What 3,332 answered calls say about after hours

In short: The number the answering-service industry quotes, that up to 65% of property management calls arrive outside business hours, has no source behind it. We run the voice agents on two lines at one Toronto operation, a property management line and a brokerage line, so we have a complete log instead of an estimate. Over one quarter, 3,332 calls: 19% arrived outside a nine to six week, and 21 calls in the entire quarter came between 10pm and 7am. The property management line, the one the 65% claim is actually about, was the lower of the two at 12.7%. The case for covering the window is still strong, for a reason nobody sells.

Key takeaways

• After hours is real and it is far smaller than advertised. 643 of 3,332 calls, 19.3%, fell outside Monday to Friday, nine to six. Against a nine to five office it is 23.3%.
• The two lines behave differently. The brokerage line ran 31.3% after hours. The property management line ran 12.7%, and its weekend was almost empty at 5.0% of calls.
• Nights barely exist. 21 calls out of 3,332 arrived between 10pm and 7am across 102 days.
• Here is the part that matters: on the property management line, the after-hours window held 12.7% of the calls and 41% of the emergencies. One in four after-hours calls carried the operation's own emergency tag, against one in twenty during business hours.
• 12% of everything that rang the brokerage line was a wrong number or a robocall. Nobody sells you a solution for that, and it is a real line in your volume.

Where the 65% comes from

If you are shopping for an after-hours answering service, one number follows you around: roughly two thirds of calls arrive outside business hours. It appears with slightly different wording on a lot of vendor pages.

We went looking for the source. On Layer3Labs' guide the line reads "Up to 65% of incoming property management calls arrive outside standard business hours", and the page carries no attribution for it. It cites Digible for a different statistic in the same article, so the omission looks deliberate rather than sloppy. Haven's after-hours guide then states that "65% of maintenance calls occur after business hours" and cites Layer3Labs.

Follow that chain and two things happened in one hop. The claim acquired a citation it did not have, and it changed category on the way: all property management calls became maintenance calls, which is a different and much smaller population.

That is not a scandal. It is what happens to a plausible figure in a category where nobody has the data, because the companies with the phone logs are not the companies writing the blog posts. What follows is one operation's log. It is small, it is one market, and every call in it was answered and recorded, which is the part most published numbers cannot claim.

The data

One Toronto property management and brokerage operation, two separate phone lines, both answered by voice agents we built and run. Every call produces a structured record with a timestamp, so the count is a count, not a sample or an estimate.

| | Property management line | Brokerage line |

|---|---|---|

| Calls | 2,142 | 1,190 |

| Average | 21.0 a day | 11.7 a day |

| Busiest hour | 11am | noon |

| Callers per Monday | 31.9 | 13.8 |

| Callers per Saturday | 5.0 | 9.1 |

Both cover 1 June to 10 September 2026, 102 days. 3,332 calls in total.

The reason this denominator is worth something: an operation that misses calls does not know how many it missed. Carrier logs show attempts, not intent, and a voicemail box that nobody filled tells you nothing about who called. Once an agent answers everything, the arrival curve stops being a guess.

After hours, measured

| Definition of the office week | Property management | Brokerage | Both lines |

|---|---|---|---|

| Outside Monday to Friday, 9 to 6 | 12.7% | 31.3% | 19.3% |

| Outside Monday to Friday, 9 to 5 | 15.8% | 36.6% | 23.3% |

Note which line is which. The claim in circulation is specifically about property management calls, and the property management line is the lower of the two by a wide margin. Roughly one call in eight, not two in three.

The brokerage number is higher for a reason anyone who has worked in real estate will recognise. Renters and cooperating agents look at listings in the evening and at the weekend, so 17.5% of brokerage calls came on a Saturday or Sunday. On the property management line the weekend was 5.0%, and Sunday ran 2.6 calls a day.

Nights are almost empty on both. Between 10pm and 6:59am, across 102 days and 3,332 calls, there were 21 calls. Six tenths of one percent.

The busiest hour outside the office week, on both lines, was 6pm. Not midnight. The hour right after the team goes home.

The finding that should decide your purchase

Volume is the wrong thing to buy on, and the log shows why.

The property management line carries an emergency routing tag, applied by the system itself rather than by anyone's later judgment, so it is a logged field and not a label we assigned. 158 of the 2,142 calls carried it.

| | Share of calls | Share carrying the emergency tag |

|---|---|---|

| Business hours | 87.3% | 5.0% |

| After hours | 12.7% | 24.0% |

41% of the quarter's emergencies arrived in the 12.7% of time nobody was at a desk. An after-hours call on that line was roughly five times more likely to be an emergency than a call during business hours.

That is the real case for covering the window, and it is not the case the category makes. You are not buying coverage because most of your volume is at night. It is not. You are buying it because the calls that land there are, per call, the ones that cost the most to miss: water, heat, a door that will not close, someone locked out. The vendor is selling you a volume story about a risk problem.

It also tells you what to build. Thirteen percent of volume does not justify a night shift, and the six calls a week landing after 10pm certainly do not. What it justifies is an always-on first answer with an emergency list your team wrote, hard-coded rather than left to a model's judgment, that rings a named person immediately and lets everything else wait in a morning queue with a callback number attached.

The assumption we got wrong

On the brokerage line, after-hours callers left a callback number 70.7% of the time. Daytime callers left one 72.5% of the time.

