Why most tech reviews go wrong
Without numbers, the techs who talk the most or complain the least look like the best ones. With too many numbers, nobody knows what matters and every tech finds one that makes them look good. The middle path is six numbers, the same for everyone, pulled from the jobs themselves, reviewed weekly and talked about in person.
The six numbers that matter
| Number | How to compute it | What it tells you |
|---|---|---|
| First-time fix rate | Jobs fixed on the first visit / all jobs that needed a fix | Diagnosis skill and truck stock |
| Callback rate | Jobs needing an unpaid return within 30 days / jobs completed | Quality of the work |
| Average ticket | Revenue from the tech's jobs / number of those jobs | Whether options are being offered |
| Close rate | Options or estimates accepted / options or estimates presented | How the tech explains and sells |
| Utilization | Hours on jobs (billable) / hours paid | Scheduling and drive time as much as the tech |
| Reviews | Reviews naming the tech, and their average rating | How customers felt |
Decide each definition once and write it down. What counts as a callback? Is a return for a part that was on order a callback? (Usually no.) Is a drain that clogs again in two weeks a callback? (Usually yes.) The definition matters more than the target.
A worked example
Example only, with made-up numbers. In one week at Harbor Line Plumbing, Taylor Brooks ran 22 service calls. Twenty needed a repair, and 18 of those were fixed on the first visit. One job needed an unpaid return. Taylor presented options on 14 jobs and 6 were accepted. Revenue on the 22 jobs was $9,240. Taylor was paid for 42 hours and spent 29 on jobs.
- First-time fix: 18 / 20 = 90%
- Callback rate: 1 / 22 = about 4.5%
- Average ticket: $9,240 / 22 = $420
- Close rate: 6 / 14 = about 43%
- Utilization: 29 / 42 = about 69%
None of these is good or bad on its own. They only mean something next to the same tech's last eight weeks and next to techs doing similar work.
Make the comparisons fair
- Compare like with like. A tech who runs drain calls all day will have a lower ticket than one who runs replacement estimates. Group techs by the kind of work, or compare each tech only with their own history.
- Use enough jobs. One week with eight calls is noise. Look at a rolling four to eight weeks.
- Separate what the tech controls. Utilization drops when dispatch sends a tech across town three times. Fix the board before blaming the tech; see how to reduce drive time between jobs.
- Watch pairs of numbers. A high ticket with a rising callback rate can mean overselling. A high first-time fix with a low ticket can mean nobody is offering options.
- New techs get a ramp. Measure them against a training plan, not against your ten-year lead.
Ride-alongs
A ride-along is the owner or a lead spending a few calls with a tech. Do it when a number moves, and also on a schedule so it is not a punishment. Watch the greeting, how the tech finds the problem, whether options are explained or just listed, the cleanup, and the ask for payment and a review. Afterwards, name one thing that went well and one thing to try next week. Not five. For building those habits, see how to train new technicians.
A sample weekly scorecard
Example only. Each tech sees their own row and the shop average. Arrows show the trend over the last four weeks.
| Tech | Jobs | First-time fix | Callbacks | Avg ticket | Close rate | Utilization | Reviews |
|---|---|---|---|---|---|---|---|
| Taylor Brooks | 22 | 90% up | 1 | $420 | 43% | 69% | 3 |
| Sam Rivera | 19 | 84% flat | 2 | $510 | 50% | 64% | 2 |
| Jordan Lee | 25 | 92% flat | 0 | $365 | 31% | 74% | 4 |
| Shop average | 22 | 89% | 1 | $425 | 41% | 69% | 3 |
Read it like a coach. Jordan fixes things first time and rarely comes back, but closes fewer options: a ride-along on how options are presented. Sam sells well but has two callbacks: look at those two jobs before anything else.
Run the weekly meeting in 10 minutes
- Share the scorecard before the meeting, not on a screen in front of everyone.
- One-on-one, look at the tech's trend, not the ranking.
- Ask what got in the way: parts, dispatch, a hard customer.
- Agree one thing to work on.
- Recognize wins out loud, in front of the crew.
If you pay on performance, tie pay to numbers the tech controls and that cannot be gamed alone. See how to pay technician commissions.
How SWITCHBOARD handles this
In SWITCHBOARD the numbers come from the jobs themselves: time is clocked from the job in the tech app, Good, Better and Best options and signatures live on the estimate, and crew performance, commissions and job margins after labor, parts and drive time are reported from the same records, so nobody has to fill in a spreadsheet on Friday.
Questions people ask
What is a good first-time fix rate?
It depends on the trade and the mix of work. Track your own rate per tech over several weeks and work on the trend rather than chasing an outside number. See how to improve first-time fix rate.
How often should I review technician performance?
Look at the numbers weekly, talk one-on-one briefly every week or two, and do a longer review a few times a year.
Should I rank techs against each other in public?
Usually not. Share each tech's own row and the shop average privately, and recognize wins in public.
Is utilization the tech's fault when it is low?
Often it is not. Drive time, gaps in the schedule and parts runs all lower utilization. Check dispatch before blaming the tech.
How do I count a callback?
Write a definition once, for example an unpaid return for the same problem within 30 days, and apply it the same way to every tech.