Blog
SyX Solutions · July 21, 2026 · 13 min read
You already suspect the phone is costing you money. Somebody called while you were under a house or on a roof, and by the time you looked there were four missed calls and no voicemails. The question is not whether that is bad. It is how bad, in dollars, well enough that you could defend the number to your accountant. This article walks the arithmetic all the way through, labels every input, shows every division, and tells you where the estimate is weakest.
What this covers
A missed call is worth your average job value multiplied by the share of missed callers who would have booked. Multiply that by how many calls you miss and you have your annual loss. Three inputs, all of them yours: missed calls per week, average job value, and booking rate. Nothing else belongs in the formula.
Written out, the whole model is two lines:
That is the entire thing. Any article handing you a single universal figure for what a missed call is worth has quietly assumed your job value and your booking rate on your behalf, and assumed them to make the number impressive. The missed call calculator runs these same two lines against your inputs.
We are not publishing the percentages you have seen everywhere, because we could not trace them to a primary study we were able to read. That includes the three in almost every answering-service pitch deck: the share of callers who reach voicemail and never call back, the share of buyers who purchase from whichever business responds first, and the share of inbound calls that businesses miss. Those figures circulate constantly, get re-cited across marketing pages, and lose their original source along the way.
When a page tells you that some fixed share of missed callers never return, it has just supplied your booking rate for you, without knowing whether you sell $200 tune-ups to homeowners who will happily wait until Thursday or $40,000 replacements to contractors holding four other bids. One number cannot describe both.
Here is the calculation at two very different scales, every input labelled and every division shown.
Residential plumbing service calls, appliance repair, lawn maintenance. Lots of calls, modest tickets.
Jobs lost per year: 624 × 0.25 = 156. Revenue lost per year: 156 × $450 = $70,200. The value of a single missed call: $70,200 divided by 624 = $112.50. You can get there directly too, because $450 × 0.25 = $112.50. The per-call figure never depended on call volume at all. Volume only scales it.
Roof replacement, full HVAC changeouts, underground utility work. Half as many calls, far bigger jobs, and a lower booking rate because big-ticket buyers shop.
Jobs lost per year: 312 × 0.15 = 46.8. Revenue lost per year: 46.8 × $6,200 = $290,160. Value of a single missed call: $290,160 divided by 312 = $930. Again the direct route agrees: $6,200 × 0.15 = $930.
| Input | Example A | Example B |
|---|---|---|
| Missed calls per week | 12 | 6 |
| Missed calls per year | 624 | 312 |
| Average job value | $450 | $6,200 |
| Booking rate (assumed) | 25% | 15% |
| Value of one missed call | $112.50 | $930.00 |
| Annual loss | $70,200 | $290,160 |
Read the bottom two rows again. Example B misses half as many calls as Example A and loses more than four times as much money. A single missed call is worth roughly eight times more there. One universal per-call figure would have been wrong for both owners, and wrong in opposite directions.
The two-line model is deliberately conservative and it leaves money on the table in two ways. Neither is easy to quantify honestly, so we are describing them rather than assigning them a number.
The first is the caller who does not come back. The model treats a missed call as a coin flip on one job. In practice, a homeowner who reached somebody else has acquired a plumber, or a roofer, or an HVAC company, and that is the number they call next time and the name they give their neighbour. Whether that is worth two more jobs or twenty depends on your repeat rate, and your invoicing history is the only place that answer lives.
The second cuts the other way, which is why it belongs here rather than in a sales pitch. A missed call is only truly lost once the caller finds somebody who picks up. Until then the job is still yours to recover. That is why the booking-rate input is a fraction and not one hundred percent, and why the recovery window is short. Call back in four minutes and you are frequently still in the running. Call back the next morning and the customer already has a technician scheduled.
A better voicemail greeting and a faster callback discipline both run after the caller has started dialling the next number. The fix is the call being answered live, the first time, which is the whole reason voice agents exist as a category.
The person most likely to grab the ringing phone in a service business is the person whose hour carries the highest documented market price. That is the awkward arithmetic at the centre of "I will just answer it myself".
In the Bureau of Labor Statistics May 2025 wage survey, the mean annual wage for roofers was $58,140 across national employment of 135,490. For heating, air conditioning and refrigeration mechanics and installers it was $64,780 across employment of 409,670.
