After-Hours Call Coverage: What to Do With the 128 Hours You're Closed

First, the arithmetic of being closed

A week has 168 hours. A business open nine to five, Monday to Friday, staffs 40 of them. That is 23.8 percent of the week covered and 76.2 percent not — 128 hours in which the phone rings into whatever you left behind.

Generous hours barely dent it. Open seven to seven, Monday to Saturday, and you have staffed 72 hours: still 96 hours closed, still 57 percent of the week. Covering all 128 closed hours with employees at 40 hours each takes 3.2 full-time people, before anyone takes a holiday, which is why almost nobody solves after-hours coverage by hiring for it.

But closed hours are not interchangeable, and treating them as one block is the most common and most expensive framing error in this category. Four distinct periods behave completely differently:

• The shoulder, roughly 5pm to 9pm on weekdays. This is when most after-hours call volume lands for most consumer-facing businesses. Callers are awake, deliberate, and shopping. They will ring the next result on the page without a second thought.
• Overnight, roughly 10pm to 6am. Very low volume for almost everyone, and dominated by genuine emergencies in the trades where emergencies exist. For a dental practice or an accounting firm, this window is mostly spam and wrong numbers.
• Weekend daytime. For some categories this is a second business day in everything but name. For others it is dead. Your own log will tell you which, and it is not predictable from your industry alone.
• Holidays and unplanned closures. Low in volume, disproportionately high in damage, and the period most likely to be configured wrong or not configured at all.

The practical consequence: buying round-the-clock coverage when your problem is Tuesday at 6pm means paying for 128 hours to fix four. Buying nothing because "we don't get overnight calls" means ignoring the shoulder, which is where the money is. The shape of the problem comes before the shape of the solution.

What the after-hours statistics actually say, and what they don't

One number dominates this topic: that 62 percent of calls to small businesses happen outside business hours. It is repeated across most of the first page of search results for every after-hours query, and it is a misreading.

The 62 percent figure comes from a study by 411 Locals, a local-marketing vendor, which monitored 85 businesses across 58 industries for 30 days. What it measured was whether calls were answered: 37.8 percent reached a live person, 37.8 percent went to voicemail, and 24.3 percent got no response at all. The 62 percent is the two failure buckets added together. It is a statement about calls that went unanswered, not about calls that arrived after closing. The data was collected between 2013 and 2015 and published in 2016, and it is routinely re-dated in secondary roundups. Somewhere in the chain of republication, "unanswered" became "outside business hours," and the mutated version now anchors an entire category of content. The missed-call cost guide traces this number and its neighbours in more detail.

The evidence that does hold up says something more useful and less convenient: the after-hours share is a property of your category, and it varies enormously. BrightLocal's analysis of 45,264 local-business listings across 36 industries reported it by category rather than as an average, and found restaurants taking 51 percent of their calls after 5pm and 32 percent at the weekend, locksmiths taking 31 percent at the weekend, and bars taking 36 percent on Fridays and Saturdays. The spread across categories is roughly fivefold. There is no single after-hours percentage that describes small businesses, and any article offering you one has stopped checking.

Two further findings matter more for this decision than any volume statistic, because they describe what happens after you miss the call. Pew Research Center, in a survey of US adults fielded 13–19 July 2020, found that eight in ten Americans say they do not generally answer their cellphone when an unknown number calls; only 19 percent said they generally do. Hiya's State of the Call 2026, based on a survey of more than 12,000 consumers across six countries, puts the share of calls from unknown numbers that go unanswered at 86 percent.

That is the case against voicemail, stated properly. It is not that voicemail is unprofessional. It is that voicemail's recovery path — you ring them back in the morning — routes through the single least reliable event in modern telephony. At 9am you are an unknown number.

Measure your own closed-hours volume in one week

This takes about an hour, spread over a week, and it replaces every borrowed percentage above with a number that is actually about your business.

