Skip to main content

AI PHONE & LEAD RESPONSE · September 2026 · ~11 min read

Measuring your actual missed call rate before you fix it

Pull thirty days of call detail records from your carrier or call tracking, count every inbound call, and count how many a human answered. The gap between those two numbers is your missed call rate. It takes about an hour, and it is the only number that should decide whether you buy anything.

On this page

Nobody does this. What happens instead is that a vendor quotes a scary industry statistic, the owner recognizes the feeling described, and a purchase gets made against a number that was never about their business. The scary statistic is usually the same one, and you have almost certainly been shown it.

01

Where does the "62% of business calls go unanswered" number come from?

A 2016 blog post by a local SEO reseller called 411 Locals, published on 18 January 2016 and still live at 411locals.us. The post's own description of its method, in full: it monitored the phone calls of 85 businesses operating in 58 industries for a period of 30 days, and found that 37.8% of calls were answered, 37.8% went to voicemail, and 24.3% got no response at all.

Add the second and third figures. 37.8 plus 24.3 is 62.1. That addition is the entire basis of the most-quoted phone statistic in local marketing.

Look at what is missing. Eighty five businesses across 58 industries is roughly 1.5 businesses per industry. No country is stated. "Monitored" is never defined. There is no raw data and no way to tell an unanswered call from a call that rang a closed business at 2am. And the same post pivots into recommending that you hire a virtual secretary, which is a service the publisher sells.

The attributions you will see are false. Vendor pages credit the figure to Forbes, to a Ruby Receptionists study, and at least one to a ServiceTitan analysis of more than 50,000 contractor phone lines. ServiceTitan has published no such figure. A statistic with three mutually incompatible parents is telling you something. The date gets misreported too, usually as 2023 or 2024. It is January 2016.

The deeper problem is not the sourcing. A single cross industry missed call rate cannot be true of anything. Your business is not the average of a dentist, a divorce attorney, a plumber, and a realtor.

02

What should you compare yourself to instead?

Industry specific benchmarks, and only as context for the number you measured yourself.

CallRail, which sells call tracking software, published a benchmark report in January 2025 based on 1.1 million de-identified conversations. Its missed call rates: health care 32%, legal 28%, home services 14%, real estate 9%. CallRail's definition matters here, and it is stated: a missed call is an unanswered inbound call that resulted in a detected voicemail or was detected as abandoned.

A dentist at 30% is normal for the category and still bleeding. A realtor at 30% has something badly broken.

Dental has the best data in the category. Patient Prism scored 11,552,668 patient calls across 8,280 dental locations in calendar 2025 and found that 31 of every 100 calls were abandoned before reaching an agent. That is a queueing failure, not a nobody-picked-up failure, and the two have different fixes. Patient Prism's cohort is made of practices that already bought call intelligence software, so the true market number is plausibly worse than 31%, not better.

Restaurants show the widest spread, and it is entirely within the vertical. Revmo AI analyzed 12,091 real restaurant call recordings and published answer rates by segment: quick service 59.9%, fast casual 75.3%, full service 91%, pizza 93.1%. A quick service operator misses four calls in ten. A pizza shop misses seven in a hundred. Both are restaurants.

"Restaurants miss X% of calls" is not a meaningful sentence. The segment decides the number, not the vertical.

One more distinction, because vendors get it wrong constantly. ServiceTitan's platform data puts the call booking rate at a typical home services shop at 42%. That is not a 58% miss rate. CallRail's 14% counts answered calls. ServiceTitan's 42% counts booked jobs out of answered calls. Conflating the two is how the folk numbers get inflated.

Use all of these as a sanity check on your own measurement, never as a substitute for it.

03

How do you actually measure it?

Three sources, in order of quality.

Call tracking software. If you already have it, this is a report you can run today. It shows answered, missed, abandoned, and duration for every inbound call.

Your carrier's call detail records. Every phone provider can produce a CDR export. Ask support for thirty days of inbound records with disposition and duration. It is not always in the dashboard, and it is always available if you ask.

A manual log. If neither exists, tape a sheet next to the phone for two weeks. Date, time, answered yes or no. Writing it down changes behavior, so collect a third week after everyone stops noticing.

Then split the result three ways, because the total hides everything useful. By hour, by day of week, and by whether the number was new or an existing customer. A rate concentrated between 11am and 1pm is a lunch coverage problem with a free fix. The same rate spread evenly is a capacity problem you have to buy your way out of.

Where this measurement breaks down

Four places, and all four inflate or deflate the number in ways that change the decision.

Disposition codes lie. Most carriers mark a call "answered" the moment any device in the ring group picks up, including your own voicemail system. Check what your provider means by the word before you trust the column.

Hold abandonment hides inside "answered." A call picked up in two rings and parked for four minutes until the caller gave up is a lost customer recorded as a success. CDRs show duration. Anything short with no follow-on activity deserves a look.

Simultaneous calls disappear. If a second caller hits a busy line and your system returns a busy signal rather than queueing, some carriers never log the attempt. Ask specifically whether busy-signal attempts appear in the export.

