How Many Calls Is Your HVAC Shop Missing? What It Costs

It's 2:40 on a hot afternoon. You're in an attic with a condenser that won't start, and your phone buzzes in your pocket for the third time in ten minutes. You let it ring. By the time you climb down, two of those callers have left no voicemail. One of them was probably a no-cool call, and by now they have probably booked whichever shop picked up.

Most owners know this happens. Few know how often, and almost nobody knows what it costs them. This article covers both. It's built on the best published numbers we could find, with an honest note on where each one comes from, followed by a one-week method for measuring your own.

What the published numbers say

Three sources come up again and again. All three are worth reading, and all three come from companies that sell to the trades, so treat them as directional rather than gospel.

What was measured	HVAC	Plumbing	Source
Calls answered by a person	34%	74%	Invoca, Home Services Lead Conversion Benchmarks 2026 (trade split as reported by Skygnosis)
Calls that turned into a booked job	38%	43%	ServiceTitan platform data, as summarized by Skygnosis
Calls booked at shops with fewer than 5 techs (all trades)	24%		Same ServiceTitan data
Businesses that don't ask a phone lead to book the job (all home services)	55%		Invoca 2026

Two numbers stand out.

The size gap. In ServiceTitan's data, shops with fewer than 5 technicians book 24% of calls. Shops with 25 or more book 59% (per Skygnosis). The obvious explanation is staffing rather than skill: a big shop has someone whose whole job is answering the phone, and a small shop has an owner on a roof.

HVAC vs plumbing. Plumbing callers reached a person 74% of the time, compared with 34% for HVAC. We don't know why. It may be the call mix, the shops in Invoca's sample, or how calls get routed. It's a big enough gap to be worth checking against your own shop.

Read these numbers with care

Invoca sells call tracking and AI call tools. Its figures are averages across its own customers, who tend to be larger paid-media advertisers rather than typical two-truck shops. Invoca's own breakdown also shows how much call length matters. Across home services, 52% of all calls reached a person, 65% of calls over 15 seconds and 73% of calls over 30 seconds. Some "missed" calls are misdials, spam and hang-ups.
The ServiceTitan figures reached us second-hand. We took them from Skygnosis's write-up (Skygnosis builds phone and booking systems), not from a ServiceTitan report we could open ourselves. ServiceTitan's data comes from its own customers, and those customers skew larger than the average shop.
Neither source measured your shop. These are benchmarks, not your number. The rest of this article is about getting your number.

Why most owners don't know their number

Your phone system knows how many calls you missed, but nobody looks at that report. Voicemail makes the problem look smaller than it is. You see the callers who left a message, and you never see the ones who hung up and called the next company. It also blends together missed calls you couldn't avoid (two calls at once) and ones you could (lunch, the owner on a job, after hours with nobody on call).

Count your missed calls in one week

You don't need new software for this. You need your call log and about 30 minutes.

Export last week's inbound calls. Most VoIP providers, call-tracking tools (CallRail and similar) and business carriers can export a CSV from the web portal. If you only have a cell phone, scroll the recent-calls list and write the calls down.
Remove the noise. Delete calls from your own techs, suppliers and known spam numbers. If your export shows call duration, set aside unanswered calls that rang for only a second or two, since those are often misdials.
Collapse repeat callers. One customer who tried three times is one missed opportunity, not three.
Split by time. Count missed calls during business hours and after hours separately, because they need different fixes.
Check what happened next. For each missed caller, check whether you called back, how long the callback took, and whether they booked. If you can't tell, that's useful to know too.

At the end you should have four numbers: total real inbound callers, missed during hours, missed after hours, and missed callers who later booked. Repeat this for a second week if the first was unusual (a heat wave, a holiday).