• Export a full recent week of call records with timestamps. Almost every VoIP system and business mobile line supports this. Avoid a holiday week and avoid your seasonal peak.
• Sort every call into the four periods: staffed hours, the weekday shoulder, overnight, and weekend. Do this by timestamp, not from memory. Owners are reliably wrong about this, usually in the direction of overestimating overnight.
• Within the closed-hours calls, classify each one as new business, existing customer, supplier or internal, or spam and wrong numbers. This step is the one most often skipped and the one that changes the answer most. A pile of 40 after-hours calls that turns out to be 11 new enquiries is a different business case entirely.
• Check for urgency. How many of those closed-hours new-business calls were time-critical — a no-heat call in January, a flood, a dental injury — versus a routine booking request that could have waited until morning? Urgent and routine calls need different handling, and mixing them is what produces both missed emergencies and pointless 2am phone calls to the owner.
• Call your own number as a customer would. At 8pm on a Tuesday, at 10am on a Saturday, and on the next public holiday. Count the rings before anything answers, listen to the whole greeting, and check whether the mailbox is even accepting messages. Businesses discover full mailboxes and greetings naming staff who left two years ago almost every time they run this.
• Check the recovery you already have. Of last month's after-hours voicemails, how many were returned, how fast, and how many of those callbacks were actually answered? That final number is the callback problem measured in your own data rather than in a survey.

Three things corrupt this measurement. Forwarding the main line to a personal mobile hides misses from the business line's log entirely. Separate tracking numbers for ad campaigns split the record across systems. And some phone systems count a call answered by voicemail as answered, which quietly zeroes out the exact thing you are trying to count — check the report's definition before trusting it.

The five options, and what each one actually does

Voicemail. Free, already installed, and not literally nothing: a caller who leaves a message has told you they want to be reached. The cost sits in the recovery path, for the reason above — your callback arrives as an unknown number, hours later, competing with spam for two seconds of attention, and possibly carrying a carrier spam label you don't know about. Voicemail is a defensible choice when closed-hours new-business volume is genuinely tiny and nothing is ever urgent. It is a poor choice everywhere else, and it is the default almost everywhere.

Forwarding to a person on call. The cheapest coverage that involves an actual human, and the one most trades already run informally. Its costs are real but rarely counted: the on-call person's evenings, the inconsistency between how each person handles a call, the absence of any record if the call goes to a personal phone, and the slow attrition of the arrangement as whoever holds the phone starts letting unknown numbers ring out. It works well as a second rung for genuine emergencies. It works badly as the primary answer to routine after-hours volume.

A human answering service. Live operators, usually a shared pool working from a script. Pricing is typically per minute rather than flat, with market rates in September 2026 running roughly $0.75 to $2.00 a minute and mid-range providers around $0.90 to $1.25. After-hours calls commonly carry a premium of 25 to 50 percent over daytime rates, federal holidays are commonly billed at 1.5 to 2 times the standard rate, and overage minutes beyond your bundle often run 30 to 75 percent above your in-package rate. After-hours-only plans at small-business volumes commonly land somewhere around $135 to $400 a month. The structural limitation is not price, it is scope: most answering services take a message and relay it. Very few can write into your actual scheduling system, which means the booking still happens tomorrow, by you.

An AI receptionist. Answers on the first ring, at any hour, on as many simultaneous calls as arrive, usually on a flat monthly subscription rather than per minute — which matters most in exactly the hours a human service charges a premium for. Where it is connected to a real calendar it can complete the booking rather than record a request for one. Its limitation is the mirror image of its strength: it does only what its rules permit, so the rules are the product. An AI receptionist deployed without a considered permission model and escalation ladder is worse than voicemail, because it sounds authoritative while being wrong. The comparison guide covers the three staffed options side by side, and the cost guide covers what the monthly bill actually contains.

Missed-call text-back. Strictly speaking not coverage at all — it is recovery, triggered after a call has already been missed. It is the cheapest partial fix available, it lands in a channel people actually read, and it carries your business name rather than an unfamiliar number. It cannot answer a question the caller asks out loud, and it is subject to a timing constraint covered below that most write-ups of this topic ignore entirely. The missed-call text-back guide covers what the sequence has to do to be worth sending.

These are not mutually exclusive, and the businesses that solve this properly usually run two. A common working combination is an AI receptionist for the shoulder and weekends, a named on-call person as the escalation rung for genuine emergencies, and text-back as the catch for anything that still slips through.

