Missed-call math

What a Missed After-Hours Call Actually Costs a Small HVAC Shop

By Daniel Graham, founder of AgentHive Inc, Palm Coast, Florida. 25 years in enterprise IT and telecom.

You know the feeling. You wake up, check your phone, and see three missed calls from numbers you don't know. One left a voicemail. Two didn't. By the time you call back, one of them says, "We already got somebody, thanks."

It's easy to shrug that off. But it's worth doing the math once, because a missed call costs more than the one repair. Here's how to work it out for your shop.

A quick note on the numbers: every figure below is an example made up to show the math. They are not industry averages. Swap in your own numbers from your call log and your invoices. That's the only version that counts.

Step 1: Count the calls that were real jobs

Not every missed call is lost money. Some are wrong numbers or robocalls. Count only the calls that would have turned into work.

Example: say you miss 4 after-hours calls in a typical week. Going through the voicemails and callbacks, about half are real service requests. That's 2 real jobs a week.

Now the harder question: how many hired somebody else before you called back? "We're all set" or no answer usually means they moved on.

Example: say 1 of those 2 hires another shop. That's 1 lost job a week, or 52 lost customers a year.

Step 2: Put a dollar figure on the first job

Pull your average repair ticket from your invoices. Use repairs, not installs. A night call is about something that's broken.

Example: say your average repair ticket is $450.

1 lost job a week × $450 × 52 weeks = $23,400 a year in lost revenue.

But revenue isn't what you keep. Parts, labor and fuel come out of it, so run it through your gross margin.

Example: say you keep 40 cents of every repair dollar after parts and direct labor.

$23,400 × 0.40 = $9,360 a year in lost gross profit.

Step 3: Add the work that would have come after

The first repair is rarely the whole relationship. A customer you fix tonight might sign up for maintenance, call you when the system needs replacing, and tell the neighbor. When they hire someone else at 11 p.m., all of that goes with them. Keep these guesses conservative.

Maintenance agreements. Example: say 1 in 4 repair customers signs up for a maintenance plan at $200 a year. Out of 52 lost customers, that's 13 plans, or $2,600 a year in revenue. And it comes back every year they stay with you.

Replacements. Example: say 1 in 20 repair customers eventually replaces their system with whoever fixed it last. Out of 52 lost customers, that's 2 or 3 replacements you never get to bid. At an example price of $9,000 each, that's $18,000 to $27,000 in future revenue. Not all this year, but real.

Referrals. Worth something, but you can't count them, so leave them out of the total.

These follow-on numbers are revenue, so run them through your margin too before you compare them to anything. The exact figure depends on your shop. The point is that the call is the front door to a customer, not a single job.

Step 4: Count what it costs you to catch them

The other side of the ledger is what you already pay to cover nights, written down or not.

Your own sleep. If you're the night line, you pay in bad mornings. Don't call it free.
An on-call tech. Whatever you pay in on-call or callout pay.
An answering service. Check your current invoice, the monthly total, not just the plan price.
Voicemail. Free, and it's how the calls in Step 1 got lost.

Step 5: Work out your break-even

Here's the test for any fix, whether it's a person, a service or software. First, find your gross profit per saved job.

Example: $450 ticket × 0.40 margin = $180 gross profit per repair.

Divide the monthly cost of the fix by that number. That's how many saved jobs a month it takes to pay for itself. If it's lower than the jobs you're losing now, the fix pays. If it's higher, keep what you have. If the fix has a one-time setup cost, divide that by the same number to see how many saved jobs it takes to earn it back.

In the example, 1 lost job a week is about 4 a month. A fix that needs 2 saved jobs a month to pay for itself would clear that bar. One that needs 6 wouldn't.

Run your own numbers

Fill these in from your last month or two:

Real jobs lost to another shop per week: ____
Average repair ticket: $____
Gross margin on repairs: ____%
Follow-on revenue you'd expect per customer (maintenance, replacements): $____
What you pay now to cover nights: $____ a month

Then work it out:

Lost repair profit a year: line 1 × line 2 × 52 × line 3.
Lost follow-on profit from one year of lost customers: line 1 × 52 × line 4 × line 3.
Break-even for any fix: its monthly cost ÷ (line 2 × line 3) = saved jobs a month it needs.

If you don't track missed calls yet, start a log this week: date, time, what they needed, and whether you got the job.

If the math says it's worth fixing

If your number is bigger than you'd like, you can take the phone yourself, rotate it through a tech, use an answering service, or use an AI phone desk.

One option is First Deploy AI from AgentHive Inc. It's an after-hours desk for HVAC, plumbing and electrical shops: a voice line that answers night leads, books them and follows up. Setup is $1,750 (start with $875, the balance at completion), then $250 a month per company from go-live. Run it through the same test. With the example's $180 per repair, the $250 a month is covered by fewer than two saved repairs a month ($250 ÷ $180 = about 1.4), and the $1,750 setup is earned back by fewer than ten ($1,750 ÷ $180 = about 9.7). To see how it would fit your shop, book a free 30-minute call at firstdeploy.ai.

Either way, run the numbers. Your call log will tell you more than any sales pitch.