One of four things, and three of them leave you with nothing to work with in the morning. The line rings out, it goes to voicemail, it forwards to a cell phone in someone’s pocket, or something answers and writes down who called and why. Which one happens at your number tonight is a setting, and most owners inherited it rather than chose it.
This post uses federal time-use and wage data, Google’s own documentation, and prices printed on vendors' pricing pages with the date they were read. Where we do arithmetic, we say so, and the inputs are published so you can redo it.
What happens to a call after hours at a service business?
It follows one of four paths. Three of them end with the caller gone and no record that they existed; the fourth ends with a name, a number, a reason, and a next step.
Path one: the line rings until the caller hangs up. Path two: voicemail answers, and whether the caller leaves one is their decision, not yours. Path three: the call forwards to a personal cell, which is the least predictable of the four, because the outcome depends on whether the owner is at dinner, asleep, or on a roof. Path four: something answers, identifies the business, captures the caller’s details and reason, separates an emergency from a quote request, and tells the caller what happens next.
The honest limit of the available data is worth stating up front. Invoca’s 2026 benchmark report, built on anonymized data from over 70 million calls and 600 million minutes of conversation across 10 industries, found that 56% of calls to businesses are answered by a person. Filtering out misdials and quick hangups, the rate rises to 65% for calls lasting longer than 15 seconds and 71% for calls over 30 seconds. Those are all-hours figures across that customer base, not an after-hours measurement, and nobody should present them as one. They establish that unanswered calls are ordinary, not that yours cluster at night.
How many hours a week is your business actually closed?
If you are open 8 to 5 Monday through Friday, you are open 45 of the 168 hours in a week. That leaves 123 hours, about 73% of the week, on the other side of your business hours. That is calendar arithmetic, ours, not a claim about your demand.
The number matters because of what those 123 hours contain: every weekday evening, every weekday morning before eight, every overnight, and both full weekend days. A plan that covers "evenings" covers maybe 20 of them. Worth saying plainly: a big window is not the same as a busy window. The only thing that tells you whether anybody is calling in those 123 hours is your own call log, with hour stamps, for a full week.
Calendar arithmetic. Ours, not a measurement.
View data
| Item | Value |
|---|---|
| Hours in a week | 168h |
| Open | 45h |
| Outside business hours | 123h |
When is your customer free to call you?
Mostly when you are closed, because they are at work during the hours you are open. The Bureau of Labor Statistics' American Time Use Survey found that 81% of employed people worked on an average weekday in 2025, compared with 30% on an average weekend day.
The same survey found that on days they worked, full-time employed people worked an average of 8.1 hours, rising to 8.5 hours on an average weekday and falling to 5.5 hours on an average weekend day. Those are 2025 annual averages from roughly 6,100 individuals interviewed, each reporting a single diary day by telephone.
Two caveats belong with that number, both from BLS itself. First, remote work moves some of this: 35% of employed people did some or all of their work at home on days they worked, against 70% who did some or all of it at their workplace, so a minority now has daytime flexibility it did not have a decade ago. Second, ATUS operations were suspended from October 1 through November 12, 2025 during the federal shutdown, and BLS states plainly that it is not possible to quantify the effect of the missing diaries on the 2025 estimates. The weekday-versus-weekend gap is large and long-standing, but it is not a precision instrument this year.
What this supports is narrow and useful: the hours when your customer can comfortably make a personal phone call overlap heavily with the hours you have nobody on the line. It does not tell you how many of them do.
Why does your call report make the after-hours window look empty?
Frequently because the measurement is switched off in that window, by design and on good advice. If your calls come from search ads, you may have told Google to stop showing your phone number when you are closed.
Google’s documentation for call assets states that under business hours you can set numbers to appear only when your business can accept calls, and that under advanced options you can schedule a call asset to show only on certain days and hours, for example only when your business is answering phone calls. Invoca’s 2026 report recommends the same discipline, listing "pause ads when you’re closed" among its tactics and noting there is no reason to pay for calls that cannot be answered unless an agent is handling them.