We expected a gap. The assumption everyone works from, us included, is that a caller who reaches something other than a person in the evening is a caller halfway out the door. Over a quarter, that assumption is wrong by two percentage points, which is to say it is wrong.

What did change after hours was who was calling. Cooperating agents were 43.9% of daytime brokerage callers and 34.7% of evening and weekend ones. The evening skews toward the public: renters and buyers looking at a listing after dinner, the callers most likely to try the next number on the page if nobody picks up.

The property management line tells the same story from a different angle. One in five of its calls was labelled negative in sentiment, against one in sixteen on the brokerage line, and the share rose slightly after hours. People do not call a property manager in the evening because they are browsing.

What the calls actually were

Brokerage line, where the category labels are cleanest:

| Reason | Share |

|---|---|

| Asking for a specific person | 28.7% |

| Booking or asking about a showing | 14.3% |

| Is this still available | 12.3% |

| Wrong number or robocall | 12.2% |

| Maintenance or a property issue | 11.2% |

| Everything else | 21.3% |

The largest single category is not a question at all. Nearly three calls in ten were someone trying to reach a named person. On a brokerage line the main job of whatever answers the phone is routing, and most of the value is in getting the message to the right agent with a number attached, not in answering anything.

One more thing the vendor pages do not mention: an eighth of that line's volume is junk. 145 calls, 12.2%, were wrong numbers, robocalls or silence, and after hours the share rose to 14%. If you are pricing a per-call answering service, that is one call in eight billed for nothing.

What we would tell an operator with this in front of them

Cover 6pm to 9pm first, then Saturday, then decide about nights. That is where the volume is. Overnight coverage buys you a handful of calls a quarter, and you can add it later for very little once the rest is running.

Buy for the emergency concentration, not the volume. If your line looks like the property management one, the argument is that a quarter of what arrives in that window is urgent, not that most of your calls arrive in it. Price it against the cost of one missed flood, not against call counts.

Do not pay per call without asking about junk. At 12% wrong numbers and robocalls, a per-call or per-minute contract is charging you for a meaningful slice of nothing. Ask any vendor how they bill a robocall and watch what happens.

Measure your own line before you believe any of this, including this. Two lines at one operation in one city over one summer, and they disagreed with each other by a factor of two and a half. A student-housing portfolio in August, a condo board in February, a plumbing company in a cold snap: all different curves. The method is the transferable part. Answer everything, log everything, count it after ninety days.

Method, and what this is not

Every call to both lines is answered by a voice agent, which produces a structured record: timestamp, caller intent, whether a callback number was captured, a sentiment label, and on the property management line an emergency routing tag. Records were pulled for 1 June to 10 September 2026, deduplicated on timestamp, and reduced to a set of derived fields with no caller name, number, email or address attached. Local time throughout, so daylight saving is not a factor inside the window.

Timestamps, callback capture, the emergency tag and the sentiment label are read straight off the logged fields. Reason categories were assigned per record and are the softest numbers here, which is why the category table above covers only the line where the labels were cleanest and why no share is quoted to a decimal it cannot support.

What it is not: it is not multifamily, it is not the United States, and it is not a large sample by the standards of a real study. It is one operation. We are publishing it because a specific, limited, honestly bounded number beats a round one with no source, and because the operations that hold this data usually never publish any of it.

If you want the same view of your own phone line, that is the first thing an audit produces, and the call is free. How the phone layer gets built is on our property management answering service page and AI voice agents. What the after-hours build looks like in a property management operation specifically is in an AI answering service for property management.

Frequently asked questions

What percentage of property management calls come in after hours?

On the property management line we measured, 12.7% of calls arrived outside Monday to Friday nine to six, and 15.8% outside a nine to five week. The brokerage line at the same operation ran 31.3%, and the two together 19.3%. The widely quoted 65% figure traces back to a vendor page that states it without a source. Your own number depends on your market, your property type and when your office closes, and the only reliable way to get it is to answer and log every call for ninety days.

Do property management emergencies really happen in the middle of the night?

Rarely in the small hours, and often in the evening. 21 calls out of 3,332 arrived between 10pm and 7am across 102 days. But 41% of the property management line's tagged emergencies arrived outside business hours, and an after-hours call there was about five times more likely to be an emergency than a daytime one. The middle of the night is quiet. The evening and the weekend are not, and that is where the risk sits.

Is an after-hours answering service worth it?

For a property management line with a similar emergency profile, yes, and the case is about risk density rather than call volume. For a brokerage line it is a capture argument instead: a third of calls, mostly renters and buyers in the evening, who will try the next number if nobody answers. Before signing anything, get the hourly distribution of your own calls and ask the vendor how they bill wrong numbers.

Why are these numbers lower than the ones on other websites?

Two reasons worth separating. The first is that most published figures are estimates produced by companies selling the remedy, and at least one of the most-quoted ones carries no source at all. The second is genuine: different populations give different answers, and our own two lines differed by a factor of two and a half. A maintenance-only line in a 5,000 unit multifamily portfolio is not a brokerage line in Toronto. We would rather publish a small honest number with the window and the method attached than repeat a large one we cannot trace.

Can we get this analysis for our own phone line?

Yes, and it is the first thing worth doing. It needs every call answered and logged, which is what the voice agent layer does on day one, and then ninety days of patience before anyone draws conclusions. The discovery call is free and thirty minutes, and if your volume is low enough that the answer is do not buy this, that is the answer you will get.