What a missed call costs you

There is no trustworthy industry dollar figure for this, and we won't invent one. You can work out a defensible estimate from your own numbers:

Monthly cost of missed calls ≈ missed real callers per month × share that were new-job inquiries × your booking rate on answered inquiries × your average ticket × your gross margin

Missed real callers per month: from your one-week count, multiplied by about 4.3.
Share that were new-job inquiries: listen to a sample of your voicemails, or call a few of the numbers back. Some missed calls are existing customers, vendors or sales pitches.
Booking rate on answered inquiries: of the people who reached you and wanted work done, how many did you book? Your own books or FSM reports are better than any benchmark here.
Average ticket and gross margin: from your accounting. Use gross margin rather than revenue, because a job you never did also has no parts or labor cost.

A worked example with invented inputs, just to show the arithmetic: 40 missed real callers a month × 50% new-job inquiries × 60% booking rate × $450 average ticket × 45% gross margin ≈ $2,430 of gross profit a month at risk. Change any input and the answer moves a lot, so the result is only as good as the numbers you put in.

Two honest adjustments:

Not every missed caller is lost. Some leave a voicemail and wait, and some call back. Subtract the missed callers who later booked, which you counted in step 5 above.
The cost depends on your ad spend. If you pay per lead or per click, each missed call also wastes the money you spent to make the phone ring. Use your own cost per lead from your ad dashboard.

Fixes, from cheapest to most involved

Call back faster. The cheapest fix is a rule: every missed caller gets a callback within a set time during business hours. It costs nothing but attention.
Look at the booking gap, not just the answering gap. If Invoca's finding applies to your team and calls get answered without anyone asking to book the visit, that's a coaching problem. Fixing it is cheaper than any software.
Text back automatically. A missed-call text ("Sorry we missed you, what's going on? We'll get you on the schedule") gives callers who would otherwise hang up and try the next shop a reason to wait. It also gives you something in writing.
Handle after-hours intake. After hours, a text conversation can collect the problem, name and address, put a request into your field-service software and page whoever is on call for real emergencies.
Answering service or more office staff. This works, but it costs the most, and a traditional answering service usually takes a message rather than booking the job.

The recipe: missed-call text-back with after-hours intake

We built options 3 and 4 as a downloadable n8n workflow for shops on Jobber: Missed-call text-back + after-hours AI intake.

When a call is missed, the caller gets a text within seconds, and only one per caller per 12 hours.
After hours, an AI model collects the problem, name and address by text and creates a request in Jobber. A burst pipe at 2am pages the on-call number. Anyone reporting a gas smell gets safety instructions instead of a booking.
Running costs on the vendors' published pricing come to roughly $8 to $17 a month at 60 missed calls, plus n8n and a one-time $19.50 US texting registration (A2P 10DLC). The recipe page shows the arithmetic.
What we tested: the whole flow ran in n8n against mock versions of Twilio and Jobber, with a small local AI model. It has not run against a live Twilio or Jobber account. The recipe page lists exactly what you need to check on yours, including one Jobber field we took from a community schema copy rather than from Jobber's docs.

If your answered calls aren't turning into bookings either, the call scorecard recipe scores recorded calls on "did we ask for the booking?" and sends a weekly coaching summary.

What a text can't do

A real emergency needs a person. Treat text-back as a safety net under your on-call process, not a replacement for it.
US business texting needs registration. Texts from a local number to US customers are blocked until your A2P 10DLC brand and campaign are approved, and approval takes time.
Text only people who called you, and honor STOP. Check your carrier's and your messaging provider's rules before you turn it on.

Sources

Invoca, The Invoca Home Services Lead Conversion Benchmarks Report (published July 2026). Call-tracking vendor; figures are averages across its customer base. invoca.com
Skygnosis, The calls you never knew you missed (Aug 2026). Secondary write-up of ServiceTitan booking-rate data and Invoca's trade split; Skygnosis builds phone and booking systems. skygnosis.com
Twilio, Jobber, OpenAI and n8n pricing pages, as linked and dated on the recipe page.

How many calls is your HVAC shop missing, and what it costs