Are most of the calls you miss just spam anyway?

You climb out of a crawlspace and there are six missed calls on the screen. Two are flagged as likely spam, one's a number you don't recognize from two states over, and three are local numbers you've never seen. No voicemails. It's tempting to decide the whole list was junk and get back to work.

Here's the short answer. Some of them are, and no published benchmark can tell you how many on your line. But the spam share isn't what a missed call costs you. Junk adds calls to your list. It doesn't take real callers off it (on a single-line phone it can even add to them), so two shops that each miss ten real customers lose the same work, whether their lists are mostly robocalls or barely any. What decides the cost is how many real callers are on your list, and a two-week callback test will show you.

Where we stand, before the argument starts. Willison answers phones for a living, so we have an interest in this answer being "no, they're customers." Where the honest answer is "partly yes," we say so, and one of the outcomes at the end is that you don't need to buy anything.

Why does it feel like every missed call is junk?

Because a lot of what rings is junk, and you've personally picked up enough of it to know. The volume is real and national. The FTC's Do Not Call Data Book for fiscal year 2025 counts over 2.6 million Do Not Call complaints in the year, and over 258 million numbers on the registry.

Those numbers prove the flood exists. They don't tell you what share of your own line it is. They're consumer complaints about calls people received, not a measurement of any business's inbound calls, and anyone turning them into "most of your calls are spam" is stretching them past what they say.

There's also a cost to the junk you do answer, and it's the real reason owners stop picking up. Every pitch you take on a roof is a climb down for nothing. That cost is real. It's just a different cost from the one this post is about.

Myth: no voicemail means it wasn't a customer

This is the belief doing the most damage, because it feels like evidence. A real customer would have left a message, right?

Mostly, they don't. Per Invoca's platform data (2024), fewer than 3% of callers who get pushed to voicemail leave a message. An empty voicemail box is what a normal day looks like whether your missed list was all robocalls or all homeowners.

Be careful with that in both directions. Silence doesn't prove the caller was a customer either. It just means the voicemail box can't answer the question you're asking it. We dig into the voicemail half on its own in do customers leave a voicemail when you don't answer.

Myth: most of what rings is junk, so most of what I miss is junk

That only follows if your misses are a fair sample of your calls. There's no reason to assume they are, and which way they lean depends on why you missed them.

Blind misses are the ones you never saw: you were on a roof, under a house, on another call, asleep. They follow your busy hours and the weather, not the robocall schedule. The day the AC calls stack up is the day you can't get to the phone. So blind misses plausibly lean toward your busiest hours. Nobody has measured whether those hours carry more real callers or just more calls, so don't assume a tilt either way. Your blind and chosen marks in the test below will show it.

Chosen misses are the ones you saw and let ring. Split those in two. The ones your carrier stamped as likely spam are your likeliest junk, though labels miss some dialers and occasionally flag a person on a fresh number. The unflagged unknown numbers are a different pile. That's where robocalls on spoofed local numbers sit, and it's also where every first-time customer sits, because a new customer is an unknown number by definition. Letting that pile ring is a screen on caller ID, and caller ID can't tell you what someone wants.

One more correction to the belief: "not a lead" doesn't mean "junk." The tech's customer asking when he's coming, a supplier with a delivery problem, a sub cancelling tomorrow. None of those are new work, and none of them are spam either.

Myth: if it's mostly spam, the misses barely cost anything

This one sounds like arithmetic, which is why it sticks. It's actually the wrong arithmetic. What a missed call costs depends on the count of real callers you missed, not on what percentage of the list they make up.

Take two made-up shops to show the shape of it. Shop A misses 40 calls in a month: 30 junk, 10 real. Its list is 75% spam. Shop B misses 12: 2 junk, 10 real, so 17% spam. Shop A's owner feels better about it. Both lost the same ten callers.

The same logic cuts the other way, and it should. If your count of real callers is small, the loss is small, however clean or dirty the list looks. Junk can't make a short list of real misses expensive.

And on a single-line phone, junk can add real misses rather than just sit beside them. The two minutes you spend getting rid of a listing pitch are two minutes when the next caller gets voicemail.

The part of the belief that's true

A raw missed-call list does overstate lost customers. Some of it is plain noise.

Invoca's 2026 home services benchmarks find 52% of all calls answered by a person, 65% for calls lasting longer than 15 seconds, which Invoca says filters out misdials and quick hangups, and 73% for calls over 30 seconds. The answer rate climbs as the short calls come out, so the short calls are disproportionately the unanswered ones. Some of what shows up as a miss on your phone was a pocket dial, a wrong number or a dialer that never had a person behind it.

Repeat rings inflate the list too. One person calling three times is one caller. But read the burst, don't just collapse it: three rings in two minutes from the same number can mean somebody has a problem right now.

And not every real caller was a job. In the same Invoca report, 38% of calls answered by a person were leads, and 45% of those leads converted on the call. Those are shares of all answered calls, not of confirmed customers. And if you let unknown numbers ring, the calls you answered were partly pre-screened, so your missed list probably runs junkier than that 38% implies. Your own test settles it, not the benchmark. If you want to turn real misses into jobs, how many jobs a two-truck shop loses to unanswered calls walks through that math.

