What to record on a work order so the KPIs don't lie

Every KPI the manager asks about in the meeting starts on a work order filled out in the field. If the work order goes in with guessed times, generic equipment and "broke" as the cause, MTBF, MTTR and backlog come out wrong along with it, and so do the decisions made on top of them.

This article is for the people who fill out the work order: the technician, the mechanic, the electrician. It isn't about paperwork. It's about what each field turns into later, and about the few habits that make the number reflect what actually happened to the equipment.

From work order to KPI: the path of the data

The path is always the same: maintenance does the job, the technician records what was done, someone (or a system) organizes the records and the KPI comes out at the end. There's no shortcut. A KPI is never better than the data that went in at the start.

Here's where each field you fill in ends up:

Work order field	Becomes
Time of failure and time opened	MTTR and detection time
Start and end of work	MTTR, wrench time
Type (corrective, preventive, predictive)	% corrective, plan compliance
Equipment (tag)	MTBF per asset, equipment with repeat failures
Who worked and for how long	Labor hours, backlog, productivity
Cause and fix	Failure analysis, preventive plan
Parts and materials	Maintenance cost per asset
Link to a previous work order	Rework

Notice that almost every KPI depends on two or three fields at once. One wrong field is enough to spoil the math.

The 7 mistakes that spoil KPIs the most

1. Times written down from memory

The work order is filled out at the end of the day and the times come out "round": 8:00, 10:00, 2:00. MTTR ends up measuring the memory of whoever wrote it down, not how long the machine was down. Record the time when it happens: arrival, start, finish. On a phone, that takes one tap.

2. A stop that isn't a failure counted as a failure

MTBF and MTTR only consider stops caused by failures. A machine stopped for lack of scheduled production, setup, product changeover, preventive work or inspection doesn't count. If you open a corrective work order for everything that stopped the machine, MTBF drops and MTTR rises without anything having broken. When in doubt, record the real reason for the stop and make the classification clear.

3. Wrong job type

During a preventive task you found a bearing with play and replaced it on the same visit. That's corrective work found during preventive maintenance. If everything goes on the same preventive work order, the failure disappears from the equipment's history and the % corrective looks better than it is. The cleanest option is to open a corrective work order linked to the preventive one, or to flag the finding in a dedicated field.

4. Generic equipment

"Pump", "boiler area", "building B elevator". Without the asset tag, you can't calculate MTBF for each piece of equipment or find out which one keeps breaking. That repeat offender is exactly the equipment that deserves the most attention, and it's the one that disappears when the work order points to the area instead of the asset.

5. Hours instead of labor hours

Two technicians worked together for 4 hours on a gearbox. The equipment was down for 4 hours, but the job used 8 labor hours (one labor hour is one person working for one hour). Both numbers matter: the first goes into MTTR, the second into backlog, productivity and the planned vs. corrective split. Record who took part and how long each person worked.

6. "Defect" as the cause

"Broke", "defect", "replaced". That kind of record doesn't help anyone prevent the next failure. Write what you saw and what you did: "coupling-side bearing seized from lack of lubrication; replaced it and relubricated the opposite side". With that, you can run a root cause analysis and adjust the preventive plan.

7. A return with no link

The same failure came back two days later and someone opened a new work order, with no connection to the previous one. The rework becomes invisible: it looks like a new failure, not a job that had to be redone. When the problem comes back, link the new work order to the previous one.

Wrench time

One of the KPIs that depends on your time records is labor productivity, also called wrench time: the share of the shift you actually spend doing maintenance work.

This number isn't for pressuring the technician. It shows how much time the team loses to things outside its control. Here's an 8-hour shift (480 minutes), fictional but common:

Activity	Minutes
Doing the work	270
Travel between jobs	60
Fetching parts and tools from the storeroom	50
Waiting for the area or equipment to be released	40
Filling out paperwork and passing on information	30
Meetings, safety briefing and other	30
Total	480

Productivity is 270 ÷ 480 = 56%. The other 44% isn't slacking off. It's parts that weren't staged, an area that wasn't released and paperwork that could have been filled out on a phone. When you record these reasons, management can see where to invest: part kits staged before the job, clearance arranged with operations, routes that group nearby jobs.

The minimum every work order needs

• Equipment by tag, not by area.
• The right job type: corrective, preventive, predictive, inspection, improvement.
• Times recorded as they happen: failure (if corrective), arrival, start and finish.
• Who worked and for how long each.
• What was found and what was done, in a short, specific sentence.
• Parts and materials used, with quantities.
• A link to the previous work order, when it's a return.
• The reason for any wait, when the job stopped for parts, clearance or approval.

It looks like a lot, but in a well-built app most of it is one tap: the time with a button, the equipment read from a QR code, the type and reason picked from a list. In taskfield, the technician records all of this on a phone, even with no signal, and the system builds the KPIs on its own. Nobody has to retype the same work order into a spreadsheet.