The decision nobody writes down: what coverage is allowed to do

Whatever answers at 8pm needs a written permission model. Almost nobody writes one, which is why after-hours coverage so often either under-delivers or creates a mess that costs more to clean up than the calls were worth. There are three tiers, and they carry very different risk.

Tier one, inform and record. Confirm hours, location, parking, whether you take a given insurance, what to bring. Take a message with a callback number and a stated reason. Low risk, low value, and the ceiling on what a message-taking service can offer.

Tier two, transact. Book, reschedule, or cancel against the same live calendar your staff use. Quote from a fixed published price list, never from judgment. Collect the details that let the job actually start in the morning. This is where nearly all of the value in after-hours coverage sits, because it converts an enquiry while intent is live instead of adding it to a callback queue.

Tier three, wake somebody up. Reserved for genuine emergencies, defined below. High value when correct and high cost when wrong, in both directions: a missed emergency damages a customer relationship and sometimes a building, and a false alarm at 3am erodes the on-call rota faster than anything else.

For tier two, four things have to be decided explicitly and in advance. Which service types are bookable without a human — a routine cleaning, yes; a complex multi-visit treatment plan, probably not. How far ahead the calendar may be opened at night, which is usually narrower than your daytime booking window. Whether an after-hours booking requires confirmation in the morning before it is treated as firm. And what must never be quoted, priced, or promised outside staffed hours, which for most businesses includes anything involving insurance coverage, warranty status, or a price that depends on seeing the job.

One failure mode is worth naming because it is common and entirely avoidable: giving after-hours coverage permission to book, but pointing it at a calendar that is not the one your staff work from. Monday morning then arrives with double bookings, and everyone concludes that automation doesn't work when what actually failed was an integration decision. The buyer's checklist has the specific questions that separate real two-way calendar access from a scheduling link in a text message.

Emergency triage and the on-call ladder

If your business has emergencies, after-hours coverage is a triage system before it is a booking system, and it needs to be designed as one.

Start by writing down what urgent means in your own words, in terms a person with no training could apply from what the caller says. For plumbing: active uncontrolled water, sewage backup, no water at all. For HVAC: no heat below a stated outside temperature, no cooling with a vulnerable occupant, any gas smell — which is not a plumbing or heating call at all but a utility emergency, and should be routed as one. For dentistry: facial swelling, uncontrolled bleeding, trauma to a permanent tooth. For property management: flood, fire, no heat, lockout, anything affecting habitability. Vague criteria produce either a woken owner every night or a missed emergency, and usually both in the same month.

Then build the ladder, and write down all four parts of it. Who is rung first, and on what number. How long before the call escalates if that rung does not answer — sixty to ninety seconds is typical, and it must be measured in rings, not in hope. Who is rung second. And, the part almost always left undefined, what happens when nobody on the ladder answers at all.

That last one is the real failure mode. The default behaviour of most systems is to leave the caller with a promise of a callback that no named person is on the hook for. Decide instead what the caller is actually told: a specific alternative (an emergency line, an urgent care route, a partner firm), or an honest statement that the message has been logged and someone will call at a stated time. An honest "we will call you at 7am" is better than a callback promise that quietly evaporates, and far better than an automated system improvising reassurance.

Two safety rules apply regardless of industry. Anything that is a medical emergency, a fire, or a gas leak goes to emergency services, and an automated first line should say so immediately and plainly rather than attempting to assess it. And no automated system should be the last link in a chain that ends with a person in danger — an escalation that is not answered must fail loudly to a human, not silently to a message log. Where health information is involved, the HIPAA guide covers what changes about handling and storing those calls.

Finally, review it. Log every after-hours escalation for the first month and read the log weekly: how many fired, how many were genuine, how many were answered on the first rung, and how long each took. Triage rules written once and never revisited drift out of alignment with reality within a season, and the rota is usually the first thing to quietly break.

The texting trap: you can't legally text everyone back at 2am

The standard advice for after-hours calls is to text the caller back immediately. It is good advice, and it comes with a constraint that almost no article on this topic mentions.

