Callback Rate: Why Your Repeat Trips Are a Systems Problem, Not a Technician Problem
Between 4% and 6% of service calls end in a second trip for the same complaint. That is where most contractors landed when ACHR News polled HVAC shops about callbacks, and industry consultant Ruth King’s response was blunt: a 5% average is “way too high,” and the target should be under 2%. That poll ran in 2003. When I pull callback rate numbers with trades owners today, most of them are still sitting in that same band, and most of them did not know it until we looked.
Callback rate is the cleanest single number a trades owner has for one question: does the work come out right when you are not standing in the driveway? It is also the number owners most often misread. They see a callback, they think of a tech, and the conversation ends there.
Here are the questions I hear most from owners of 10 to 40 truck shops once callbacks start eating the margin, with straight answers.
What counts as a callback, and how do I measure it?
A callback is a return trip, at your cost, to fix the same complaint on a job you already closed. Not a new problem at the same address. Not a follow-up visit you sold. The same complaint, back again.
The formula is simple: callbacks divided by completed jobs over the same period. The discipline is in keeping it consistent.
- Pick a window and never change it. Thirty days is the most common. If you switch between 30 and 90 days, your trend line becomes fiction.
- Split service from install. An install callback and a repair callback have different causes and different costs. Blending them hides both.
- Tag every callback with a cause. Diagnosis, workmanship, part failure, missing part, customer education, or wrong information at booking. Six codes are enough. Without a cause code, you will only ever have an argument.
- Track by tech, by job type, and by day of week. The pattern usually shows up in one of those three cuts before it shows up anywhere else.
Most field service software will produce this number if someone tells it what a callback is. The software is rarely the obstacle. The definition is.
What does a callback actually cost me?
More than the drive. Run the math on a shop doing 6,000 service calls a year at a 6% callback rate: that is 360 unpaid return trips. At roughly two hours each with drive time, you are giving away about 720 labor hours, which is a third of a full-time technician’s year. Those hours also displace paying calls on a busy day, and the customer who had to take a second morning off work is now shopping for your replacement.
The construction side of the trades has measured this for decades. The Construction Industry Institute put direct rework at 2.4% of total contract value on standard industrial projects and 12.4% on civil and heavy industrial work, against typical pretax margins of only 3% to 4% cited in the same ENR report. Read those two numbers side by side. On a thin-margin job, rework is not a cost of doing business. It is the difference between a profitable year and a free one.
If you already run job costing for your contracts, callbacks are one of the leaks it should be catching. Most shops never attach the return trip to the original job, so the original job looks profitable and the callback disappears into overhead.
Isn’t this just a bad technician problem?
Sometimes. Rarely. W. Edwards Deming, the statistician who taught postwar Japanese manufacturers how to build quality in, estimated in Out of the Crisis that 94% of troubles and possibilities for improvement belong to the system, which is management’s responsibility, and only 6% to special causes. You can argue about the exact split. You cannot argue with the direction once you have looked at the data.
I learned this outside the trades first. In IT operations at a large telecom, we tracked repeat tickets: the same fault reopened within 30 days of being closed. The first instinct every time was to find the engineer who closed it. Almost every time we dug in, the cause sat upstream. A closure test that was optional. A ticket handed between shifts with half the notes missing. A definition of “resolved” that lived in a few senior people’s heads and nowhere else. When we made the test mandatory and wrote down what resolved meant, the repeat rate dropped without anyone being disciplined.
Trades shops run on the same physics. If one tech’s callback rate is double everyone else’s for two straight quarters with the same mix of work, that is a people conversation, and the right person, right seat lens applies. If the rate is high across the board, firing your way out of it only resets the clock.
Where do callbacks actually come from?
The techs who answered that ACHR poll were candid. They pointed to rushed diagnosis from techs racing to bill hours, required checks like fuel oil and manifold gas pressure getting skipped, weak quality control and poor callback documentation. Two decades later, those same causes show up when I map callbacks in client shops. They fall into four buckets.
