A callback is the most expensive appointment on your schedule. It consumes a full slot, burns fuel and labour, earns nothing, and displaces a job you could have billed. Then it costs a second time, quietly, in a customer who is now slightly less sure about you.
And yet most service businesses cannot say what their callback rate is. It is nobody's number — the technician who goes back rarely enjoys reporting it, and the office logs it as a visit like any other.
Count them before you fix them
You need one field: whether a job was a callback, and a short reason. That is the entire measurement system, and it produces more actionable information in one month than most reporting projects.
Two numbers fall out of it. Callback rate — return visits divided by completed jobs — where anything above 5% is a system telling you something. And first-time fix rate, its inverse and the more motivating one, because people will chase a number that goes up.
Three causes cover most of them
Sort a month of logged callbacks and they concentrate almost entirely into three buckets, each with a different fix.
| Cause | What it looks like | Where it is actually fixed |
|---|---|---|
| Parts | The right diagnosis, without the part on the truck | The booking call and the van stock list |
| Diagnosis | The symptom was treated, the cause was not | Training, and a second opinion on repeat symptoms |
| Expectation | The work was correct and the customer expected something else | What was written on the quote and said at handover |
The third one surprises people. A meaningful share of "callbacks" are not defects at all — the job was done properly and the customer believed something else was included. That is not a technical failure, it is a quoting and handover failure, and it is the cheapest of the three to eliminate.
Parts: fixed on the phone, not in the van
Parts callbacks feel like a stocking problem and are usually an information problem. The technician arrives knowing the customer said "it is not heating." They do not know the make, the model, the age, or that the same unit was repaired eighteen months ago.
- Capture equipment details at booking — make, model, approximate age. Two extra questions on the call, and the technician arrives able to prepare.
- Keep equipment history on the property, not on the job. What was replaced, when, and by whom. Second visits to the same address should never start from zero.
- Stock the van from your own callback data. After a month of logs you will know exactly which five parts you keep driving back for.
A missed call answers itself in 90 seconds — while she is still holding the phone.
Built on the truck, sent before you pull away, with the chase already scheduled.
Photos and notes land on the customer record, not in someone's camera roll.
Invoice, review request, and next spring's tune-up — all without a Monday to-do list.
Diagnosis: watch the third visit
Diagnostic callbacks are the hardest to reduce and the most important to see, because they cluster. A specific fault type, a specific technician, a specific equipment model — the pattern is always there and is invisible without records.
The signal worth building a rule around is a third visit to the same property for the same symptom. At that point the diagnosis is wrong, and sending the same person a third time is not troubleshooting, it is a scheduled disappointment. Escalate it to your most experienced person automatically, before the customer has to insist.
Expectation: the callback that is not a callback
This is the pure profit fix. Nothing was wrong with the work; the customer thought the job included something it did not. Every one of these is preventable with words that cost nothing.
- 01Write exclusions on the quote
What is not included is more valuable to the customer than another line about what is. "Does not include drywall repair" prevents an entire visit.
- 02Say what "done" looks like at handover
Sixty seconds at the end of the job: what was fixed, what to expect over the next week, and what would genuinely warrant a call back.
- 03Photograph the finished work
A photo on the record settles a disagreement in one message rather than one visit, and it is already on the technician's phone.
- 04Distinguish warranty from new work in writing
The distinction is obvious to you and invisible to the customer. Naming it at handover, before it matters, prevents the argument entirely.
Handle the ones you do get properly
A callback handled well is not a neutral event — it can end with a stronger relationship than a job that went right first time. What determines that is speed and posture, both of which are choices.
- Prioritise it above new work. Counter-intuitive and correct: the customer who is already unhappy is the one who is about to write a review.
- Send the same technician where you can, unless the pattern says otherwise. Continuity reads as ownership; a new face reads as starting over.
- Do not argue the boundary on site. If there is a question about whether it is warranty, fix it and settle the classification afterwards. Winning that argument in a customer's hallway is the most expensive way to be right.
- Close it visibly. A short message afterwards confirming what was found and what was done turns the incident into evidence that you handle problems well.
What the rate tells you
Once you have three months of data, the rate stops being a scorecard and becomes a diagnostic — it points at a specific department.
- Under 3% — healthy. Look at whether you are over-cautious on scheduling instead.
- 3–5% — normal, with room. The parts bucket is usually the cheapest win here.
- 5–10% — one cause is dominating. The log will name it within a month.
- Over 10% — this is not a technician problem. It is a booking, training, or quoting problem wearing a technician's uniform.
A callback is not an accident. It is the same three mistakes, repeating, in a business that never wrote them down.
Common questions
What is a good callback rate for a service business?
Under 5% of completed jobs, with under 3% being genuinely strong. Above 5%, one cause is usually dominating — parts, diagnosis, or expectation — and a single logged field with a short reason will identify which within a month.
How much does a callback actually cost?
Roughly twice the visit. You pay for an unbilled truck roll and lose the billable job that would have filled the slot, before counting the reputational cost with a customer whose confidence has just dropped.
How do I reduce callbacks?
Capture equipment make, model, and age at booking so technicians arrive prepared; keep repair history on the property rather than the job; write exclusions on quotes and explain at handover what "done" looks like; and stock vans from your own callback data rather than by guesswork.
Should the same technician handle the callback?
Usually yes — continuity reads as ownership and they already know the property. The exception is the third visit for the same symptom, which should escalate automatically to your most experienced person, because at that point the diagnosis is wrong and repeating it is a scheduled disappointment.