of emergencies is 240 of your 400. So where did the rest go? That's your next question, and it usually points at your estimates or your wrench time assumption.
And the 0% fill-in tells you the crew stuck to the schedule they were given, so the gap sits in the loading and the estimates.
A plant that reports "our schedule compliance is 79%" has told you almost nothing. Is that 79% of a full week, or of a half-empty one? Counted in jobs or in hours?
What to do with the number on Monday
Schedule compliance, emergency work and fill-in work are where I start. Trend them every week from day 1.
Estimating accuracy comes later, once your planner feedback loop works and you trust your time writing. Better estimates then lift your schedule compliance too. I've made the wider case in why most planning and scheduling efforts fail.
I always insist you start your weekly schedule review meeting with last week's scheduling metrics.
Then do these 5 things:
- Write your definition down. Jobs or hours, which denominator, and when the week locks.
- Publish the loading figure next to it. When TIP Trailer Services loaded the schedule properly, their scheduling compliance went up 20 points.
- Add emergency work and fill-in work. You need all 3 before any of them means much.
- Take the Sunday evening snapshot. Without a frozen baseline, the metric can be edited after the fact.
- Don't pay anyone on it. Palmer and I agree completely on this one.
You'll often see it claimed that higher schedule compliance raises wrench time. No published study measures schedule compliance and wrench time on the same plants. The mechanism makes sense: an under-loaded crew fills the spare time with delays, travel and slack, which is exactly what wrench time measures. But nobody has measured it.
So treat the link as a well-reasoned argument, and load your schedules properly anyway.