One fence on all of Invoca's numbers. They're averages of calls tracked on Invoca's platform across its customers, not your line. A number that's been on your truck doors and your listing for fifteen years has had a long time to land on dialing lists, so it may draw more junk than their average. That cuts toward the belief, and it's one more reason to measure your own line instead of borrowing anybody's mix.

What does an HVAC line say about it?

HVAC is where this question bites hard, because it's one of the trades where the fewest calls reach a person. In Invoca's 2026 report, 34% of calls to HVAC companies were answered by a person, the second lowest of the nine sub-industries measured. Plumbing companies answered 74%.

Of the HVAC calls that did get answered, 45% were leads, against 38% across home services. Plumbing ran at 30%. So the HVAC calls that reach a person are, if anything, more often real demand than average, not less.

That doesn't prove the unanswered HVAC calls are leads too. Nothing measures the calls nobody picked up, and there's a second reading: if HVAC shops let more unknown numbers ring, the calls they do answer would look cleaner. Plumbing's high answer rate and low lead share fit that same reading. So the HVAC numbers don't refute the "it's all spam" belief, and they don't back it either.

How do you find out what's on your own list?

Ring them back. For two weeks, call every missed number the same day, except the ones your carrier flagged as likely spam, which you just tally. Skip international numbers and area codes you can't place. If a recording answers, hang up. Sort each one into a column:

• Still needs work. A real job, and they haven't hired anyone yet.
• Already hired someone. A real job you lost.
• Existing customer. A booked job, a warranty call, a maintenance customer. Not new work, but not free to miss either.
• Supplier, sub, personal or other.
• Junk. A recording, a disconnected line, a salesperson.
• Never called you. A confused stranger whose number was spoofed.
• Couldn't reach. Keep this as its own column, not dropped data. It could be a lost job, a spoofed number, or someone who ignores unknown callers, so don't count it toward either side.

Mark each one blind or chosen while you're at it. That tells you whether your misses come from being busy or from screening, and the fix for each is different.

Two cautions on reading it. A same-day callback may be faster than you manage on a normal week, so your "already hired someone" column can come out lower than it would without the test. And two weeks in a shoulder season says little about July. Run it in your busy season, or run it twice.

What should you do with the answer?

If the first two columns come back nearly empty across a busy stretch, your misses really are mostly cheap, and the right move is nothing at all. Keep your money. If they don't, read the blind and chosen marks before you buy anything, because they point at different fixes. Each option below has a cost, a way it fails, and a case where it's the right call.

• Screen less. Free. If most of your real misses were chosen ones, unknown numbers you let ring, the cheapest fix is picking up more of them. It costs you interruptions, and it does nothing for the calls you never saw.
• Carrier spam blocking or a call-filter app. Cheap, often free. It cuts the interruptions, which is the right fix if junk is the real problem and your real misses are few. It screens on the number, so it can't help with a call you were too busy to answer, and it can block a real caller it misjudges.
• The same-day callback habit. Free, apart from your time. It wins when your real misses are a handful a week and they aren't urgent. It's slower than the caller, so it loses whoever kept dialing, and your "already hired" column is the measure of that.
• An automatic text when you miss a call. Cheap, and it reaches the caller fast, which suits a line where most callers can wait an hour. It still asks them to wait, and a text isn't an answer to someone with water on the floor. The trade-offs are laid out in answering service vs missed-call text-back.
• A human answering service. A trained person picks up, which some callers prefer. It's a monthly bill, and how well it books depends on the script and the person on shift. It wins when your callers expect a person and your intake is simple to script.
• Hiring or answering more yourself. Someone who knows your business on the line. A wage or your own hours, and still one caller deep when the phone stacks up. It wins when the calls are steady enough to keep a person busy.
• An AI receptionist like Willison. A monthly fee. Some callers hang up on an automated voice, and it can get things wrong: mishear an address, file a job under the wrong type, or take a message that doesn't reach you. It makes sense when your real misses are blind ones that cluster on your busiest days, because Willison answers in seconds, 24/7, and takes many calls at the same time.

Willison reads the address back and won't tell a caller they're booked until the booking's written. Its monthly report shows what it booked and what it filtered as spam on your line. That's the vendor's own count, so the callback test is still worth running first.

If you'd like to hear how it handles a caller first, press play on a real call on willisonhq.com. It's a call to a sample HVAC shop, start to finish.

Frequently asked questions

Nobody can give you that number honestly. No published benchmark measures the junk share of a contractor's missed calls, and any figure you're quoted without a source is a guess. It also varies by line: an old number on truck doors and listings can draw more robocalls than a newer one. Two weeks of same-day callbacks will tell you more about your own phone than any average.

It's worth thirty seconds, because the log can't tell you who it was. A one-ring hangup can be a misdial or a dialer, and Invoca's data shows short calls are more likely than longer ones to go unanswered. But it can also be someone who heard the ring go on, gave up and moved to the next number. Skip it if it's an international number or an area code you can't place, and if a recording picks up when you call, hang up and move on.

Yes, on a phone that can only take one call at a time. While you're getting rid of a sales pitch, the next caller hears ringing or voicemail, and fewer than 3% of voicemail-bound callers leave a message, per Invoca's platform data. A business phone system with a second line, a person, or a service that takes several calls at once removes that particular cost.