Turn those into hourly figures. The convention that makes the BLS numbers reconcile internally is 2,080 working hours a year, 40 hours a week for 52 weeks. Label that as the assumption it is, because a contractor in peak season is not working a 2,080-hour year, and dividing by a bigger number produces a lower hourly figure.
Now compare that to the occupation whose actual job is answering the phone. BLS puts the mean hourly wage for receptionists at $18.97 and the mean annual wage at $39,460, across national employment of 910,180. Those two figures reconcile against each other: $18.97 × 2,080 = $39,458, which is the published annual mean to within rounding. In Texas the mean annual wage for receptionists is $35,720 across 78,330 people employed.
One caveat before anybody builds a business case on that $39,460. It is a wage figure and nothing more. Payroll taxes, benefits, paid time off, equipment, and turnover all sit on top of it, and we do not have a verified multiplier for the fully loaded number, so we are not going to state one.
What the comparison does support is narrower and still useful. When a technician stops work to take a call, the business spends an hour priced at roughly $31.14 doing work the market prices at roughly $18.97. Put volume behind it: twelve calls a day at four minutes each is 48 minutes, or 0.8 of an hour. At the HVAC hourly figure that is 0.8 × $31.14 = $24.91 a day, and across 250 working days, $6,228 a year. That counts only the clock.
Once you have your loss figure, the only question left is what the fix costs.
Ruby bills on included minutes. Published plans run $250 a month for 50 minutes, $395 for 100, $720 for 200, and $1,725 for 500. Divide price by included minutes and the effective rates are $250 / 50 = $5.00 a minute at entry, $720 / 200 = $3.60 at the plan Ruby marks most popular, and $1,725 / 500 = $3.45 at the top. Ruby states that "There are no additional or hidden fees for activation, onboarding, setup, customization, or coverage during certain periods". The overage rate is not published on that page, which is worth asking about before signing.
PATLive also bills on minutes and does publish its overage. Pay as you go is $75 a month with no included minutes at $2.60 per minute. Its most popular plan is $460 a month for 200 minutes with additional minutes at $2.20, an effective $460 / 200 = $2.30 a minute before any overage. The top published plan is $1,170 for 600 minutes. Bilingual receptionists are $20 a month and multiple scripts are $20 a month per script.
Smith.ai bills per call instead of per minute: $300 a month for 30 calls with overage at $11.50 a call, $810 for 90 calls, and $2,100 for 300 calls with overage at $8.50. Divide and included calls cost $300 / 30 = $10.00 each at entry and $2,100 / 300 = $7.00 each at the top. Smith.ai explains the model in its own words: "Billing per client call saves you money, and there's no mystery accounting involved. You're trusting us to represent you. You wouldn't rush a client off the phone and we don't either."
Now put those next to your own figure. Example A valued a missed call at $112.50 against a per-call answering price of $7.00 to $11.50. Example B valued a missed call at $930. The comparison is not close in either case.
Smith.ai's pricing FAQ states that "Even our most expensive plan, the pro plan, is about $33,000 less than the yearly salary of an in-house receptionist." Work backwards. The Pro plan is $2,100 a month, or $25,200 a year. Add the claimed $33,000 gap and the implied receptionist salary is $58,200. BLS puts the mean annual wage for receptionists at $39,460 and the median at $38,010. The implied figure sits roughly 47 percent above the national mean wage.
Read that fairly rather than as a gotcha. As a fully loaded cost of an employee, wages plus payroll taxes plus benefits plus equipment plus paid time off, $58,200 is defensible. The issue is that it is described as a salary, and a salary and the total cost of employing somebody are two different numbers. Carry the lesson into every sales conversation you have this year, ours included: when a vendor compares itself to hiring, ask which figure they used.
This is an estimate, and you should present it to yourself as a range rather than a single confident figure. The three inputs are not equally reliable.
Missed calls per week is your strongest input. Your carrier logs every inbound call, so this is measured, not assumed. Pull ninety days rather than one week, because a single week during a heat wave will not describe your year, then divide the total by roughly 13.
Average job value is nearly as strong. It comes off your invoices. The one trap is mixing job types. If you run $200 service calls and $9,000 replacements out of the same phone number, a blended average describes neither. Run the calculation separately for each type and add the results.
