Utilization, productivity, efficiency: three technician metrics owners mix up

Ask three service managers what "technician utilization" means and you'll get three answers, and at least one of them will actually be describing productivity. The words get used interchangeably on dashboards, in pay plans and in sales pitches for software. They aren't the same thing. They measure different parts of a tech's week, they're controlled by different people, and paying on the wrong one can make a shop worse while the number on the wall gets better.

This post gives precise definitions and formulas for utilization, productivity and efficiency, shows how they multiply together into one identity, works through three technicians' weeks to show how the metrics can point in opposite directions, and ends with an opinion on which to pay on.

Three definitions, precisely

The definitions below are the ones we recommend. Other shops and vendors use different ones, which is exactly why you should write yours down.

Utilization: how much of the paid day is billable

On-job hours run from Arrived to Completed on billable work. Paid hours are every hour on the timesheet: drive, on-job, shop time, supply-house runs, meetings, training and paid waiting.

The denominator has to include drive time. Under federal wage rules, travel from job site to job site during the workday must be counted as hours worked.¹ A utilization figure that leaves out drive hours is measuring a smaller day than the one you pay for.

Don't confuse it with booked capacity, which is booked hours ÷ bookable hours on the dispatch board. That's a forward-looking scheduling number. Utilization is what actually happened. The arrival windows post explains why booking the board to 100% backfires.

Efficiency: actual time against standard time

If your flat-rate price book says a condenser fan motor replacement takes 1.5 hours and the tech does it in 1.2, that job ran at 125% efficiency. Across a week, sum the standard hours of every task sold and divide by actual on-job hours.

Efficiency only works if your standard hours are honest. A price book with padded task times makes everyone look efficient; one with tight times makes everyone look slow. Efficiency is as much a test of the price book as of the tech.

Productivity: output per paid hour

The Bureau of Labor Statistics defines labor productivity as the efficiency with which goods and services are produced via labor hours, often referred to as output per hour.² For a technician, output is best measured in money.

Revenue per paid hour is easy to compute and easy to inflate with expensive equipment. Gross margin per paid hour (revenue minus materials minus the tech's own loaded labor cost) is harder to game and closer to what the business keeps.

The identity that ties them together

These three metrics aren't independent. Revenue per paid hour breaks down exactly into three factors:

The third factor, realized rate, is revenue per standard hour sold: what your pricing and selling actually captured for each hour of book time. It moves with price book rates, discounts, option mix and add-on sales.

The identity matters because each factor has a different owner:

When productivity drops, the identity tells you where to look. A tech whose productivity fell because their utilization fell probably had a dispatch or demand problem, not a performance problem.

A worked example: three technicians, one week

All numbers are hypothetical, except the wage inputs. Each tech is paid at the national mean HVAC mechanic wage of $31.14 an hour from the BLS May 2025 survey,³ with time and a half over 40 hours and benefits loaded at the BLS employer cost ratio for installation, maintenance and repair workers (benefits were 31.4% of compensation in June 2026).⁴ For context, production and nonsupervisory employees at specialty trade contractors averaged about 39.5 paid hours a week in August 2026, per BLS.⁵

Worked example. Wage and benefit rates from BLS; everything else hypothetical. Check the identity: 0.62 × 1.11 × $203 ≈ $140.

If you ranked these three on utilization, Tech B wins and Tech C comes last. Rank them on what they produce for the business and the order flips. Tech B spends more of the day on jobs, but takes longer than the standard on them and sells less per book hour. Tech C has a lower utilization, perhaps because of longer drives or a training morning, but works fast, diagnoses accurately and presents options well.

Neither ranking tells the whole story. Maybe Tech B is newer and should be slower. Maybe Tech C got the best calls. That's the point: one metric alone tells you almost nothing about a technician.

How each metric gets gamed

Every one of these metrics can be pushed up in ways that hurt the business. This isn't a character flaw in technicians. It's a well-documented property of measurement. The social scientist Donald Campbell argued in 1979 that the more a quantitative indicator is used for decision-making, the more it is subject to corruption pressures and the more it distorts the processes it's meant to monitor.⁶ The anthropologist Marilyn Strathern later put a related idea in the form most people know:⁷

Here's how that plays out with each metric.

