What is the real cost of sending after-hours calls to voicemail?

It's 6:40 on a Thursday and your phone lights up in the truck while you're loading a ladder. You let it ring. The greeting picks up, tells the caller you're open at eight, and gives them the beep. You just made a decision about that job, and it probably wasn't the one you thought you were making.

Plenty of shops run this way after five, and for a good reason: nobody can sit at a desk answering a phone at ten at night. Restoration and the emergency-plumbing crowd are the exception, on call by definition. The part worth checking isn't the decision to stop answering. It's the belief sitting underneath it, which is that voicemail holds the call until morning.

It doesn't. Voicemail doesn't hold an after-hours call, it ends it. Invoca's platform data puts the share of voicemail-bound callers who leave a message at fewer than 3%, so the overwhelming majority of your after-hours calls leave you a bare number in a log and nothing else: no name, no job, no idea whether it's still worth calling. The real cost isn't the messages you return in the morning. It's the calls you can't tell apart, booked that same evening by whoever picked up.

Where did the "voicemail will catch it" habit come from?

It came from a time when it was true. For decades, leaving a message was simply how you reached a business that was closed. Everybody did it, nobody thought twice about it, and an answering machine was a genuine upgrade on a phone ringing into an empty office.

What changed isn't the voicemail. It's the caller. The person holding the phone at 6:40 has a competitor's number one tap away, a map with six more, and no particular reason to wait on you. The habit stood still while the alternative to it got very good.

If the call is really urgent, won't they leave a message?

This is the load-bearing belief, and the evidence doesn't support it. The share of voicemail-bound callers who leave a message is under 3% across the board, and that figure is an average over everybody, urgent or not. Nobody has measured it by urgency. But it's hard to argue that the person with a real problem is the patient one, waiting through a greeting to narrate it to a machine.

Picture what that person is actually doing. Half the house has no power and they can smell something hot, or there's water coming through a light fixture. If there's any real question about electricity or gas, the utility and the fire department are the right call, not a contractor.

Most of those calls aren't that, though. What the caller wants is somebody to pick up, take it seriously, and get them on the schedule. What they get is a recording, an opening time, and a tone. The next thing they do is hit back and tap the next result.

The ones who do leave a message aren't much of a consolation either. Most voicemail-bound callers never get a call back, and a message sitting in a box overnight is competing against the company that answered at 6:41. We went through what actually happens to those callers in more detail.

Isn't calling back first thing in the morning fast enough?

Fast enough compared to what? A call at seven in the evening, returned when you open at eight, has been sitting for thirteen hours. In the customer's head that isn't a callback. That's the next day.

The best evidence on response speed is old, and it measures web-form leads rather than phone calls, so take it as a direction rather than a law. A study reported in the Harvard Business Review piece on online sales leads, covering 1.25 million leads at 29 B2C and 13 B2B U.S. companies, found firms that made contact within an hour were nearly seven times more likely to qualify the lead than firms that waited even an hour longer, and more than sixty times more likely than those that waited a day or more.

Worth being precise about what that measures: qualifying a lead, meaning getting a real conversation with the person who decides, not closing a sale. Which is exactly what an after-hours call is. Somebody wanted a conversation, and thirteen hours later they've had one with somebody else. We went deeper on how fast a contractor should call a new lead back.

HBR's own audit in that same article planted a test lead with 2,241 companies to see how long each took to answer. Harvard Business Review reported an average of 42 hours among the companies that answered within 30 days, and 23% that never answered at all. Nobody sets out to be in that last group. You get there one "I'll deal with it in the morning" at a time.

Aren't after-hours calls mostly junk anyway?

There's something real underneath this one. Unwanted calls are a genuine nuisance at national scale: the FTC's Do Not Call Registry Data Book for fiscal year 2025 logged over 2.6 million Do Not Call complaints. If your evening ring is usually a robocall, that isn't paranoia.

But look at what that data actually is. Those are consumer complaints about calls people received, on a registry that doesn't even cover business lines. Nothing in it tells you what share of your shop's inbound is junk, and nothing else does either, because a missed-call log can't tell you what a call was. The number you're making this decision on is a number you have never measured.

The deeper problem is that voicemail doesn't screen anything. It treats the robocall and the homeowner with a flooded basement exactly the same, and hands you the same result for both, which is nothing. Screening is a job for something that picks up, listens, and can tell the difference. We covered how an AI receptionist screens spam without dropping real customers separately, and all of it starts with somebody answering.

Does the phone even ring enough after five to matter?

A meaningful share of calls land outside normal business hours. There's no reliable public figure for exactly how many, and any specific percentage you see quoted on this is worth distrusting, so the only number that means anything here is your own.