Booking rate is the assumption, and it is where nearly all of the error lives. The best proxy is your own answered-call booking rate: take every call you did connect over the last ninety days and divide the ones that became paying jobs by the total. It is imperfect, because a caller who reached you may differ systematically from one who did not. So run the number twice, low and high. In Example A, a 10 percent booking rate gives 624 × 0.10 × $450 = $28,080, while 30 percent gives 624 × 0.30 × $450 = $84,240. Same call log, same ticket size, a threefold spread. Any tool that hides that spread is not being straight with you.
One more limit. The model counts revenue, not profit. For the number that actually hits your bank account, multiply the annual loss by your gross margin on that work. It will be smaller and truer.
Pull ninety days of call records, pull your average invoice, pick a low and a high booking rate, and run both. Run it through the missed call calculator if you want it done in one screen instead of on the back of a work order.
Then compare the low end of your range, not the high end, against the published prices above. If the conservative version of your loss still dwarfs the cost of having the phone answered, you have your answer and you did not need anybody's statistic to get there. If it does not, you have learned something equally valuable and saved yourself a monthly bill.
If the number is large enough to act on, what we build is simple to describe: a phone that gets answered and books the job. It answers 24 hours a day. It reads real availability, so the times it offers out loud are times it can actually book. It writes a real appointment record into the calendar. It routes by service area rather than round-robin. It discloses recording in its own first words. Every plan includes a block of minutes rather than a fictional unlimited allowance, setup is $0, the minimum term is three months, and minutes past the block are $0.50. Price is quoted on the call, because the build is scoped to your services. For a specific trade, see the contractors page, the HVAC write-up, or the Jason Underground case study, a delivered build for a Greater Houston underground utility contractor.
Want to hear it before you talk to anybody? Call our own front desk at (346) 514-5030. That line is answered by the same kind of agent we build, and we run our own business on it. Then book a 30-minute discovery call at syx.solutions/book and we will scope it against your actual call volume and services.
There is no universal figure. One missed call is worth your average job value multiplied by the share of callers who would have booked. At a $450 job and a 25 percent booking rate that is $112.50. At a $6,200 job and 15 percent it is $930.
Because we could not trace them to a primary study we were able to read. The claims about voicemail callers who never return, buyers who purchase from whoever responds first, and the share of calls businesses miss are repeated constantly without a readable source behind them. We do not publish figures we cannot check.
Your own. Take the calls you did answer over the last ninety days and divide the ones that became paying jobs by the total. Then run the calculation at a lower and a higher figure, because this input carries almost all of the uncertainty.
No, and the calculation should not assume it does. A call is only lost once the caller reaches somebody else who picks up. That is what the booking-rate fraction represents, and it is why the recovery window is measured in minutes rather than days.
Published market prices vary widely by billing model. Ruby lists $250 a month for 50 included minutes up to $1,725 for 500. PATLive lists $75 a month pay as you go at $2.60 per minute up to $1,170 for 600 minutes. Smith.ai bills per call, from $300 a month for 30 calls to $2,100 for 300. Compare those against your own loss figure, not against a vendor's ROI claim.
Your hour has a documented market price. BLS puts the mean annual wage at $64,780 for HVAC technicians and $58,140 for roofers, roughly $31.14 and $27.95 an hour at 2,080 hours. The mean hourly wage for receptionists is $18.97. The owner or technician grabbing the phone is usually the most expensive person available to answer it.
As accurate as your inputs. Call count and average job value are measured from records you already hold. Booking rate is assumed, and it is where the error lives. Treat the output as a range, decide against the low end, and multiply by gross margin for profit impact rather than revenue.
Effective per-minute and per-call figures here are our own division of a published price by its published included volume. Wage figures are wages only and do not represent an employer's fully loaded cost of employment.
Keep Reading
This article is part of the SyX Solutions blog. Next, read why every call you missed last week went to the guy who picked up, or see how an answering service and an AI receptionist actually compare on price. If you would rather fix the number than measure it, start with missed call automation or the AI voice agent build.
Run your number, then fix it
Book a 30-minute discovery call. We scope it against your real call volume and your real services, and you get a price on that call. Want proof first? Call our own front desk at (346) 514-5030 and listen to the same kind of agent we build.