The guards share a theme: pair every metric with one that moves the opposite way when it's gamed. Utilization with efficiency. Efficiency with callbacks. Revenue with margin and complaints.

What to do when each one slips

Because each factor has a different owner, each slip has a different first move.

• Utilization falls across the whole crew. It's almost never the techs. Check demand first: call volume, booking rate and unsold estimates. In shoulder season, the fix is filling the board (maintenance visits, follow-ups on open estimates), not pushing techs to move faster. Then check drive time per job; a board that routes techs across the metro can cost an hour a day.
• Utilization falls for one tech. Look at what's on their timesheet that isn't on-job: supply-house runs point at truck stock, long gaps point at dispatch, shop time points at a side task someone handed them.
• Efficiency falls for one tech. Ride along. Watch the diagnosis and the work. Check whether they're missing a tool or a part that the faster techs carry. If one task type is always slow for everyone, the standard time is wrong, not the crew.
• Realized rate falls. Check discounts and price overrides first, then whether options are being presented. A tech who quotes a single option every time will have a lower realized rate than one who presents three, with the same skill and speed.
• Gross margin per hour falls while revenue per hour holds. Materials or labor cost moved. Look at supplier price changes that never made it into the price book, and at overtime.

Which one to pay on

This is opinion, with reasons.

Don't pay technicians on utilization. Techs control only part of it. Dispatch, call volume, weather, parts and drive distance control the rest. Paying on it rewards techs for stretching jobs and punishes them for slow weeks they didn't cause. Utilization is a management metric: it tells you whether the office is feeding the field. Put it on the dispatcher's and service manager's scorecards.

Don't pay on efficiency alone. It depends on standard times the tech didn't write and rewards speed over care. Use it to coach, and to find which price book times are wrong.

Pay on productivity, measured as gross margin, with quality guards. Gross margin per paid hour (or a commission on gross margin, which is the same idea) rewards the things a tech controls: working well, diagnosing correctly, presenting options and selling what the customer needs. Guard it by netting out callbacks and refunds, and by giving fair credit for maintenance and warranty work so nobody avoids it.

Whatever you choose, it has to sit on top of legal minimums. Hours worked include drive time between jobs,¹ commission plans have their own overtime math, and some states add rules on top of federal law. The technician pay post covers how hourly, commission and hybrid plans work within wage law.

Measuring it without a spreadsheet war

Three data sources make these metrics reliable:

1. 1

   Time from job statuses

   Dispatched, En route, Arrived and Completed taps give drive and on-job time without anyone filling in a timesheet by hand. Check that techs tap reliably before you trust the numbers.

2. 2

   Standard hours on every task

   Every flat-rate task in the price book needs a standard time. Tasks without one make efficiency incomplete; track the share of on-job hours with a standard.

3. 3

   Costs on every job

   Materials at real cost and labor at the loaded rate per job turn revenue into gross margin. Without job costing, productivity defaults to revenue and the gaming starts.

In TradesWork, drive and on-job time are recorded automatically from those job statuses on the timesheet, and approving a timesheet week writes labor cost to each job, so gross margin per job uses real labor rather than an estimate. See timesheets and payroll and performance pay.

app.tradeswork.app · Timesheets with payroll export

Review the three metrics monthly, by technician, side by side, with the identity in mind. Weekly is too noisy for individual techs: one bad install day swings everything. The Monday KPI scorecard post covers which numbers belong on the weekly owner view instead.

A short glossary to pin on the wall

• Utilization: on-job hours ÷ paid hours. The office's number.
• Booked capacity: booked hours ÷ bookable hours. Tomorrow's board, not last week's work.
• Efficiency: standard hours sold ÷ actual on-job hours. The coaching number, and a test of your price book.
• Realized rate: revenue ÷ standard hours sold. The pricing and selling number.
• Productivity: gross margin ÷ paid hours. The pay number, with guards.
• Billable hours: the on-job hours that end up on an invoice.

Owners who keep these straight stop arguing about whose number is right and start asking better questions: is the board feeding the field, are the techs good at the work, and is the price book capturing what the work is worth. Those are three different problems, and now they have three different numbers.