The TCPA's call-time rules, at 47 C.F.R. § 64.1200(c)(1), prohibit telephone solicitations before 8 a.m. or after 9 p.m. in the recipient's local time, and text messages are treated as calls for this purpose. Read strictly, the restriction covers solicitations, which is why an immediate, genuinely responsive reply to a call the person placed themselves at 11pm is a different thing from an 11pm promotional message. That distinction has not stopped a substantial volume of litigation arguing otherwise, and the industry petition asking the FCC to rule that prior consent disapplies the quiet-hours provision, filed in March 2025, remains unresolved. Several states impose narrower windows than the federal rule, including Florida's 8 a.m. to 8 p.m.

The configuration that follows is simple enough to state in one line: the immediate responsive text follows the caller, and everything scheduled afterwards — the nudge, the reminder, the re-engagement message — waits for the permitted window in the recipient's local time, defaulting to the narrowest window when the time zone cannot be determined. A system that fires a "just checking in" follow-up at 7am because that is when the job queue runs is the exact thing the rule is about. The TCPA guide covers consent, carrier registration, and the state-by-state variation in full.

None of this is a reason to avoid texting after hours. It is a reason to configure the immediate reply and the scheduled follow-ups as two separate things with two separate rules, which is a five-minute decision at setup and an expensive one to retrofit.

Holidays, closures, and the edge cases that quietly break coverage

After-hours coverage usually fails first at the edges rather than in the middle, and the edges are predictable enough to handle in advance.

Holidays have to be loaded, not assumed. Whatever answers your phone needs an actual calendar of the days you are closed, including the ones that are not federal holidays — the week between Christmas and New Year, regional holidays, the day after a long weekend that your team takes and the calendar doesn't know about. A system that treats a holiday Monday as a normal Monday will cheerfully book appointments into an empty building.

Unplanned closures need a switch and a named person who can throw it. Snow, a burst pipe in your own premises, an outbreak of flu through a small team: the question is not whether these happen but whether the person who finds out first at 6am can change what callers hear from their phone, without logging into a desktop tool nobody has the password for.

Failover needs testing. If your internet or VoIP provider goes down, where do calls go? Most systems have an answer to this configured by default, and most businesses have never tested what it actually is. Find out on a quiet afternoon rather than during an outage.

Greetings drift. The single most common after-hours defect found on an audit is a seasonal greeting still playing months later — a holiday message in February, a closure notice from a renovation that finished last year. Put a recurring reminder on whoever owns it.

Published hours have to match everywhere. Your Google Business Profile, including its special-hours entries for holidays, your website, and your voicemail greeting should say the same thing. A customer told you are open who finds you closed is a worse outcome than one who knew; and inconsistent hours across your listings are a local-search problem as well as a customer-service one.

Multi-location businesses need a routing rule for the hours when only some locations are closed, and a decision about whether a caller to a closed location is offered the open one. Both answers are defensible. Not deciding is not.

The break-even, and when the answer is to buy nothing

The arithmetic is the same shape as the general missed-call formula, restricted to closed hours, and it takes five inputs you now have from the measurement exercise.

• A: closed-hours new-business calls per month, counted from your log rather than estimated.
• L: the share of those callers who do not come back to you. This is the genuine assumption in the model. AutomateLine's own missed-call calculator states a band of 30 to 55 percent in its small print rather than presenting a single number as research, and that band is a reasonable starting point.
• C: your conversion rate from a reached enquiry to a booked job, which your CRM knows.
• V: the average value of a booked job, or of a customer's first year if the work repeats.
• R: the recovery rate — the share of those calls the coverage actually saves. Nothing recovers all of them, and any vendor implying otherwise has told you something about the vendor.

Monthly value = A × L × C × V × R. Weigh that against the all-in monthly cost of the option, including after-hours premiums, holiday rates, setup, and per-number fees.

Worked as an illustration with made-up inputs, clearly labelled as such: a business logging 30 closed-hours new-business calls a month, at a 40 percent non-return rate, a 45 percent conversion rate, and a $450 average job, has roughly $2,430 a month at risk. At a 50 percent recovery rate the coverage is worth about $1,215 a month. Against a flat subscription in the low hundreds that is a clear decision. Against a per-minute service at, say, 30 calls averaging four minutes at $1.10 plus a 35 percent after-hours premium — about $178 a month before holidays and overage — it is also a clear decision, but a different one, and the gap between them narrows fast as volume rises. Run it at both ends of the non-return band. If the decision only works at the optimistic end, it is a thin decision and worth saying so out loud before signing anything.

