Most missed-call statistics in circulation cannot be traced to a study. This report keeps only the figures with a named sample and a reachable source, states how old and how large each one is, and lists the popular numbers we removed and why.
62%
Of calls to small businesses unanswered (411 Locals, 2016, 85 businesses)
78%
Consumers who have abandoned a business after an unanswered call (CallRail, 2025)
82%
Would call a competitor instead (CallRail, 2025)
51%
Restaurant calls arriving after 5 PM (BrightLocal, 45,264 listings)
How many calls actually go unanswered
One frequently cited measurement exists, and it is smaller and older than its reputation suggests. Everything else on this question is either category-specific or unsourced. We give the sample size next to each figure so you can judge it.
62%
Of calls to small businesses went unanswered: 37.8% answered, 37.8% to voicemail, 24.3% no response at all
411 Locals, January 2016 - 85 businesses across 58 industries over 30 days, run by a marketing vendor. Small, old and not independent, but it is the study the 62% figure actually comes from.
91.0% / 59.9%
Inbound calls answered by full-service vs quick-service restaurants
Revmo AI analysis of 12,091 restaurant call recordings. Revmo sells restaurant voice AI and the full methodology sits behind an email gate, so treat it as the best available measurement rather than an independent one.
73 sec
Average hold time for full-service restaurant callers who were put on hold
Revmo AI, 12,091 call recordings
No figure
Industry-wide missed-call rate for healthcare, home services, legal or beauty. No first-party dataset we could find measures it, and the ranges that circulate name no sample, so this report does not carry one.
Use your own call log instead: count what rings out or hits voicemail, divide by total volume.
What callers do when nobody picks up
This is the best-evidenced part of the subject, because a survey can ask it directly. Note the gap between what people say they do and what platforms record them doing: we report both rather than the flattering one.
78%
Say they have abandoned a business after an unanswered call
CallRail, September 2025 - survey of 1,000 US consumers
82%
Say they would call a competitor if a business did not answer
CallRail, September 2025
21%
Say they call another business immediately, without waiting at all
CallRail, September 2025
42%
Say they leave a voicemail when a business does not pick up. This is stated intent, not measured behaviour, and the two must not be mixed.
CallRail, September 2025. The widely repeated counterpoint - that a message is actually left on only about 2% of calls - was attributed to Invoca platform data, but no reachable primary publication carrying that figure could be found, so this report gives no measured-behaviour figure at all.
When the calls arrive
Call timing is category-specific to the point where a blended figure is misleading. The one large first-party dataset reports by category, and the spread across categories is roughly five to one.
51% / 32%
Restaurant calls arriving after 5 PM, and at the weekend
BrightLocal study of 45,264 local-business listings across 36 industries
34% + 8%
Locksmith calls arriving after 5 PM, plus a further share before 9 AM
BrightLocal, 45,264 listings
11%
Healthcare calls arriving off-hours or at the weekend (10.0% weekday off-hours, 1.3% weekend)
Hyro analysis of 300,000 patient calls, 2023
94%
Of all local-business calls arrive Monday to Friday, across all categories in the sample
BrightLocal, 45,264 listings
Widely quoted figures this report does not use
These numbers appear on hundreds of vendor pages, including earlier versions of this one. Each was traced to its origin and each failed. They are listed here so the trace does not have to be repeated.
85%
"Never call back after one unanswered call." Traces to an uncited 2016 vendor blog post, not a study.
Every citation chain with a live hyperlink lands there. The BIA/Kelsey, Forbes and "Numa 2021 report" attributions all fail on inspection: none of those documents contains the claim.
80% / 86%
"Callers who reach voicemail leave no message." Traces to a 2014 Forbes contributor advertorial.
The behaviour is real. The number is not measured, and no measured figure available to us can be put in its place. See the stated-intent card in the second section.
~2%
"Voicemails are actually left on about 2% of calls." Attributed across this site to Invoca platform data. No such publication could be found.
Both URLs previously cited for it, invoca.com/blog/small-business-missed-call-statistics and invoca.com/blog/the-call-economy-statistics, return 404 and have no Internet Archive capture at any date, while invoca.com/blog itself is archived repeatedly. A moved page leaves a trail; these left none. Removed as unverifiable rather than re-sourced, and deliberately not replaced with a substitute number.
$1,200
"Average revenue per missed call." No study we could locate produces this figure.
Revenue per call is specific to your ticket size and close rate. Calculate it from your own numbers.
62%
Frequently miscast as a call-back figure. It is not: it is the share of calls left unanswered.
Stated correctly in the first section above. Swapping the claim type is the most common error on this subject.
Methodology
Every figure above was traced to a named primary source before it was included, and the sample size is stated next to it rather than left implied. Where a source is a vendor with a commercial interest in the answer, including 411 Locals, Revmo AI, CallRail and Hyro, we say so on the figure. Where surveys and platform data disagree, both numbers appear and we do not average them. An earlier version of this report attributed figures to Forbes, Forrester, BIA Advisory, Accenture Healthcare, ADA, AAHA, Service Titan, Clio and Invoca; those attributions could not be verified against any published document and have been removed rather than re-sourced. In the Invoca case the two cited URLs return 404 and have never been captured by the Internet Archive, so the figure was dropped without a substitute: where the measured half of a question has no source, this report says so instead of supplying a number. No figure here is our own measurement, and this report claims none.