First-Time Fix Rate: Sizing the Callback Leak and Where AI Can Help

First-time fix rate (FTFR) is the percentage of service calls you resolve on the first visit. A vendor benchmark discussed below (Aquant, cross-industry field service, not a contractor-only sample) reports a middle band around 77%. Callback rate is a different measure: the share of completed calls where the customer needs a return visit for the same problem, usually within 30 days, after the job was closed as fixed. A call that is not fixed on the first visit is not automatically a callback. Many are planned second visits, for example a part that has to be ordered or a quoted repair the customer approves later. The 77% vendor FTFR and the separate 4-6.5% callback planning range below have different definitions and evidence; one cannot be derived from the other. A callback usually bills nothing, and its cost lands in tech time, truck cost and, when your board is full, the paying job you couldn't run instead. At a 20-tech shop, that adds up every month.

If you're trying to get more out of the crew you already have, especially when hiring is constrained, first-time fix is where I'd start. It's the cleanest measure of whether your skilled hours come back as margin or get spent twice.

The benchmarks (and the cost)

One recent vendor dataset is Aquant's 2026 Field Service Benchmark, which Aquant describes as nearly 30 million service events across 161 organizations and $8.3B in service cost. Its February 19, 2026 release reports these values over three years. This verifies what the vendor published, not an independent audit of its raw data. The population spans field service industries, not a separately measured HVAC, plumbing or electrical sample, and is not Level data:

Separate planning ranges retained from this page's earlier worksheet are 7-12%, 4-6.5% and 3.5% or less for callbacks within 30 days. Their published provenance has not been established, so they are illustrative comparison bands, not Aquant findings or measured contractor percentiles. Validate your own callback definition and distribution before setting a target.

Read that last row again. In that vendor population, failed visits are reported at about a quarter of service cost in the middle band and nearly half at the bottom. Failed visits are a broader category than callbacks. Treat the row as a reason to measure your own failed-visit cost, not as your number.

The per-callback cost is less settled. A figure of roughly $650 for a typical two-hour callback is commonly attributed to the Air Conditioning Contractors of America, covering tech time, overhead and the foregone billable call. We have not located or linked the underlying ACCA source. On this page, $650 is an illustrative planning assumption for a two-hour residential or light commercial callback, not a measured average. Build your own number with this table:

Keep the last row separate. If you put it into the per-callback cost, do not add revenue from freed hours again later.

Fictional example (illustrative assumptions only): a 20-tech shop runs 400 service calls a month, or 4,800 a year. At an assumed 7% callback rate, that is 28 callbacks a month and 336 a year. At the $650 planning assumption, that is $18,200 a month, or $218,400 a year. A 20% callback rate (80 a month, $52,000 a month) would be a stress case above the illustrative comparison bands, not a typical shop.

Five callback causes to code in your own log

This is the insight that reframes the whole thing. Use these five diagnostic categories when reviewing callbacks:

1. Incomplete diagnosis, the tech treated the symptom, not the root cause.
2. Missing parts, the right part wasn't on the truck, so the "fix" was partial.
3. Poor handoff, the next tech on the next visit started from zero.
4. Inadequate documentation, no diagnostic trail to follow.
5. Knowledge gaps, the junior tech didn't know what the senior tech knows.

Four of the five causes on this list are information failures, not hands-and-skill failures. That is a count of categories, not a measured share of callbacks. We have no source for how often each cause occurs, and the mix will differ by shop, trade and call type. The check is your own log: code every callback for 60 days with one primary cause, then count. Where the repair was within the tech's ability but the information needed to get it right didn't reach him in time, the fix is an information fix, not a training or hiring fix. A counterexample: if your log shows most callbacks trace to one install crew's workmanship or one defective part batch, the fix is supervision or a supplier claim, not better information.

The senior-tech bottleneck, in one number

Aquant reports measuring the "skills gap" as the difference in first-time fix rate between a company's top technicians and everyone else. In its vendor population (cross-industry, not a contractor-only or Level sample, with values confirmed in its February 2026 release, though raw data were not independently audited), that gap at best-in-class teams is 2.9 percentage points. At underperformers, it's 10 points. Your own check: FTFR for your top three techs versus everyone else, on the same job types over the same 90 days.

That's your senior-tech bottleneck quantified. When critical knowledge lives in two or three people's heads, everyone else's first-time fix sags, work routes back through your best techs, and they spend their scarce hours rescuing other people's calls instead of doing high-margin work. Spreading senior technicians' knowledge across the crew is one intervention to test, alongside parts availability, equipment mix and dispatch. Compare matched call types before crediting a smaller gap to the intervention.

What AI actually does here

This is the cleanest example of "AI for the trades AI can't replace" in practice. AI diagnostic tools don't make the repair. They are designed to get the right information to the person making the repair. These are the mechanisms such tools aim for, and each one is a claim to test in a pilot, not a guaranteed result:

• Surfaces the probable fault before the tech leaves the shop, based on the equipment, the symptom, and the history of similar calls, so the diagnosis starts ahead, not from scratch.
• Confirms the parts the likely repair needs, so the truck is stocked for the actual job and you skip the return trip for a part run.
• Generates a structured diagnostic trail every subsequent tech can follow, killing the handoff and documentation failures.
• Turns your best tech's fixes into reusable knowledge the junior techs can lean on, aiming to shrink the gap between your best techs and everyone else.