And the case vendors rarely volunteer: sometimes the answer is to buy nothing. If your closed-hours new-business volume is a handful of calls a month, none of them urgent, and your customers reliably ring back in the morning, the honest fixes are free. Extend your published hours to match when you actually answer the phone. Re-record a greeting that says when you will call back and then keep that promise. Turn on a text-back so the caller gets something in the thirty seconds after they hang up. Then re-measure in a quarter. Buying a 24/7 system to cover four genuine enquiries a month is a worse decision than voicemail, and the common mistakes guide covers what tends to happen when automation is bought before the problem is understood.

What to change on Monday

• Export last week's call log and sort the closed-hours calls into new business, existing customers, suppliers, and spam. One hour, and it decides everything else.
• Call your own number at 8pm and on a Saturday morning. Listen to the entire greeting and check the mailbox is not full.
• Write down, in one page, what after-hours coverage is allowed to do: what it can book, what it must never quote, and what counts as urgent in your business.
• Write the escalation ladder, including who is rung second and what the caller is told when nobody answers at all.
• Load the holiday calendar for the next twelve months, and name the person who can flip the closure switch from their phone.
• Split the immediate text reply from every scheduled follow-up, and hold the scheduled ones to the permitted window in the recipient's local time.

Sources

• Hours arithmetic in this guide is simple division: 168 hours in a week, 40 staffed for a nine-to-five weekday schedule, leaving 128 hours (76.2 percent) unstaffed. No source needed and none implied.
• The 62 percent figure: 411 Locals, study of 85 businesses across 58 industries monitored over 30 days, data collected 2013–2015 and published 2016 — 37.8 percent of calls answered live, 37.8 percent to voicemail, 24.3 percent no response. It measures unanswered calls, not after-hours calls. The publisher is a local-marketing vendor, and the study is frequently re-dated in secondary roundups.
• Call timing by category: BrightLocal, Google My Business Insights Study, analysis of 45,264 local-business listings across 36 industries — restaurants 51 percent of calls after 5pm and 32 percent at the weekend, locksmiths 31 percent at the weekend, bars 36 percent on Fridays and Saturdays.
• Answering behaviour: Pew Research Center, "Most Americans don't answer cellphone calls from unknown numbers," published 14 December 2020, from a survey of US adults fielded 13–19 July 2020 — eight in ten say they do not generally answer unknown numbers, 19 percent say they do.
• Unknown-number answer rate: Hiya, State of the Call 2026, based on a survey of more than 12,000 consumers across the US, Canada, the UK, Spain, France, and Germany — 86 percent of calls from unknown numbers go unanswered.
• Answering-service pricing: market ranges observed across published 2026 vendor pricing guides as of September 2026 — roughly $0.75 to $2.00 per minute with mid-range providers around $0.90 to $1.25, after-hours premiums commonly 25 to 50 percent, federal-holiday rates commonly 1.5 to 2 times standard, overage minutes commonly 30 to 75 percent above the in-package rate, and after-hours-only plans commonly $135 to $400 a month at small-business volumes. These are market ranges, not quotes, and they move; get your own written quote including premiums and overage before comparing anything.
• Texting hours: Telephone Consumer Protection Act, 47 U.S.C. § 227, with FCC rules at 47 C.F.R. § 64.1200(c)(1) setting the 8 a.m. to 9 p.m. local-time restriction on telephone solicitations. The Ecommerce Innovation Alliance petition asking the FCC to disapply quiet hours where prior consent exists was filed in March 2025 and remains unresolved. Florida's Telephone Solicitation Act sets a narrower 8 a.m. to 8 p.m. window. None of this is legal advice.
• The 30 to 55 percent non-return band and the $35 owner-hour figure are AutomateLine's own stated calculator assumptions, published in the calculator's small print, not research findings.
• Worked examples in this guide use illustrative inputs, labelled as such. They are not AutomateLine client results.