The rushed call
The Construction Industry Institute’s rework research found scheduling was the single largest contributor at 25.4%, ahead of materials and design. In a service shop that means an overbooked board. When dispatch stacks eight calls on a tech who can do six well, the last two get diagnosed in the driveway with one eye on the clock. If your dispatching system rewards filling slots over finishing jobs, it is manufacturing callbacks.
The broken handoff
FMI and PlanGrid surveyed nearly 600 construction leaders and found that miscommunication and poor project data account for 48% of all rework on U.S. jobsites. In a service business, the handoffs are the CSR who books the call, the tech who runs it, and the installer who hands the customer to service. Every one of those is a chance for “the unit is short cycling on the second floor” to become “AC not working.”
The missing part
The tech found the problem, did not have the part, and installed something “to get them through.” That is not a workmanship failure. It is a truck stock and parts process failure, and it shows up as a callback three weeks later.
The closeout nobody checks
This is the biggest bucket in most shops I see. There is no written definition of done. Each tech has a personal version of the final checks, built from whoever did the training. The good ones do all of it. The rushed ones do what they remember.
One more cause hides behind the other four: how you pay. A straight per-call or flat-rate plan with no quality measure tells a tech that speed is the job. If you are restructuring comp, tie part of it to the callback number, as we covered in technician pay structure.
If you want an outside set of eyes on where your callbacks are coming from, or you want the tool to track it yourself, both options are below.
What does fixing it look like in a real shop?
A residential HVAC and plumbing company I worked with, about 22 trucks and a little over $7M in revenue, had never measured callbacks. The owner guessed 3%. We pulled 90 days out of their field software, defined a callback as a return for the same complaint within 30 days, and got just under 7%.
The breakdown changed the conversation. Callbacks clustered in two job types (intermittent no-heat calls and water heater replacements) and on Mondays, when the board carried the weekend backlog. The tech the owner was ready to let go had an above-average rate, but that tech was also the one dispatch sent every intermittent call to, because nobody else on the roster had as much experience. Adjusted for the work assigned, the rate was fine.
Four changes, over two quarters:
- A cause code on every callback, entered by the service manager, not the tech.
- A one-page closeout checklist for each of the top five job types, written by the two best techs and reviewed by the leadership team before it went live.
- A Monday dispatch cap, with the overflow pushed to Tuesday and the customer told so at booking.
- Callback rate on the weekly scorecard, with a goal and one named owner.
By the end of the second quarter, the rate was holding under 3%. Nobody was fired. The owner got back roughly a technician’s worth of capacity the company had been giving away for free.
Who should own the callback number?
One person. Usually the service manager, never “the team.” The Process brief in the 90u Library puts it plainly: if multiple people try to own a process, nobody truly owns it, and confusion spreads. The same rule applies to the number that tells you whether the process works.
Put callback rate on your weekly KPI scorecard with a goal, split service from install, and let it go red when it misses. A red number does not get debated in the meeting. It drops to the issues list and gets worked through with IDS: identify the real cause from the codes, discuss it, decide on one fix. Shops running their leadership meeting in Ninety.io can tie the scorecard measurable, the issue, and the documented closeout process together in one place, so the checklist the techs follow is the same one leadership reviews.
The closeout checklists themselves belong in your documented core processes. If you have not built those yet, creating SOPs your crew will actually use is the place to start, and field crew accountability covers the agreement conversation when someone keeps skipping a step.
What should I do this month?
Pull the last 90 days of completed jobs and count the returns for the same complaint within 30 days. Do it by hand if you have to. Then tag each one with a cause. You do not need a consultant or new software to do that, and the first look at the real number is usually enough to end the “it’s just a couple of guys” story.
Then pick the biggest cause bucket and make fixing it a Rock for next quarter, with one owner. Not all four buckets. One.
Callbacks are one of the most visible signs of a shop that still runs on the owner’s memory instead of a system. If that sounds familiar, the bigger picture is in our guide to systemizing a trades business, and scaling an HVAC business shows where the service manager seat fits once you are ready to hand this number to someone else.
If you would rather work through your callback data with someone who has done it in shops like yours, book a call. If you would rather build the scorecard and process yourself, start the free trial.
