How Much Do Missed Calls Cost a Heating and Cooling Business?

Work it out from four numbers: missed calls, the share that were new work, the share you never won back, and your average first-job value. In the illustrative six-van firm below, 36 missed calls in a cold-snap week meant about seven lost jobs, roughly $2,700 in first visits, plus service plans and replacements that never started.

Before you panic at your phone report, know that most missed calls are not lost jobs. Repeat attempts from the same caller, suppliers, sales calls and existing customers asking where the engineer is can make up half the list. Price any fix against the new-work calls you failed to recover, or you will buy a solution sized for a problem twice as big as yours.

Follow me on Instagram@sagnikteaches

The four numbers behind a missed call's price

Lost first-job revenue per week =
  missed calls
  x share that were new work
  x share not recovered by a callback
  x share that would have booked if answered
  x average first-job value

• Missed calls. Your phone system's call report shows unanswered inbound calls. Count unique numbers, not attempts, because a worried customer with no heating may ring four times in ten minutes.
• Share that were new work. You only get this by checking the numbers against your customer list and listening to voicemails for a week. There is no shortcut.
• Share not recovered. Of the new-work callers, how many did you call back and book? The rest are the real loss. Speed matters here: someone with no cooling in a heatwave rarely waits an hour.
• Share that would have booked. Your normal booking rate for answered calls from new customers, which your job-management software can usually show you.
• Average first-job value. The first invoice from a new customer, usually a diagnostic visit and repair, not an installation.

Why a heating and cooling firm loses more per missed call

Three things push the cost above what a decorator or a landscaper loses from the same unanswered ring.

Connect on LinkedInSagnik Bhattacharya

Urgency. A boiler, furnace or air-conditioning failure is often an emergency to the caller. They are not comparing quotes; they are ringing down a list until someone answers. A missed call from a no-heat customer is usually a job gone to whoever picked up next, often within minutes.

Subscribe on YouTube@codingliquids

What follows the first job. A repair customer can become a maintenance-plan customer paying every year, and an ageing system you repaired this winter is a replacement lead worth thousands next year. The first invoice is the smallest part of what the call was worth.

Seasonal spikes. The calls arrive in waves. The first cold week of autumn or the first hot week of summer can double your inbound volume while your engineers are all out, which is exactly when the office stops answering.

A six-van firm's cold-snap week, costed

The firm below is made up, but its numbers are the kind a small operator will recognise. It has six engineers, one office coordinator and the owner. In the first cold week of the season the phone report shows 260 inbound calls and 36 unique numbers unanswered.

A normal week is quieter: about 20 missed numbers, nine of them new work, five not recovered, and three lost jobs worth $1,140. If the firm has twelve peak weeks and forty normal weeks a year, the first-job loss is 12 × $2,660 plus 40 × $1,140, which is $77,520.

Now add what those customers would have become. That is 204 lost customers a year (84 in peak weeks, 120 in normal ones). If 15% would have taken a $180-a-year maintenance plan and kept it for four years, that is 31 plans worth $22,320. If 4% needed a replacement system within two years at $9,000, that is eight installs worth $72,000. Total revenue at stake: about $171,800 a year.

Revenue is not the true cost, because each job has materials and labour. But the vans, the engineers' salaries and the office are already paid for, so the margin on jobs you fail to book is high. At an assumed 40% gross margin across the lot, missed calls are costing this firm roughly $68,700 a year in gross profit. Put your own figures in and your answer might be a fifth of that or double it. The point is to have a number before choosing a fix.

The same sum for a one-van owner-operator

The answer changes with the shape of the business, which is why nobody's industry average is much use. Take a one-engineer firm where the owner answers the phone himself. He misses 12 calls a week because he's under a boiler or on a roof. Six are new work, he recovers two with callbacks from the van, and of the four left, three would have booked at an average of $320. That's $960 a week in a busy spell. His phone is unanswered for a different reason than the six-van firm's: not an overloaded office, but no office. So the fix is different too. An AI receptionist or answering service that takes details and books routine visits into his calendar matters more to him than to a firm with a coordinator, and the cost of a missed call is a bigger share of his week.

Missed calls you already paid for

If you advertise, some of those calls cost money before anyone picks up. Suppose search ads bring in calls at an average of $40 each. Ten missed ad calls a week is $400 spent to send customers to a competitor, before counting the lost job. Most call-tracking tools and ad platforms can show which calls came from ads, and these are the ones to answer first. It also changes the maths on fixes: an answering service that saves five ad calls a week has recovered $200 of ad spend alone.

The hours when heating and cooling calls go unanswered

Missed calls cluster. Pull a week of call data, sort by time, and you will usually find a pattern like this one from the same illustrative firm:

The fix follows from the pattern. If most misses are after hours, a daytime temp won't help. If most are at 8 a.m., a full after-hours service is more than you need.

Count your own missed calls in five working days

1. Export the call report from your phone system or VoIP provider for the last full week: time, caller number, answered or not, duration.
2. Remove repeat attempts so each number appears once per day.
3. Match numbers to your customer list. Your job-management software or CRM can flag which are existing customers.
4. Listen to the voicemails and tag each unknown number: new repair, new install enquiry, supplier, sales, wrong number.
5. Log callbacks: how many minutes until you called back, and whether they booked.

A filled-in log for one morning looks like this:

Day  Time   Number   Answered  Callback  Type               Outcome
Mon  07:52  ...4471  No        38 min    New - no heat      Booked elsewhere
Mon  08:05  ...4471  No        (repeat)  -                  -
Mon  08:11  ...9820  No        12 min    Existing - ETA     Reassured
Mon  08:26  ...3306  No        15 min    New - AC quote     Booked survey
Mon  08:40  ...7715  No        none      Sales call         -
Mon  12:47  ...6093  No        55 min    New - boiler leak  No answer, lost

Two things show up on a single morning: the no-heat caller was gone after 38 minutes, while the quote enquiry was happy to wait. That tells you urgent calls need an answer, and quote requests need a fast callback.

