Missed Calls Cost HVAC Companies: The Evidence + Math

One methodological detail bounds that 34%. Invoca reports the home-services answer rate three ways by call length: 52% of all inbound calls, 65% of calls over 15 seconds, and 73% of calls over 30 seconds. Short rings include misdials and hangups. A raw missed-call count pulled from your carrier portal sits at the 52% end of that spread; the number describing callers genuinely trying to reach you sits nearer 73%.

Vintage matters as much as denominator. The same vendor has published four home-services answer figures in three years:

Those four are different vintages, different denominators and one trade-level cut, not contradictions to be averaged. Blending them into a single "X% of calls go unanswered" is how the bad numbers get made.

The most-repeated bad number shows the mechanism. Voksha's missed-call guide asserts that "62% of small business calls go unanswered during business hours," credited to BIA/Kelsey with no link. The figure traces, via Numa's round-up of business-phone statistics, to a 2016 411 Locals study: "411 Locals monitored phone calls of 85 businesses for 30 days," spread "across 58 industries," and "on average, they were only able to answer 37.8 percent of all inbound calls," which Numa headlines as "Small and mid-sized businesses miss about 62% of calls." Eighty-five businesses, one month, ten years ago, no trade breakdown.

Answer rate and booking rate are different metrics, and conflating them doubles the apparent problem. Answer rate is the share of inbound calls a person picks up; booking rate is the share of calls that become jobs. ServiceTitan's aggregated platform data, published October 2022 from more than 3,000 trade businesses in the US and Canada, put the June 2022 all-trades call booking rate at 42% and the HVAC figure at 38% — plumbing 43%, electrical 41%, garage door and water treatment 31%. The snapshot is one month, June, and ServiceTitan reports only that "HVAC booking rates varied with the seasons, with higher booking rates in shoulder seasons and lower booking rates in summer," never naming which months sit on which side of that line.

Two findings inside that dataset matter more than the headline. Shops with 25 or more technicians booked 59% of calls; shops with fewer than five booked 24%. Booking rates also collapsed by hour: large businesses fell from 61% at their morning peak to 21% after 6 p.m., small businesses from 26% to 9%. Bill Powers, ServiceTitan's Senior Industry Advisor, names the fix in staffing terms on that same page: "The faster a business can get to a dedicated call taker, the faster they will grow."

That 59% also corrects the widely quoted line that "average shops book 42% while elite shops hit 90%." ServiceTitan quotes three of its advisors who "all feel that a 90% call booking rate is attainable" — an opinion about a ceiling, not a measured cohort — and publishes the same 90% in its KPI chapter as a CSR target: "Call conversion ratio — ≥90%." The 59% is the highest cohort average in the June 2022 data, and one named business sits above it: Black Diamond Plumbing and Mechanical, "among the top 25 companies overall with a call booking rate in the 77% range." So 59% caps the averages, not the shops, and 90% remains an aspiration with no published cohort behind it.

SearchLight Digital's HVAC and plumbing advertising benchmark supplies the best-documented HVAC inputs for the first method, tracking $14.9M in Google Ads spend across 816 contractors and 8,077 campaigns in January 2026. Its HVAC General category shows a 36.0% book rate against a $3,102 average ticket, and AC Repair 37.0% against $3,174. Read the denominator before reusing either: SearchLight defines the metric on its own page as "the percentage of leads that result in a booked appointment" — leads, not answered calls. Caveats: one agency's own client book, one month, a heating month, and tickets attributed to paid-search customers rather than to all work.

The second method reads the price off the auction: what HVAC advertisers bid to buy an emergency call. Google Ads auction data pulled through DataForSEO on 17 August 2026 for the US [VENDOR — DataForSEO]:

Advertiser bids are not revenue, and none of this is a Google ranking metric. The auction sets a floor: a contractor competing on "emergency ac repair" spends tens of dollars per click for one phone call, and that click is spent whether or not anyone picks up. Invoca states the consequence in its 2026 report: "If that call goes unanswered, the budget that drove it is spent, and the job goes to the next company on the search results."

The dollar figures dominating this SERP have no such grounding. Three of them fail a source check:

Nextiva's page — the number-one organic result for this query — labels its own uncertainty: it publishes home-services per-call values of "$300–400, but sometimes upward of $1,200" alongside the admission that "voicemail and missed-call statistics vary, with many being vendor-published. This can make it challenging to offer definitive facts." It is still an unsourced range.

missed calls per week × 4.33 weeks × booking rate × average ticket = revenue lost per month

The 4.33 converts weeks to an average month (52 ÷ 12). Nothing else in the formula is an industry average, which is why the output is defensible when the inputs are yours. Where each of the three numbers lives inside your own systems is set out on the calculator page; what that page does not carry is the published figure sitting next to each input, and the denominator each figure is built on:

That middle row is the trap. The formula wants booked jobs divided by answered calls, and no vendor we could reach publishes that number for HVAC. Invoca's report adds a second filter: only 38% of answered home-services calls are classified as leads at all — the report's own gloss on that filter is "which calls are real job opportunities" — 45% in HVAC, so a leads-denominated book rate and a calls-denominated booking rate are not interchangeable.