Both recommendations are sensible on their own terms. Together they produce a reporting artifact: the window looks quiet because you stopped buying visibility in it and stopped placing a number there to call. That is a fine way to control spend and a terrible way to size the opportunity. If you want to know what the window holds, look at the organic line, the one on your Google Business Profile and your truck, over a week, before changing anything.
What does covering the after-hours window cost?
On published pricing pages read October 1, 2026, round-the-clock availability is not the expensive part. It is included at both ends of the market, and what the plan actually meters is volume.
Ruby’s plans and pricing page lists 24/7 live answering, including after-hours, weekends and holidays, as a feature of every plan, then prices the plans by receptionist minutes: $250 a month for 50 minutes, $395 for 100, $720 for 200, and $1,725 for 500. Rosie lists 24/7 availability on all plans and prices them at $49 a month for 250 included minutes, $149 for 1,000, and $299 for 2,000. Smith.ai lists 24/7 AI call answering among its core features, including on a free tier covering 25 calls a month at $3.00 per call beyond that, with paid tiers at $150 a month for 75 calls and $500 for 300.
Read those two facts together, because the implication is easy to miss. "24/7" describes when the service is awake. It does not add minutes. The 50-minute plan that used to cover 45 open hours a week now covers 168, out of the same 50 minutes. If the after-hours calls are real, the bucket is where you will feel them, not the coverage line. We went through the full pricing landscape and what each plan counts in what an AI receptionist costs per month.
What would staffing those hours yourself cost?
More than any plan above, for a slice of the window. The Bureau of Labor Statistics puts the median hourly wage for receptionists at $18.27 in May 2025, with the lowest 10% under $13.83 and the highest 10% above $24.01, across an occupation of 947,500 jobs.
Now the arithmetic, which is ours. Extending coverage from 5pm to 9pm on weekdays adds 4 hours a day, 20 hours a week, roughly 87 hours a month. At the median wage that is about $1,580 a month in wages alone, before payroll taxes, benefits, or the cost of recruiting somebody who wants an evening shift. And it buys 20 of the 123 hours. The remaining 103, including both weekend days, stay exactly as they are tonight.
Our arithmetic on published inputs: BLS median receptionist wage (May 2025) and vendor list prices read October 1, 2026. Not equivalent products.
View data
| Item | Value |
|---|---|
| Evening shift, 87 hours, wages only | $1580 |
| Live answering, 200 included minutes | $720 |
| AI answering, 1,000 included minutes | $149 |
Those three bars are not the same product and should not be read as a ranking. The evening shift buys a person who can hear that something is wrong and decide what to do about it, for four hours a day, five days a week. The metered plans buy availability across all 168 hours and judgment only inside whatever you wrote down in advance. Which one fits depends on how many of your after-hours calls are routine, and you cannot know that until you have listened to a few.
What should actually happen to an after-hours call?
Four things, in this order: the caller gets a response inside the first minute, their name and number and reason get captured, an emergency gets separated from everything else and escalated, and the rest get a specific next step with a time attached.
The first one is the only one that has to happen live. Everything after it can wait until morning, provided it was written down. In practice that means:
- Something answers and identifies the business by name. A caller who hears a generic greeting at 9pm often assumes they dialed wrong.
- The reason gets captured in the caller’s words, not a category. "Water coming through the kitchen ceiling" routes differently than "plumbing."
- One triage question, asked early. Is this an emergency or can it wait until business hours? The answer decides whether a phone rings in somebody’s bedroom.
- A next step with a time on it. Not "someone will be in touch." Closer to "a technician will call you between 7 and 8 tomorrow morning." A time you can keep beats a time that sounds good.
- A record that exists in the morning. Name, number, reason, time of call, and what was promised. If it only lives in a voicemail nobody plays, the call did not happen.
Notice that none of this requires a decision about vendors. It is a specification for the window, and it is worth writing down before you shop, because it is the thing you will be comparing plans against.
What should you do first?
