After-hours calls and what they are worth

After hours is the least examined part of most phone operations. The calls arrive when nobody is there, and the only trace is a voicemail count nobody reads as a lost revenue report. It is also the easiest thing in this category to fix, because the bar you are clearing is silence.

How many calls are actually coming in after hours?

More than owners guess, and less than the vendors selling night coverage claim.

Start with what is actually measured, because most of what circulates is not. Slang AI, working from millions of full-service restaurant calls in its August 2025 phone report, found that only 66% of restaurant calls occur during business hours, that 26% are either simultaneous or after hours, and that roughly 71% of all calls are directly tied to revenue. Maple, reporting on 1.2 million calls across more than 1,000 US locations, puts it more plainly: the average paying restaurant on its platform receives roughly one in five of its calls outside the hours it is open to take them.

Health care looks similar from a different angle. Patient Prism scored 11,552,668 dental calls across 8,280 locations in calendar 2025 and attributed 9.3% of all lost opportunities to connection failures, meaning voicemail, disconnects and hold hangups, explicitly citing staffing, routing and after-hours coverage as the causes.

All three are vendors reporting on their own platforms, and all three disclose their samples, which is what separates them from most AI vendor blogs.

The defensible planning number, and label it as an assumption rather than a finding, is that 20 to 35% of inbound calls to a consumer-facing local business arrive outside staffed hours. Restaurants sit at the higher end because evening and weekend demand is the business. Emergency trades sit higher still.

Then stop using the assumption and get your own number. Ask your carrier for thirty days of inbound call detail records and bucket them by hour. You are looking for four windows: before opening, after closing, weekends, and holidays.

A total is useless. The shape tells you what to buy. Heavy 6pm to 8pm volume calls for an extended evening, not a twenty-four hour system, and those are different prices.

Are after hours callers worth more than daytime callers?

Sometimes, and the honest version is narrower than the pitch.

Nobody has published a study comparing after-hours call value to daytime call value in a local business. The claim gets made constantly and there is no dataset underneath it.

The competitive argument needs no statistic. At eight in the evening most of your competitors are not answering either. During business hours you compete against every open business in your market. After hours you compete against whoever picks up.

The demand evidence is real for home services. Housecall Pro's 2026 Home Services Report, combining a homeowner survey with Q1 2026 platform data, found that a sudden problem is what most often makes a homeowner pick up the phone at 58%, ahead of a smaller issue worsening at 51% and routine maintenance at 47%. In the same report, 72% said they would pay more to resolve an emergency within 24 hours, and repairs accounted for roughly 87% of plumbing jobs and 88% of HVAC jobs. Unplanned failures drive the call, speed carries a premium, and failures do not respect your posted hours.

What the evidence does not support is a blanket claim that a night caller is worth more than a day caller in every category. A restaurant taking reservation calls at 9pm is not capturing higher-value covers, it is capturing covers it would otherwise lose. It is a volume argument, not a value argument, and the two produce different budgets.

What does the arithmetic actually look like?

What should happen to a call at nine at night?

Something specific, and it should be decided in advance rather than left to a default.

Book, if you can. An appointment written into the calendar while the caller is on the phone is the outcome worth paying for, because it removes the morning callback, and the morning callback deserves more suspicion than it gets. A TNS survey from July 2022 found that 75% of Americans never answer calls to their wireless phone from a number they do not recognize. Your callback is an unrecognized number. Capturing a name and ringing back is materially weaker than it was when the folklore around it was written, which is the argument for resolving the call while you still have the person.

The Patient Prism funnel shows what happens when you do not. Of every 100 dental calls, 21 book and 13 walk away. Only 4.8 of those walkaways receive a followup call, and only 0.7 come back and book. Fewer than one caller in a hundred who walked away ever walked back through the door. That is measured behaviour across eleven and a half million calls, not a survey of what people say they would do.

Capture, if you cannot book. Name, number, what they need, and a specific callback promise. "Someone will call you before nine tomorrow morning" is a commitment a caller can plan around. "As soon as possible" is heard as nothing.

Escalate, if it is urgent. Define urgency in advance with a word list and route those calls to a real cell phone. An agent that qualifies an emergency caller through five intake questions is doing damage.

Nine at night is also when a system is most likely to overstep, because the caller is pushing and there is no human in the room. The list of things an AI phone agent should never say does not relax after closing time.

Why is after hours the right place to start?

Because it is the lowest risk deployment available and the learning is real.

The alternative is voicemail, so the comparison is generous. Volume is low, which means you can read every transcript for a week without it becoming a job. And your daytime operation is untouched, so a bad week costs you calls that were already being lost.

Start there for two weeks. Read everything. Fix the script. Then extend into the hours where you know you miss calls during the day.

The pattern also tells you whether you have a bigger problem, and this is where the advice most often breaks down. Maple's data shows 68% of restaurant calls arrive during the lunch and dinner rushes, 28% at lunch and 40% at dinner. The phone peaks exactly when the floor is least able to answer it. If that is your shape, night coverage is solving the wrong hour. The fix is capacity between 5pm and 9pm, and that is a different purchase.

A rise in after hours volume is also worth reading as a signal. It sometimes means your business is losing customers on the phone during the day and people are calling back when they get home.

Is there a cheaper option than a full voice agent?

Yes, and for a lot of businesses it is the right first move.

An automatic text back on a missed call costs a fraction of a voice agent. It does not book anything and it does not answer questions, and it puts a live thread in front of the caller within seconds, which is the part that matters most at night.

It also carries obligations that voice does not. Under the FCC's revocation rules, in force since 11 April 2025, a consumer may revoke consent by any reasonable method, and the caller must treat any wording a reasonable person would understand as a stop request as valid. Not just the word STOP. Requests must be honored within ten business days at the outside. Statutory damages run at $500 per message, trebled to $1,500 for willful violations. Confirm your own program with counsel, not with a vendor.

The engineering consequence most deployments miss: a spoken opt-out on a voice call has to write into the same suppression list your texting platform reads. One table, keyed on the phone number, shared across voice and SMS. Ask your vendor whether that is wired before you switch anything on, and get the answer in writing.

Run the text back for a month and count replies. High reply volume proves the demand exists. Low reply volume saves you a subscription.

The other cheap move is fixing what people find before they dial. If evening callers are asking questions your site should answer, that is a content problem. Ranking for a service you never wrote a page about is a common gap, and publishing pricing on your site removes one of the most common reasons people call at all.

What to do this week

Request thirty days of inbound call detail records and bucket them by hour and day.

Total the after hours calls, strip the noise, and run the five step arithmetic above against your own close rate and average ticket. Halve the result before you compare it to a price.

Decide the three outcomes for a night call: book, capture, or escalate, with the exact wording for each. Write your urgency word list and the cell number those calls go to.

Then call your own line at nine at night and listen to what a customer currently gets. Most owners have never done this and are surprised by it.

Questions about what your nights are worth?

Email me at eric@seod.com with your open hours, your average job value, and your rough close rate on inbound calls. I will run the five steps above on your numbers and send back the break even figure for a month of coverage.

It takes me about ten minutes. If the number says do nothing, I will tell you to do nothing.

More on the surrounding decisions sits in the AI phone and lead response library.