Maintenance technician KPIs: nine metrics worth tracking

Most maintenance scorecards measure one thing: how many work orders somebody closed.

That number is easy to produce and it teaches the wrong behavior, because the fastest way to close more work orders is to close them badly and let them come back as new ones.

1. Callback rate is the one to add first

A work order closed on Tuesday that reopens on Friday for the same issue was not closed.

Callback rate is the percentage of completed work orders that generate a repeat visit for the same problem. It is the most honest quality signal in maintenance and most portfolios have never had the number.

Read it per technician and per category together. A tech with a high callback rate on plumbing and a normal rate everywhere else needs training on one thing, not a performance conversation.

2. First-time fix separates a training gap from a truck-stocking gap

Closely related and genuinely different.

A callback means the fix did not hold. A failed first-time fix means the tech could not complete the job on the first visit, usually because the part was not on the truck.

That distinction decides who owns the problem. Low first-time fix with normal callbacks is a parts and inventory issue. Normal first-time fix with high callbacks is a technique issue.

GetDone reports first-time fix on its own analytics tab for that reason.

3. Average days to close, read per category

The portfolio number is nearly useless. A plumbing backlog and a slow technician look identical in an average.

Read it per tech and per category. Then compare the same category across techs at the same property, which is the only comparison that controls for the property's condition.

4. Open backlog age, not backlog count

Forty open work orders is not information. Forty open work orders where nine are over 90 days is.

The 90-day view exists in GetDone specifically because the oldest open item is almost always the interesting one. A work order nobody has touched in three months is either not a real work order or it is a problem nobody wants.

5. Completion rate means nothing without its denominator

Percentage of assigned work completed in the period.

The reason to keep it despite its gameability: 40+ properties in the reference portfolio run at 100% work order closure. That is a meaningful operational standard, and it only means something because the closures are checked against callbacks.

6. Without adoption, the other eight metrics do not exist

This is the metric that makes the other eight possible, and it is the one most portfolios skip.

If techs do not log work, you have no data, and a performance program built on partial data is worse than none because it rewards whoever documents most rather than whoever performs best.

94% of maintenance techs at ResProp use GetDone. The other end of that distribution matters too: the lowest-adoption property in the portfolio sits at 46%. The average hides it, and the 46% is the property worth looking at.

7. High-risk workflow compliance

Water intrusion, pool, and light inspections are the work that becomes a claim.

The metric is not how many were completed. It is what share were completed with the required photos and approvals attached. A high-risk work order closed without documentation is a liability that looks like a completed task.

Before the workflows were enforced in our own portfolio, 116 high-risk work orders were open with 90 of them aged past two days.

8. Turn contribution, per tech

Make-ready is the maintenance work that lands directly in NOI, so the tech's contribution to turn time is worth isolating from general work orders.

At $50 a day for a vacant unit at $1,500 rent, two days off a turn is $100 back per turn. Attributing that per tech tells you who to learn from rather than only who to coach.

9. Resident-facing responsiveness

Time to first response, distinct from time to close.

A resident whose ticket is acknowledged the same day and resolved in four days is usually more satisfied than one resolved in two days with no communication. That is a real finding about human beings and it does not show up in closure metrics at all.

What measurement does to a team, which is the part that matters

Being direct about this, because it is where maintenance performance programs go wrong.

Used to find who needs a part in their truck, this data helps. Used to rank people weekly in a shared channel, it will cost you your best techs, who are the ones with options.

Two rules that seem to hold. Give techs their own numbers before you give them to anybody else, which is why GetDone has a per-technician My Analytics view. And never compare across properties without controlling for asset class and age, because a 1966 Class C generates different work than a 2025 lease-up and a comparison that ignores that is just a statement about the buildings.

The data comes from work orders techs already close. It is not location tracking and it is not a timer. Keeping it that way is most of what makes it acceptable to the people being measured.

Which of the nine does your team not currently have?