The arithmetic below runs at an illustrative 36% booking rate against those two very different published tickets. Thirty-six percent is a placeholder, because nobody publishes a rate on this formula's denominator; your own rate rescales every cell in proportion. Cells are rounded to the nearest dollar:

The spread is the point: five missed calls a week costs $35,073 or $290,124 a year, driven only by which published ticket you plug in. Run the formula on your own three numbers.

Five steps make the inputs defensible rather than flattering:

1. Use a service ticket for service calls and a replacement ticket for replacement calls. Blending them, as the $3,102 column does, imports install revenue into a formula fed by repair calls.
2. Compute booking rate on answered calls only. Dividing booked jobs by all inbound calls double-counts the misses the formula is already pricing.
3. Keep short rings out of the count. Invoca publishes its answer rate at 15-second and 30-second thresholds precisely because raw inbound counts carry misdials, hangups and repeat robocalls.
4. Run it twice, on a shoulder month and a peak month. ServiceTitan measured HVAC booking rates higher in shoulder seasons and lower in summer; call volume by season is not published there.
5. Price the fix against the annual figure, not the monthly one. Coverage is bought as a subscription; the loss accrues every week you don't.

<!-- component: calculator — embed MissedCallCalculator (shared component) with the three inputs above; defaults blank, no baked-in industry averages; deep-links to /tools/missed-call-calculator -->

The nearest checkable 85% measures something else. Numa's round-up sources its voicemail section to Numa's own 2020 platform data and to two 2013 eVoice surveys — "67% of people don't listen to voicemail from business contacts," and only 18% listen to voicemail from numbers they don't recognise. Its Vonage 2019 entry sits in a different section and measures IVR behaviour, from 2,010 survey responses: "85% of consumers have abandoned a call after reaching an auto attendant." The two 85% figures are not the same measurement, and we could not reach a study behind the voicemail one.

Pew Research Center measured the behaviour on the other end of the callback. Pew surveyed 10,211 US adults from 13–19 July 2020 through its American Trends Panel and found that 19% of US adults say they generally pick up cellphone calls from unknown numbers. Sixty-seven percent say their practice is not to answer but to check a voicemail if one is left; 14% ignore both the call and the voicemail.

Your callback arrives on the homeowner's phone as an unrecognised number, and 19% generally answer those — while the caller who reached your voicemail at 9 p.m. with a dead furnace dials the next company.

The same Pew work cuts against the generational story attached to this topic. Voksha cites "a 2024 Pew Research survey" finding that 72% of people under 40 prefer not to leave voicemails; we could not locate any such Pew survey. What Pew did publish points the other way: "Americans ages 18 to 29 are more likely to take calls from unknown numbers than those in older age groups."

Answering is only half the failure. Invoca scored home-services call handling on six separate measures: 81% gave a proper greeting, 72% a proper closing, 62% obtained caller information, 58% scored as an excellent call, 45% converted the lead to a booked job, and 45% asked for the sale or appointment. The last two land on the same number, but the report treats them as distinct measures and nowhere states they describe the same calls. Its own research highlight states the business-level version instead: "55% of home services businesses don't ask leads to buy or book the job."

1. Change the routing before you buy anything. Ring groups, sequential and simultaneous routing, and an after-hours destination cost nothing beyond configuration on most VoIP systems. Invoca's own answer-rate playbook makes "Implement Ring Group Call Routing" its fifth strategy and pairs "Create after-hours alternatives" with "Implement dayparting in all paid media campaigns" under its third. Invoca ranks differently from us: its 2026 report puts AI agents first, "the single biggest lever for improving answer rates and speed-to-lead." We order by cost, and routing is the free move — our guide to running the closed hours covers what to point that overflow destination at.
2. Buy metered human coverage if your call mix is heavy on judgment: distressed callers, commercial accounts, contract terms. Human services bill per call or per minute, so the bill peaks in the weeks the calls are worth most. The meters, overage clauses and break-even volumes sit in our per-call versus per-minute versus flat comparison at 120 calls.
3. Buy flat AI coverage if your volume is mostly routine no-heat and no-cool work arriving outside staffed hours. Flat pricing means a cold-snap week and a quiet April week bill the same, and concurrency is a configuration limit rather than a hiring one. The calls where an AI receptionist still hands off to a person are worth reading before you buy.

Our own position in that list, stated plainly as a bias: HVAC Foundry sells the third option — an AI answering agent aimed at HVAC calls, quoted per shop rather than metered per call. If the number your calculator produced is small, the first fix is the whole answer and you should not buy anything. Our quote is set by your call mix and your escalation rules, so there is no sticker of ours for the cost side of that comparison — ask for one through the contact form. We publish no answer rate, coverage percentage or booking outcome of our own for it.