We called 10 Chicago plumbers at 8:32 AM. Two could come, one today.
Ten plumbing listings in ZIP 60625 were dialled at 8:32 AM on a Wednesday. Seven reached a human voice, two confirmed they could come out, and one could make it the same day. Not one named a price.
It was 8:32 on a Wednesday morning in Chicago, and a homeowner with one clogged toilet needed a plumber the same day. It is the easy hour for the emergency claim: every office has just opened, the phones are staffed, and the workday still has ten hours in it. Boocki found twenty plumbing listings in ZIP 60625, dialled ten of them in one burst, and asked each the same two things. Could they handle a clogged toilet, and how soon could they come. The polite answer is not the same as an appointment.
- 20
- plumbers foundZIP 60625, 18 names
- 10
- called8:32 AM CT, one burst
- 2
- confirmed they could comeof the ten
- 1
- could come the same day“in an hour, couple hours”
How to read this
Four phrases carry the piece, and the first two are the easiest to confuse.
- Reached a human voice: a person at the provider’s end spoke. Not a voicemail greeting, not a recorded menu, not a carrier announcement.
- Confirmed they could come: the product’s own availability field was set to true because the provider said so on the call. It is a stricter test than a person picking up.
- Same day: the provider named a window inside the day of the call. A window of “tomorrow” is still a yes, but it is not the same promise.
- Named a price: a number a provider said out loud on the phone. Not an invoice, not a range, not an average.
The method is one job, one ZIP code and one frozen search phrase. The ZIP is 60625 (Chicago); the phrase is plumber toilet clog unclogging, filed on the “as soon as possible” setting. Ten calls started within three seconds of each other at 8:32 AM CT on Wednesday 29 September 2026, so no provider was reached in a different hour or a different mood.
The claim being tested
Half the list carried the “Open 24 hours” label, and the request was filed as urgent. Those two facts together make a promise a homeowner can check: at 8:32 AM on a weekday, how many of these lines would answer with a person, and how many of those people would commit to a visit.
A listing’s hours are a scraped field, not a switchboard. The useful question is not whether the sign says open. It is what happens when the phone is answered, which is what the next table records.
The ten calls, one by one
Everything below reconciles to this table. No provider is named: the recordings contain the homeowner’s street address, a callback number and the first names of the people who answered, so the call numbers are the identifiers here. The “Confirmed” column is the product’s own availability field, not our reading of the transcript: Yes is a provider who said they could come, No is one who declined, and None is a call that never got an answer.
| # | Listed | What happened | Length | Confirmed | Price |
|---|---|---|---|---|---|
| 1 | 7 AM to 9 PM | A recorded menu, then a rep picked up; the call ended as the request was being made | 0:56 | None | None |
| 2 | 7 AM to 9 PM | A second listing for the same business; a recorded menu, then a rep who named a different company, and the call ended at the introduction | 1:09 | None | None |
| 3 | 8 AM to 4 PM | The line opened, and no conversation was captured | 0:15 | None | None |
| 4 | Open 24 hours | A recording, then the line dropped as the agent introduced itself | 0:10 | None | None |
| 5 | Open 24 hours | A rep answered, and the call was cut off mid-introduction | 0:30 | None | None |
| 6 | Open 24 hours | A rep confirmed the job, asked to inspect the toilet before quoting, and gave a window | 1:18 | Yes | None |
| 7 | 8 AM to 8 PM | A rep confirmed the job and gave their soonest window | 1:32 | Yes | None |
| 8 | Open 24 hours | A rep named a two-hour window, then refused to serve a caller they learned was an AI assistant | 2:03 | No | None |
| 9 | Open 24 hours | A voicemail greeting took the call | 0:03 | None | None |
| 10 | 7 AM to 5 PM | A rep asked who was calling, then said they were not interested | 0:43 | No | None |
One row needs a note before the summary. Calls 1 and 2 went to two provider rows that share a business name, so the same business was dialled twice, and the second call reached a rep who introduced a different company name. A listing’s phone line is not always the listing’s brand. Beyond that, the list held twenty rows under eighteen names, so it was not deduplicated before dialling.
The funnel
The same ten calls, narrowed six times. Every row is the sum of rows in the table above, and the “Calls” column says which, so nothing here is an orphan number.
| Step | Calls | Count | Share |
|---|---|---|---|
| Plumbing listings found | n/a | 20 | n/a |
| Called | 1 to 10 | 10 | 100% |
| Connected by the carrier | all but 9 | 9 | 90% |
| Reached a human voice | 1, 2, 5, 6, 7, 8, 10 | 7 | 70% |
| Gave any arrival window | 6, 7, 8, 10 | 4 | 40% |
| Confirmed they could come | 6, 7 | 2 | 20% |
| Said they could come the same day | 6 | 1 | 10% |
| Named a price | None | 0 | 0% |
The 90% is the number an ordinary call log would print as “answered,” and it is the least useful row here: it counts a three-second voicemail greeting (call 9) the same as a two-minute conversation. The 20% is the tier that matters, because it is the only one a homeowner can act on. Note that the arrival-window row and the confirmed row do not match. Calls 8 and 10 named a time or engaged with the question before declining, so they sit a rung above where the availability field puts them.