Composition matters more than count, and the evening mix is genuinely mixed. Some of the 7:15 calls are homeowners who just got in from their own job and finally had a free minute, and those will happily wait until Wednesday. Some of them cannot wait at all. The problem is that a missed-call log gives you no way to tell which is which, so voicemail loses both at the same rate.

It's not as though daytime is solved either. Invoca's 2026 home services benchmarks found that 52% of calls to home services businesses are answered by a person, and that counting only the calls that ran long enough to be a real attempt lifts it to 65% past 15 seconds and 73% past 30 seconds. So the honest read is that the daytime picture is better than the raw figure suggests. After five, when nobody is staffing the phone at all, it goes the other way by design.

What does the voicemail habit actually get right?

More than the pitch usually admits.

The staffing instinct is correct. You can't put a person at a desk at two in the morning for a handful of calls. Median pay for a receptionist runs $18.27 an hour, about $38,010 a year, per BLS OEWS 2025 figures, and that covers one person on one shift. Nights and weekends on top of it is a second hire and a third.

And the owner who watches the log and calls back from the truck at 7:05 is doing something that genuinely works. It's free, it's fast, and a real callback in twenty minutes beats most of what's sold as a fix for this. It holds right up until the evening you're on a ladder, on another call, or asleep, which is the same gap that sent the calls to voicemail in the first place.

Not every evening call is an emergency either. Some are quote questions, some are scheduling, some are a supplier who forgot what time it is. Those genuinely do wait until morning, and treating every 7pm ring as a lost job overstates the case.

And voicemail isn't useless. It's a fine fallback and it costs nothing to have. The mistake was never having it. The mistake is letting it be the only thing there.

So what do you do instead at seven at night?

Start by separating two things that got welded together: answering a call, and responding to it. Nobody has to be awake for a call to be answered. What has to happen at 7pm is that a real conversation takes place, the details get captured, and the caller has a reason to stop looking.

Willison answers every inbound call, 24/7, 365, in seconds, effectively on the first ring. It greets the caller, asks what's wrong, qualifies the job, and books it straight to the calendar. When a storm or a neighborhood outage stacks the calls on top of each other, it answers them at once: the tenth caller gets picked up as fast as the first, no busy signal, nobody sitting on hold.

Then you decide what reaches you tonight and what waits for morning. That's a rule you set once, in daylight, instead of a judgment call you make half asleep. Willison can text you the details of a call as it comes in if that's what you want, and leave you alone until morning if it isn't.

On a restoration call it does the same work in a different tone: answer, calm the caller down, triage the loss (water category, source, how long it's been running, the address), and hand your team the details. It doesn't dispatch a crew, send a truck, or tell a homeowner when somebody will arrive. It makes sure the call is yours to answer.

It also doesn't decide who gets out of bed. Whether tonight's emergency puts somebody in a truck at 2am is still your on-call rotation and your call to make. What changes is that the job is captured, qualified and on the calendar while the caller is still on the phone, instead of sitting in a log as a number you'll squint at over coffee.

And none of it fixes a calendar that's already full. If your first opening is nine days out, answering at 7pm just gets you a faster no. Some callers will also work out they're talking to software, and a few of those will hang up, though a caller you can see and call back is more than a beep leaves you. We wrote up how that actually goes on the phone separately. The question was never whether every call ends perfectly. It's whether the call reaches you at all, or ends at a tone.

You can test the first thirty seconds of it yourself in about two minutes. Talk to the live Willison demo right in your browser on willisonhq.com and give it the call you sent to voicemail last Thursday.

Frequently asked questions

It helps the caller who can wait, and it's a courtesy worth keeping. What it can't do is hold anyone. Somebody watching a stain spread across a ceiling hears "we open at eight" as a no and moves on before the beep. Treat the greeting as manners rather than coverage, and don't let a well-written one talk you into believing the hours are handled.

It beats silence, because the caller gets something back within seconds instead of nothing at all. Where it runs out is the conversation. A text can't ask what's wrong, can't tell an emergency from a quote request, and can't put an address and a time on your calendar. It opens a thread and then waits for a stranger to keep typing at eight at night.

Plenty of owners do, and it works right up until it doesn't. Forwarding assumes you're free to pick up, and the hours you're forwarding are exactly the hours you're driving, eating, asleep, or already on another call. When you can't take it, the caller lands on your personal voicemail instead of the shop's. Same outcome, one number further down the chain.

Pull a month of call records from your carrier or your tracking number, then filter for inbound calls after five that you never returned and that left no message. Nobody can quote you that figure, because it belongs to your phone, your market and your season. Count it once and the whole argument stops being theoretical.