Planning Work Orders and Preventive Maintenance

Every building has more work than hours. A beginner works whatever is loudest. A good mid-level tech can look at 40 open requests, say which 10 matter this week, and defend that order to a manager. You also start shaping the preventive maintenance (PM) program itself, instead of just following it. This is planning work, and it decides whether your days are calm or spent chasing failures.

What You Will Be Able to Do

• Set a work order priority by consequence, not by who is asking
• Ask the intake questions that reveal a real emergency
• Work out your crew's real weekly capacity and backlog
• Rank assets by how much their failure would hurt
• Build and tune PM tasks and frequencies

Priority Means Consequence

A priority answers one question: what happens if this waits? How loud the caller is does not change the answer. Neither does their job title.

Most sites use four or five levels. Your site sets the names and the response targets, so learn them. This table shows a common shape.

Two questions move a request up or down:

1. Is it getting worse? A slow drip on a ceiling tile is routine until it sits above a server room or an electrical panel. Then it is urgent or an emergency.
2. What does it touch? A failed light in an office is routine. A failed light in an exit stairwell is a life safety item.

Questions to Ask at Intake

A short call now saves a wrong trip later. Ask these before you set a priority:

• What exactly are you seeing, smelling, or hearing?
• Where is it? What is below it and next to it?
• When did it start? Is it getting worse?
• Is anyone at risk? Has anything been shut off?
• Can you send a photo?

Worked example. A front desk worker calls in a "small leak" in a closet. You ask what is below it. The closet sits above the main electrical room. That makes it a level 1 no matter how small it looks. You stop the water first, protect the gear, and keep out of standing water until power is confirmed safe.

Planning Against Real Hours

The most common planning mistake is treating 40 paid hours as 40 hours of work. Tech time goes to travel, parts runs, meetings, and interruptions. The hours actually spent doing work are often called wrench time.

Worked example. You have 3 techs at 40 hours a week. Your site's records show about 70 percent of the time is hands-on. That gives 3 x 40 x 0.70, or 84 hours of real work each week. Your records also show about 24 hours a week go to emergencies and same-day calls. That leaves 60 hours to plan. If the PM schedule needs 38 hours this week, only 22 hours remain for routine repairs. Your percentages will differ, so use your own CMMS data.

Do not plan to 100 percent of those hours. Leave room for the call you cannot predict.

Measuring Backlog

Backlog is all approved work that has not been done, counted in hours and turned into weeks. Weeks are easier to read than a count of tickets.

Worked example. Open routine work adds up to 210 hours. Your crew can give routine work 22 hours a week. 210 divided by 22 is about 9.5 weeks. If a tenant's request waits nearly 10 weeks, the crew is too small for the demand, or too much is being let in, or too much time is lost to failures. A backlog of a few weeks is usually healthy, because it lets you plan. A backlog that grows month after month needs a decision from management. Show them the number.

A Good Weekly Plan

1. Pull all open level 2 and 3 work, and all PMs due.
2. Group jobs by area, so one trip does several things.
3. Check parts, keys, access, and any shutdowns before you schedule. A job that cannot be finished is not ready.
4. Assign the job to the person with the right skills and authorization.
5. Book tenant notices for anything that shuts off water, power, or air.
6. Review on Friday what got done, and why the rest did not.

Keep a status for work that is waiting, such as "Waiting on parts" or "Waiting on contractor". Then your open list shows the truth about what can be started.

Building a PM Plan

A PM plan should not be a copied list. It should follow from what your building has, what would hurt most, and what has failed before.

Step 1: List Your Assets

Every piece of equipment that needs service gets an entry in the CMMS, with a location, make, model, serial number, and install date. An asset that is not on the list gets no PM. That is how a rooftop unit goes three years without a filter.

Step 2: Rank by Criticality

Criticality means how much a failure would hurt. Score each asset on a few simple questions, 1 for low and 3 for high:

• Safety: Could failure hurt someone?
• Operations: Would people have to stop work or leave?
• Cost: How expensive is a failure or a replacement?
• Backup: Is there no spare or second unit?

Worked example. The only domestic booster pump in a 10-story building scores 2 for safety, 3 for operations, 2 for cost, and 3 for no backup, for a total of 10. A small exhaust fan in a storage room scores 1, 1, 1, and 1, for a total of 4. The pump gets detailed PMs, a spare parts plan, and first claim on your time. The fan gets a simple check.

