Maintenance KPIs: MTTR, MTBF and what every second visit costs

Five KPIs are all you need to run field service: response time, MTTR, MTBF, preventive maintenance compliance and first-time fix rate. The first three measure speed and reliability. The fourth measures discipline. And the last one is the one that drains the most money without anyone noticing: every service request that isn’t closed on the first visit is another truck roll, another morning of the schedule and one more customer waiting. Here’s what each one measures, how to calculate it and a calculator to put a price on your second visits.

The five KPIs at a glance

MTTR and MTBF come from industrial maintenance. MTTR stands for mean time to repair and MTBF for mean time between failures; both are defined in EN 13306, the European standard for maintenance terminology. First-time fix rate is the most widely used KPI in field service.

Response time: how long it takes someone to arrive

It’s the time from when the customer reports the problem until the technician arrives. In contracts with a service level agreement (SLA), it’s usually in writing: four hours for an emergency, twenty-four for everything else, or whatever was agreed. If it’s missed, there may be a penalty.

The usual mistake is measuring it from when the office logs the request instead of from when the customer reports it. A request that comes in by phone at seven in the evening and gets entered into the system at nine the next morning has a fourteen-hour response time, even if the system says zero. The customer counts from when they called.

It’s worth keeping it separate from resolution time, which runs from when the request comes in to when it’s closed. You can arrive in two hours and take a week to resolve it. That’s where the second visit shows up.

MTTR: mean time to repair

MTTR tells you how long it takes on average to get a piece of equipment working again, from when the repair starts until it’s running. You calculate it by adding up the time of all repairs in a period and dividing by the number of repairs.

In a factory, MTTR is measured by how long the machine is down. In field service you have to decide what counts, and almost always only the technician’s time in front of the equipment is counted. That hides the second visit. Two one-hour visits give a great MTTR even if the customer went three days without heat.

That’s why it’s worth calculating both: the technician’s MTTR, to know how much work each repair takes, and the customer’s resolution time, to know how long it really takes for the problem to be fixed.

MTBF: mean time between failures

MTBF tells you how long a piece of equipment runs on average between one failure and the next. You calculate it by dividing the operating time in a period by the number of failures in that period.

It’s measured per unit or per type of equipment, never for the whole company: an average that mixes wall-hung boilers with chillers tells you nothing. It’s useful for three things:

• Finding the equipment that fails most and suggesting the customer replace it, with the history in hand.
• Tuning preventive maintenance. If a unit fails every 600 hours, an inspection every 2,000 comes too late.
• Spotting repairs that weren’t properly closed. A failure that comes back on the same unit a few days later is usually the same failure, not a new one.

Together, they give you the equipment’s availability: MTBF ÷ (MTBF + MTTR). With an MTBF of 1,000 hours and an MTTR of 3, the chiller is available 99.7% of the time. It’s the number a maintenance contract customer cares about most, even though they’re almost never shown it.

Preventive maintenance compliance

Of the inspections scheduled in a period, how many were done on time. If 120 were scheduled and 102 were done on time, compliance is 85%.

It’s the first KPI to suffer when there are a lot of breakdowns: the technician who was booked for an inspection ends up at an emergency. For thermal installations in Spain, the minimum inspection frequency is also set by the RITE in its preventive maintenance program. A preventive visit that falls through is a missed legal obligation, not just a rescheduled appointment.

First-time fix rate: the KPI that costs the most

First-time fix rate is the percentage of service requests resolved on the first visit, without going back. If 150 out of 200 breakdown requests are closed the first time, the first-time fix rate is 75%. The other 50 are second visits, and sometimes third ones.

The most cited study is by Aberdeen Group, covering 156 field service companies, mostly in the United States, and it’s from 2013. It’s old, but we haven’t found a more recent public one, or one done in Spain. What it found:

• The average was 75%: one in four service requests needed another visit.
• When it wasn’t fixed the first time, it took 1.6 more visits on average.
• 17% of companies didn’t measure it.

And what matters most: the customer notices.