Call your own main number tonight at 8pm, and again on Saturday at 10am, and write down exactly what happens. Then pull one week of call logs with hour stamps.
That is a ten-minute exercise and it answers the two questions this whole post circles. The first call tells you which of the four paths your line is on right now, which is more than most owners can say with confidence. The log tells you whether the 123-hour window has anything in it worth covering, which no benchmark and no vendor can tell you from outside. If the log is empty, you have saved yourself a monthly bill. If it is not, you now have the number the decision actually depends on, and how fast the follow-up needs to be in the morning is its own question with its own data.
If you want the rest of the path a caller takes before they ever dial, from what Google shows to what your site answers, a site check covers that layer in a few minutes.
Questions owners ask
What happens to a call after hours at a service business?
It takes one of four paths, and the path is a setting somebody chose, usually years ago. The line rings until the caller gives up; it goes to voicemail; it forwards to a personal cell that may or may not be answered; or something answers, captures the caller’s name, number, and reason, and tells them what happens next. Only the fourth leaves you with a record you can act on in the morning. Dialing your own number at 8pm tells you which one you have in about thirty seconds.
How many calls come in after hours?
Nobody can tell you from outside your own logs, and be careful with anyone who claims to. What is measurable is the size of the window: a business open 8 to 5 Monday through Friday is open 45 of the 168 hours in a week, which leaves 123 hours, about 73% of the week, outside business hours. That is calendar arithmetic, not a measurement of your demand. Pull a week of raw call logs, with the hour stamp, before you buy anything.
Is it worth paying for after-hours coverage?
That depends on a number only your call log has, which is why the first step is reading it rather than pricing plans. What published pricing pages show, read October 1, 2026, is that 24/7 coverage is not the expensive part: Ruby includes 24/7 live answering on all plans, from $250 a month for 50 receptionist minutes, and Rosie includes 24/7 answering on all plans, from $49 a month for 250 minutes. What you pay for is volume, not hours.
Why does my call reporting show almost nothing after hours?
Often because the measurement is switched off in that window on purpose. Google’s documentation for call assets states you can set numbers to appear only when your business can accept calls, and schedule the asset to show only on certain days and hours. Invoca’s 2026 benchmark report recommends the same thing, advising businesses to pause ads when closed because there is no reason to pay for calls that cannot be answered. Both are sound advice, and both make after-hours demand invisible in your own dashboard.
Should an after-hours system handle emergencies?
It should separate them, not solve them. The useful job at 11pm is triage: decide in one or two questions whether this is a burst pipe or a request for a quote on a kitchen, then escalate the first to a human immediately and give the second a specific next step with a time on it. A system that treats both the same way either wakes you up for nothing or lets a real emergency sit in a transcript until morning.
Sources
- Invoca, Lead Conversion Benchmarks Report 2026 (anonymized data from over 70 million calls and 600 million minutes of conversation, 10 industries and 7 marketing channels; figures are averages across the Invoca customer base) (2026)
- U.S. Bureau of Labor Statistics, American Time Use Survey Summary, 2025 annual averages (approximately 6,100 individuals interviewed) (2026)
- U.S. Bureau of Labor Statistics, American Time Use Survey Technical Note (sample and method) (2026)
- U.S. Bureau of Labor Statistics, Occupational Outlook Handbook: Receptionists (median pay, May 2025 wage data; 947,500 jobs in 2025) (2026)
- Google Ads Help, About call assets (business hours and scheduling settings) (2026)
- Ruby, plans and pricing page (50, 100, 200 and 500 minute plans; 24/7 live answering listed in all plans; read October 1, 2026) (2026)
- Rosie, pricing page (250, 1,000 and 2,000 minute plans; 24/7 availability listed in all plans; read October 1, 2026) (2026)
- Smith.ai, AI receptionist pricing page (free, Pro and Enterprise tiers with published per-call rates; read October 1, 2026) (2026)
The first meeting starts with findings, not promises.The free Business Health Report reads your business and shows you what it found, in days.
Get Your Free Business Health Report