Run some low-criticality items until they fail if that costs less than maintaining them. A cheap fan with no safety role may not deserve a monthly task. That is a planned decision, not neglect.

Step 3: Choose Tasks

PM tasks come from three places.

Write a task so that two different techs would do the same thing. "Check unit" tells nobody anything. A good task says what to measure and what counts as a problem:

• Record the temperature split. Write a work order if it falls outside 15 to 22 F.
• Inspect the belt for cracks and glazing. Replace it if it is worn.
• Read the filter pressure drop and replace it at the listed limit.

A number written down each visit also shows a trend. A motor running a little hotter each quarter warns you early.

Step 4: Set the Frequency

Start with what the manufacturer says. Then tune it from your own records.

• If a filter is always clean at 3 months, try 4 months.
• If a filter is always loaded at 2 months, shorten the interval.
• If a part fails between PMs, shorten the interval or change the task.

Do not stretch anything set by law, code, or an insurer. Bad conditions shorten intervals too. Roof units near a highway or a bakery exhaust need more attention.

Step 5: Build Routes and a Calendar

Group PMs by location and season, so that one roof trip covers all the roof tasks. Then spread them out, or your crew will have a 90-hour month in April and a 10-hour month in June.

Plan the season ahead of the demand. Cooling checks happen in spring, and heating checks happen in fall.

Scheduled Tests You Will Track

Many recurring tests are required, and your job is to see that they get done and that the records are kept. The rule that applies depends on your state, your insurer, and the codes your local fire authority enforces. These are common examples.

OSHA sets the extinguisher schedule for most workplaces, and it requires a record of the yearly check. Ask your fire authority or insurer which editions of the fire codes apply to you.

Is the Program Working?

Track a few numbers each month.

• PM compliance: PMs finished on time divided by PMs due. A target of 90 percent or higher is common.
• Planned versus reactive hours: More planned work means fewer emergencies. Sites often aim for far more planned hours than reactive hours.
• Repeat work orders: The same asset, same problem, within a set time. A high count means causes are not being fixed.
• Backlog in weeks: Steady is good. Growing means trouble.

Worked example. Last month your crew logged 300 hours of work: 195 planned and 105 reactive. That is 65 percent planned. If one asset caused 40 of those reactive hours, its PM needs a closer look, or it needs replacing. A number points you to the question.

Watch out for a compliance score that looks good and means little. Closing a PM without doing the work hides risk and ruins the data you plan with.

Deferred Work

Some work will be put off, and the decision to defer belongs to management. Your job is to make the risk clear. Keep a deferred list with the asset, what is needed, what happens if it fails, and a rough cost. A manager who has seen that list can make a real decision.

Common Mistakes

Check Your Understanding

1. A vice president demands that a scuffed wall in their office be painted today. A restroom on the same floor has a running toilet that is overflowing. Which comes first, and why?

The restroom. It is active water damage and a loss of function. The wall can wait, whoever is asking. Tell the vice president when the paint will happen.

2. A new supervisor books 160 hours of jobs for a 4-person team next week, because that is what payroll shows. Records put hands-on time at 65 percent. What do you tell them?

That plan will fail early in the week. 4 x 40 x 0.65 is 104 real hours, and emergencies still take a share, so book below that.

3. A director says things are improving because open tickets fell from 90 to 70. Those 70 tickets hold 210 hours of jobs, and you can spare 30 hours a week for them. Three months ago the figure was 5 weeks. What do you show the director?

210 divided by 30 is 7 weeks, up from 5. A ticket count hides how big each job is. Weeks show that demand is beating capacity, so look for what is driving it.

4. A coworker wants to change a quarterly fire pump test to twice a year because the crew is behind. What do you say?

No. A required test cannot be stretched by the crew. Only the rule owner can change it. Look for other work to move, and raise the staffing problem.

5. A filter PM is set for every 3 months, but your notes show these filters are always clogged at 2 months. What do you change?

Shorten the interval to fit the conditions, and record why. Your own history is a valid reason to change a task that is not required by law.

Key Takeaways

• Priority comes from what happens if the work waits.
• Plan against real hours, and leave room for what you cannot predict.
• Turn backlog into weeks, and watch the trend.
• Build PMs from your asset list, risk ranking, and failure history.
• Write tasks with a measurement and a limit.
• Tune frequencies from data, except for anything required by rule.