What counts as a second visit

Before measuring it, you have to decide what counts, and not change it afterward. Not every extra visit is a failure:

• They don’t count: jobs that take several visits by nature (an installation, a construction job, a commissioning), a different fault on the same equipment, or warranty work on a new installation.
• Everything else counts, including going back with the part ordered on the first visit. Being scheduled doesn’t mean it doesn’t cost money.

Some companies leave return visits agreed with the customer out of the calculation. You can, but then the number looks better than it is, and that’s exactly the most common cause.

What every second visit costs

An extra visit has a cost you pay and another you don’t see. The one you pay is the technician’s hours, including travel, and the vehicle. The one you don’t see is the schedule: while the technician goes back to a breakdown that was already looked at, they’re not taking a new request, not doing preventive maintenance and not billing a job.

How to work out the cost of one technician hour: gross annual salary × 1.3065 ÷ hours worked per year. The 30.65% is what an employer in Spain pays into Social Security in 2026 for a permanent contract (common contingencies, unemployment, FOGASA, vocational training and the intergenerational equity mechanism). On top of that comes the workplace accident rate, which depends on the activity. With €26,000 gross and 1,700 hours, it comes to about €20 an hour before accident insurance. That example is made up: plug in your own salary and your collective agreement hours.

With the starting values, a company with 200 breakdown requests a month makes about 80 extra visits: twenty technician days a month, almost one full technician dedicated to going back. They’re example numbers, but the math works the same with yours.

Why you have to go back: the seven causes

In the Aberdeen study, companies gave three main reasons for going back:

When you review second visits one by one, seven causes come up. In the order of the moment they could have been prevented:

Six of the seven are decided at the office, before the technician heads out. The sixth doesn’t appear in the studies; we found it reading customer reviews:

“A technician showed up at 9:30. He shut off the water and said he’d come back. Around 12 he left for other buildings without stopping by our home. At 6:30 nobody had come yet.”

— Public review of a maintenance company in Madrid, September 2025 (translated from Spanish)

The job was well prepared and the technician was the right one. An emergency took him away, and the customer was left worse off than before the first visit: without water.

How to measure it with what you already have

You don’t need new software. Just export the closed work orders from the last twelve months from your ERP or management software, with these columns: date the request came in, date of each visit, customer, address, equipment, type (corrective or preventive) and closing status.

1. Keep only corrective work. Inspections and installations go separately.
2. Count the visits on each work order. Work orders with more than one visit are your second visits.
3. Look for the ones that come back. A corrective job on the same equipment, or at the same address if you don’t track equipment, less than 30 days after another was closed, is usually the same failure, not properly closed. Many companies open it as a new work order, and so it never shows up.
4. Read twenty by hand. With the service request and the service report in front of you, note the cause of each one using the table above. Twenty cases are enough to see which one weighs most.

The third step is the one that surprises people most. If the equipment isn’t properly identified in the ERP, the same chiller shows up under three different names and repeat failures go unnoticed. It’s another reason for the service report to include make, model and serial number.

Measure the process, not each technician

The temptation is to pull each technician’s first-time fix rate and make a ranking. It backfires for three reasons. The most experienced technician gets the hardest breakdowns, so they go back more. Almost all the causes are decided at the office, before they head out. And a score per person means processing employee data, which at many companies in Spain has to go through the workers’ representatives.

What helps is measuring it by cause, by equipment type and by how the request came in. That tells you what to change.

The service request is where you win

Of the seven causes, four are solved by a well-taken service request: what equipment it is, what’s wrong with it, what’s been tried and when you can get in. That’s what we’re working on at Delegic: making sure every service request is complete before anyone is sent out, by phone, WhatsApp or the website. We cover it in AI agents for maintenance companies.

In short

Five KPIs are enough: response time, MTTR, MTBF, preventive maintenance compliance and first-time fix rate. The technician’s MTTR hides second visits; look at the customer’s resolution time too.

First-time fix rate is the one that costs the most. The average in the most cited study is 75%, and every request that isn’t closed takes 1.6 more visits. Six of its seven causes are decided at the office, before the technician heads out.