We have not linked or verified a specific case study for this kind of tool, so this page makes no claim about the FTFR change you should expect. If you pilot one, measure it this way:

1. Pick two or three job types.
2. Record 90 days of baseline FTFR and 30-day callback rate by tech.
3. Run the tool on those job types for the next 90 days, using the same definitions.
4. Compare against job types that did not get the tool, so a seasonal swing is not mistaken for a tool effect.

A vendor's published results describe that vendor's customers, not your shop or the trades as a whole. The point stands without a number: none of the mechanisms above require better technicians. They require better information reaching the technicians faster.

The gross profit per tech hour math

Here's why I care about this more than almost any other ops metric. A callback double-spends your scarcest asset, the skilled hour, and bills nothing for the second spend. So every avoided callback can do two things at once: it removes an unbillable return trip and, when paying work is waiting, frees that hour for a paying call.

Run it on the same fictional 20-tech shop at 400 calls/month. Cutting the callback rate from 7% (28 a month) to 3.5% (14 a month) avoids 14 callbacks a month. At the $650 planning assumption, that's $9,100 a month, or $109,200 a year. In the 20% stress case, cutting to 5% (inside the separate illustrative middle callback band) avoids 60 callbacks a month: $39,000 a month and $468,000 a year.

The $109,200 is modeled cost and opportunity exposure, not a guaranteed cash saving or gross-profit increase. Two cautions keep this honest. First, the $650 assumption already includes the foregone billable call, so do not add revenue from freed hours on top of it. If you want to show that separately, take the displaced-call margin out of the per-callback cost first. Second, that portion only becomes real money if the freed hours are actually filled with paying work. If the tech is paid for the hour either way and has nothing else to do, the cash saving is mostly truck and parts. Avoided callbacks reduce cost; they do not show up as billed revenue.

That's the whole "make more money with the same crew" thesis in one metric. You didn't hire anyone. You stopped paying for the same hour twice.

How to start

You can't fix what you don't measure, and most contractors don't track this weekly:

1. Measure first-time fix and callback rate weekly, by tech and by job type, not quarterly. Callbacks within 30 days for the same issue. If your FSM doesn't surface it cleanly, that's the gap (it's one of the numbers field software doesn't show you).
2. Stock trucks by call type, using history. In many shops a meaningful share of callbacks traces to the part not being there; your cause log from step 1 will tell you how much.
3. Pre-screen symptoms before dispatch so the tech arrives with the probable fault and the right parts.
4. Capture the senior tech's fixes as a diagnostic trail the rest of the crew can use.
5. Tie it back to gross profit per tech hour. Callbacks aren't a service-quality footnote, they can be one of the bigger leaks in your labor yield, and your own numbers will show how big.

First-time fix rate is the rare metric that's a customer-experience win, a margin win, and a capacity win at the same time. In a market where you can't hire your way out, that combination is as close to free growth as it gets.

FAQ

What is a good first-time fix rate?

Aquant's vendor-published benchmark, with its release checked but raw records not independently audited, puts the middle band at about 77% and top performers at 88%. That covers a cross-industry field service population, not a contractor-only sample. Your definition matters more than the band. Decide whether planned second visits for ordered parts count as a miss, and keep that rule fixed. Then track it weekly by technician and job type, because the aggregate number hides which techs and which call types are driving your callbacks.

What does a callback actually cost?

A figure of roughly $650 for a typical two-hour callback is often attributed to the Air Conditioning Contractors of America, covering technician time, overhead and the foregone billable call. We have not linked the underlying source, so treat $650 as an illustrative planning assumption. Because that figure already includes the foregone call, do not add lost margin on top of it. Build your own number: return-visit hours x loaded cost, plus truck cost, plus free parts, plus the margin of a displaced paying call only when your board was full.

Is a callback the same as a warranty visit?

Closely related, but not identical. A callback is a return visit for the same issue within about 30 days of a job closed as fixed. A warranty visit is defined by your warranty terms and may cover equipment failures that were not a workmanship or diagnosis miss. Most callbacks are unbillable. Most of the common causes are information failures (incomplete diagnosis, missing parts, poor handoff, weak documentation) rather than the technician's skill. How often each cause occurs at your shop is something to measure by coding callbacks, not something this page can tell you.

How does AI reduce callbacks if it can't do the repair?

AI diagnostic tools are designed to get the right information to the technician faster: suggesting a probable fault before dispatch, flagging the parts the likely repair needs, and building a diagnostic trail that spreads your best tech's knowledge to the rest of the crew. Whether a given tool does that well at your shop is something to test with the before-and-after measurement above, not assume. The repair stays human; the aim is to close the information gap. That's the idea behind the operating layer for the trades.

Why does first-time fix matter so much for a labor-short contractor?

Because a callback double-spends your scarcest asset. When you can't hire more techs, reducing avoidable return visits can free capacity within the hours you already pay for. Growth still depends on demand, pricing and whether freed hours become completed billable work. Cutting callbacks is one operating lever on revenue per truck and technician utilization.

---

If you can't say what your first-time fix rate was last week, or which techs and job types are driving your callbacks, that's the visibility gap, and it's costing you skilled hours you're paying for twice. Book a 15-minute call and we'll work through how to calculate it from your FSM job history and size what it's worth using your own costs. The first profitability audit is free.