Spam inflates everything. No published call taxonomy has a spam bucket. Maple's taxonomy of 1.2 million restaurant and local business calls puts everything outside its seven named intents into an "other" category worth 2%, which is implausibly small for spam given that local businesses publish their numbers everywhere. Treat your spam share as unmeasured until you count it yourself.

One compliance note, because measuring can quietly become recording. Call detail records are metadata. Listening to call audio is a different activity, and California requires the consent of all parties to record a call under Penal Code section 632.7. If your measurement project involves switching recording on to see what is happening, you have made a consent decision. The specifics of California consent and phone systems belong in front of your own attorney first.

04

What actually counts as a missed call?

Decide this before you count, and write it down, because it determines the number more than anything else.

My definition: a call is missed if a first time caller did not reach a human who could help them, and did not receive a callback the same business day.

That definition deliberately includes three things owners like to exclude. Calls answered and immediately put on hold for four minutes, where the caller hung up and you counted it as answered. Voicemails returned two days later, which are technically handled and commercially dead. And calls answered by someone who could not book, quote, or resolve, who promised a callback that never came.

On that second one, the genuine Harvard Business Review audit of 2,241 US companies found 37% responded to a web lead within an hour, 24% took more than a day, and 23% never responded at all. What the research shows about lead response time explains why one hour and twenty four hours are not close.

What I exclude: spam, robocalls, and existing customers with a routine question who reached voicemail and were happy to.

05

What do you do with the number once you have it?

Worked example

Multiply it into money before you shop. Here is the arithmetic with fabricated inputs so you can see the shape, and you should rerun every line with your own.

Step one, the count. Thirty days of CDRs shows 420 inbound calls. Of those, 63 were spam or robocalls, leaving 357 real calls. A human answered 300 of them.

Step two, the rate. 57 missed out of 357 real calls is a missed call rate of 16%derived. Leave the spam in and 57 out of 420 reads as 14%derived, which is CallRail's home services benchmark exactly, and wrong by two points because the denominator was wrong.

Step three, the money. Take your own close rate on inbound calls from your CRM, not from a benchmark. Say 30% of answered new-caller conversations become jobs, and your average ticket is $480. Then 57 missed calls a month, at $480 and a 30% close rate, is $8,208 a month, or $98,496 a year.

Step four, the decision. That is your annual leak and your budget ceiling in one figure. If the leak is smaller than the cheapest fix, do nothing and go work on something else. That happens more often than the category admits.

Step five, the sensitivity check, which nobody does. Rerun step three at a 15% close rate. The annual figure halves to $49,248. If your decision flips between those two numbers, go pull the real close rate before you buy anything.

ServiceTitan's platform data sizes the prize on the other side: every five point improvement in booking rate is worth roughly $100,000 in additional revenue for a shop with five to fourteen technicians, achievable with less than one additional booked call per weekday. It also explains why small operators miss calls. ServiceTitan's booking rate is 24% for shops with fewer than five technicians against 59% for shops with 25 or more. That gap is not discipline. The owner of the small shop is on a roof or under a sink.

If the leak is large, the measurement also tells you which fix. Concentrated at night means after hours coverage. Concentrated at lunch means scheduling. Spread evenly and rising means volume, and then you need to understand how call volume and overage pricing work before you sign anything.

Keep the baseline. Screenshot it, date it, file it.

The same discipline applies elsewhere. Whether financing options are visible at the right moment and which trust signals change behavior are both measurable, and both usually guessed at instead.

06

What to do this week

Request thirty days of call detail records from your carrier. Ask for an "inbound CDR export with disposition and duration." That phrase gets you to the right person faster.

Write your definition of a missed call in one sentence before you open the file, so you cannot move the goalposts once you see the number.

Ask your provider two questions in writing: what disposition code voicemail produces, and whether busy-signal attempts appear in the export.

Count, split by hour and by day, then do the money arithmetic including the sensitivity check.

If you decide to move forward, the information an AI receptionist needs before it goes live is next, and your call log is half of it already.

Be honest with yourself

When you do not need this

If your phone rings a handful of times a day and a person answers all of it, do not spend the hour. You already know.

If your measured rate comes in low, skip the purchase. A pizza shop near the segment answer rate of 93.1% will not earn back a phone system, and neither will a real estate office near the 9% benchmark. A low missed call rate is a reason not to buy, and most of this category will never tell you that.

If your business does not run on inbound calls at all, this is somebody else's problem. Plenty of businesses sell through forms, walk ins, and referral, and the phone is a courtesy line.

And if you have the number already and sit inside your category benchmark, go fix the next thing. Repeated measurement is a comfortable way to avoid a decision.

Sources

Related reading

10

Questions about your call log?

Email me at eric@seod.com with a screenshot or export of last month's call summary, with any customer numbers blacked out. I will calculate your missed call rate against a clear definition, split it by hour if the data allows, and tell you whether it is a coverage problem, a capacity problem, or not a problem.

It takes me fifteen minutes. If the answer is that your phone is fine, that is what you get back.

The rest of the AI phone and lead response library covers what comes after.

Call Eric Email Eric