The repair you have to do twice
A technician drives out, diagnoses the problem, and discovers they need a part they didn't bring. They leave, order the part, and come back days later to finish. Two trips, two windows of resident availability, two rounds of scheduling, for one repair.
That's a callback, and it's one of the most underrated cost drivers in maintenance operations. Every callback roughly doubles the labor and travel for a job that should have taken one visit. It also doubles the disruption to the resident and chips away at their confidence that you can actually fix things. A high callback rate is a quiet, persistent tax on both your budget and your reputation.
The good news is that callbacks are diagnosable. They don't happen randomly. They cluster around a handful of specific failures in the intake-to-dispatch process, and most of those failures are fixable upstream, before the technician ever gets in the truck.
Why callbacks happen
When you trace callbacks back to their root cause, they fall into a few recurring patterns.
The technician arrived without the right parts
The most common callback by far. The job description was vague, so the tech guessed at what they'd need, guessed wrong, and had to return. "Faucet problem" tells you nothing about whether they need a cartridge, an aerator, or a whole new fixture.
The wrong trade was dispatched
What looked like a plumbing issue turned out to be electrical, or vice versa. The first tech can't do the work, so a second tech with the right skills has to come out.
The diagnosis was incomplete
The reported symptom was treated, but the underlying cause wasn't, so the problem recurs. The drip was fixed but the real issue was upstream pressure, and it's back in a week.
Access wasn't coordinated
The tech showed up but couldn't get in, or couldn't reach the relevant area, so the visit accomplished nothing and has to be repeated.
Notice what these have in common: almost all of them are information failures, not skill failures. The technician was competent. The problem is they showed up under-informed, and the information they needed existed, it just didn't make it to them.
The intake diagnosis problem
The single biggest lever on callbacks is the quality of the initial diagnosis, the information gathered when the request first comes in.
When a resident says "my sink is broken," a good intake process doesn't stop there. It asks the follow-up questions that determine what the technician needs to bring: Which sink? Is it leaking, clogged, or not running at all? Where is the water coming from? How long has it been happening? Is it getting worse? Those answers turn a vague report into a job the technician can prepare for completely.
Manually, this diagnostic intake is inconsistent. A coordinator who's busy takes the request at face value and moves on. A coordinator who has time digs in. The same problem gets a thorough intake or a shallow one depending on the office's load that hour, and the shallow ones become callbacks.
Automated intake fixes this by asking the diagnostic questions every time, regardless of how busy the office is. An AI intake agent can hold a short, structured diagnostic conversation with the resident, narrow down the actual problem, and capture the details a technician needs. "It's the kitchen sink, water is pooling under the cabinet, it started two days ago and is getting worse" is a dispatchable job. "Sink broken" is a callback waiting to happen.
This is part of what Castellan does on the intake side. The AI agent gathers a complete diagnostic picture before the work order is dispatched, so the technician arrives knowing what they're walking into and what to bring. Better information in means fewer second trips out.
Right trade, right parts, right first time
Good diagnosis feeds directly into the two decisions that prevent the most callbacks: which trade to send and what they need to bring.
When the intake captures enough detail, the dispatch can match the job to the correct trade with confidence, instead of guessing and risking a wasted visit by the wrong specialist. And when the job description is specific, the technician can load the right parts before leaving, instead of carrying a generic kit and hoping.
The payoff compounds. A job that's correctly diagnosed, dispatched to the right trade, and stocked with the right parts gets done in one visit. The resident is satisfied, the technician's time is used efficiently, and you pay for one trip instead of two.
Measuring to improve
You can't reduce callbacks if you don't track them, and most operators don't track them well.
- Define a callback clearly as a return visit for the same issue, and tag those work orders so you can count them
- Calculate your callback rate as a percentage of total work orders, and watch the trend over time
- Categorize the causes, since the fix for parts-related callbacks differs from the fix for wrong-trade callbacks
- Look for patterns by vendor, trade, and property, because a callback cluster around one vendor or one building tells you where to focus
The categories matter most. If your callbacks are mostly about missing parts, the fix is better intake diagnosis. If they're about wrong trades, the fix is better job classification at dispatch. If they cluster around one vendor, that's a vendor-quality conversation. The data points you at the right intervention instead of leaving you guessing.
The trust dimension
Callbacks cost money, but they cost something harder to recover too: resident confidence.
When a technician comes out, leaves, and has to return, the resident's read isn't "they're being thorough." It's "they couldn't fix it." Two or three callbacks on the same issue and the resident concludes that maintenance in your building doesn't work, and that conclusion follows them all the way to the renewal decision. A first-visit fix, by contrast, tells the resident their problem was taken seriously and handled competently.
Getting it right the first time is one of those rare improvements that cuts costs and builds trust at the same time. The lever is almost entirely upstream, in the quality of the information gathered at intake. Ask the right diagnostic questions every time, dispatch the right trade with the right parts, and the second trip stops being necessary. Your budget thanks you, and so do your residents.