Listening to every voicemail is the slow part of the count. If your phone system transcribes voicemails, an assistant can do the first sort. A prompt:

Classify each voicemail transcript below as one of:
new - no heat/no hot water, new - cooling fault, new - install or
quote, existing customer, supplier, sales/spam, unclear.
Give the caller's number, the category and one line on urgency.
Don't guess: if it's not clear, say "unclear".
[paste transcripts with numbers and times]

Illustrative output:

...4471 | new - no heat | urgent: elderly parent, no heating since last night
...3306 | new - install or quote | not urgent: wants AC quote for summer
...9820 | existing customer | asking engineer's arrival time
...5518 | unclear | message cut off after name

Check the "unclear" ones yourself and spot-check a few others. A misfiled no-heat call is the error that matters, so read every "new" line.

Calls the phone report counts wrongly

Phone reports overstate some misses and hide others. Adjust for these before you settle on a number:

• Hang-ups within a few seconds. Often wrong numbers or pocket dials. Many systems let you filter out calls under five seconds; check a handful before excluding them all.
• Calls to engineers' mobiles. Regular customers often ring the engineer they know. Those misses never reach the office report, so ask the engineers to note theirs for the same week.
• Calls forwarded out of hours. If evening calls divert to the owner's phone, the office report may show them as answered when the owner let them ring out.
• Callers who rang back and got through. Removing repeat attempts per day handles most of these, but a caller who tries at 17:55 and again at 08:05 shows as a miss and an answer on different days. Match numbers across days too.

In the illustrative firm, the engineers' mobiles would add another four missed new-work calls on top of the 36 numbers in the table, so a count from the office phone alone understates the cost.

What each fix costs next to what it recovers

Set these against the illustrative firm's roughly $68,700 of yearly gross profit at risk and even the dearest option can pay, but only if it actually books the jobs. The detailed trade-off for this sector is in AI vs human answering services for HVAC firms, and general pricing is in what an AI receptionist costs a trades business.

What the caller hears before and after the fix

Before, the firm's after-hours caller got this voicemail greeting:

You've reached the office. We're closed right now. Please leave a message and we'll get back to you.

Nothing in that tells a customer with no heating whether anyone will come tonight, so they hang up and ring the next firm. After the change, an unanswered call triggers a text within a minute:

Sorry we missed you, our engineers are out on jobs. Reply with your address and what's wrong (no heat, no hot water, cooling not working, or a quote) and we'll call back within 20 minutes. If you smell gas, leave the property and call your gas emergency number now.

The safety line is not optional wording. Any automated reply or AI receptionist on a heating line needs a hard rule for gas smells and carbon-monoxide alarms that sends callers to the emergency service first. How to set those rules is covered in setting up an AI answering service that never misses an emergency, and the text-back setup itself in missed-call text back for a service business.

An after-hours call handled by an AI receptionist

If you choose an AI receptionist for evenings, test it with realistic calls before customers hear it. An illustrative transcript from such a test:

AI: Thanks for calling, you've reached the after-hours line.
    Is anyone at the property in danger, or can you smell gas?
Caller: No, nothing like that. The heating's just stopped and
    it's freezing. My mum's 84.
AI: I'm sorry, that sounds worrying. I can book the first
    engineer visit tomorrow at 8:00, or pass this to our on-call
    engineer tonight. Would you like the on-call engineer?
Caller: Yes please.
AI: I've sent your details to the on-call engineer, who will
    ring you back within 20 minutes. Can I confirm your address?

Two things are right: the safety question comes first, and a vulnerable person triggers the on-call route. What you'd fix: "within 20 minutes" is a promise the AI can't keep unless your on-call engineer actually commits to it, so the script should use whatever callback time is true. And the address should be confirmed before the handover, not after, so the engineer gets it in the first message. The scripting side is covered in detail in the call scripts and escalation tutorial listed at the end.

Proving the fix won back jobs

Measure the same things for four weeks after the change as in the five-day count before it:

• Unique new-work numbers not recovered per week. This is the number that costs money. It should fall sharply.
• Median minutes to callback for the ones a person still has to return.
• Bookings from first-time callers per week, from your job-management software.
• Cost of the fix per recovered job. If an AI receptionist costs $150 a month and books eight extra first visits, that is under $19 a job.

Compare like with like: a cold-snap week after the change against a cold-snap week before it, not against a quiet week. A fuller method for the first three months is in measuring an AI receptionist's return in the first 90 days. And once new customers are coming through, the maintenance-plan revenue in the example only appears if someone asks them, which is where service plan reminders for heating and cooling customers come in.

Further reads

• AI Call Summaries: Log Every Phone Enquiry in Your CRM — Log every answered enquiry so none slips after the call.
• How to Write Call Scripts and Escalation Rules for an AI Receptionist — Write the escalation rules for no-heat and gas calls.
• Can AI Answer My Business Phone? What It Can and Can't Handle — What an AI phone agent can and cannot handle.
• Jobber, Housecall Pro, and ServiceTitan AI Features Compared — Compare the AI features inside field-service software.
• How to Set Up an AI Receptionist Without Losing Callers — Set one up without losing callers in the first weeks.
• Replying to Trade Enquiries in 60 Seconds With AI — Why most automations can't reply in 60 seconds, the three builds that can, and what a first reply to a roofing or plumbing enquiry should actually say.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Smith.ai pricing pages (AI receptionist and virtual receptionist plans); Ruby plans and pricing page; Jobber help article on Receptionist powered by Jobber AI. Checked September 2026.