The eight that never confirmed
Eight of the ten calls failed to produce a yes, in six different ways. Listed in call order so they scan against the master table.
| # | Failure | Length | What the recording shows |
|---|---|---|---|
| 1 | Ended at the greeting | 0:56 | A recorded menu, then a rep came on; the call ended before the agent finished asking. |
| 2 | Ended at the greeting | 1:09 | A recorded menu, then a rep answered under a name that did not match the listing, and the call ended at the introduction. |
| 3 | No conversation | 0:15 | The line opened and nothing the carrier could transcribe was said. |
| 4 | Dropped at the opener | 0:10 | A recording, then the line dropped as the agent introduced itself. |
| 5 | Cut off | 0:30 | A recorded preamble, then a rep greeted the agent, then the call ended. |
| 8 | Declined as AI | 2:03 | A rep said the soonest was the next two hours, then refused once told the caller was an AI assistant. |
| 9 | No answer | 0:03 | A voicemail greeting answered. Nobody did. |
| 10 | Not interested | 0:43 | A rep asked who was calling, then said they were not interested and ended the call. |
Three of these are carrier-level failures (calls 3, 4 and 9), where the line connected but no person was ever in a position to answer. The other five reached a person and still produced nothing usable: three ended before the ask was finished (calls 1, 2 and 5), and two ended as refusals (calls 8 and 10). Call 8 is the one worth reading twice. It is the only call where a provider named a window inside two hours, and it is also the only call where the provider changed their answer the moment the method was disclosed.
The two yeses, and which one was first
Two providers set the availability field. Both confirmed they could do the job. Neither named a price, and only one of them could come inside the working day.
| # | What they said | Price | Arrival |
|---|---|---|---|
| 6 | Confirmed the job, asked to inspect the toilet before quoting, and wanted the address plus a tenant number | None | In an hour, couple hours |
| 7 | Confirmed the job and gave their soonest window without being pushed on price | None | Tomorrow afternoon |
“Well, I can do in a couple hours. In an hour, couple hours.”
That is the whole answer to the question this job was run to ask. Of ten plumbers called in one burst at 8:32 AM, exactly one put a same-day arrival on the table, and they did it with a hedge attached: an hour, or a couple of hours, depending on what they found. The only other confirmed provider could not come until the next afternoon. Nobody else offered a time at all.
The hedge matters more than it looks. A window of “in an hour, couple hours” is a promise to inspect, not a promise to fix, and the provider said so: they wanted to see the toilet before naming a price. For a homeowner trying to decide whether to wait, that is still the most useful thing anyone said all morning.
What “Open 24 hours” predicted
The list split evenly: five listings carried the “Open 24 hours” label and five carried ordinary daytime hours. If the label meant anything at 8:32 AM, the split would show it.
| Listed hours | Calls | Called | Reached a person | Confirmed | Named a price |
|---|---|---|---|---|---|
| Open 24 hours | 4, 5, 6, 8, 9 | 5 | 3 | 1 | 0 |
| Daytime hours | 1, 2, 3, 7, 10 | 5 | 4 | 1 | 0 |
The two halves are indistinguishable where it counts. Each produced one confirmed provider, neither produced a price, and the daytime half actually reached more people, four to three. The “Open 24 hours” label is a claim about a sign, not about a phone. On this morning, a plumber who advertised eight-to-five was exactly as likely to answer as one who advertised always.
The price nobody would name
Ten calls, four arrival windows, two confirmations, and zero prices. Not one provider said a number. Two of them explained why: they wanted to see the toilet before quoting, which is a fair answer for a job that could be a simple clog or a failed seal. The rest never got far enough to be asked properly, or ended the call first.
A homeowner can plan around “someone can be there in an hour.” It is much harder to plan around a visit with no number attached, because the decision to wait, or to keep calling, depends on a cost that nobody would put on the table. That gap is the most consistent thing in this run.
The arithmetic
The whole job cost less than the average call-out, and the reason is the same reason it produced no prices: almost nobody stayed on the line long enough to name one. Ten calls, twenty listings, eight and a half minutes of talk.
| Metric | Value |
|---|---|
| Plumbing listings found | 20 (18 names) |
| Calls placed | 10 |
| Talk time | 8 min 39 s |
| Longest call | 2:03 (call 8) |
| Shortest call | 0:03 (call 9) |
| Arrival windows given | 4 |
| Prices quoted | 0 |
| Total charged | $2.58 |
| Average per call | $0.26 |
Talk time is the sum of the Length column in the first table: 0:56 + 1:09 + 0:15 + 0:10 + 0:30 + 1:18 + 1:32 + 2:03 + 0:03 + 0:43, which is 8 min 39 s. The ledger holds a single call charge for the job, $2.58, which is the platform’s own price for that audio at $0.30 a minute, billed by the second. It is not money a provider named. The average is $2.58 divided by the ten calls placed.
What we would change next time
- Call all twenty, not ten. Half the list was never dialled, so the top of the funnel is a selection rather than the whole population, and the ten that were skipped could move any of these numbers.
- Time the run to the claim.This was filed in the small hours and dialled at 8:32 AM, the easiest hour there is. The “emergency” claim is really tested at 2 AM or on a Sunday, and ten weekday-morning calls say nothing about either.
- Disclose the assistant earlier. Call 8 named the soonest window of the day, then reversed the moment it learned the caller was an AI assistant. If the point is to find the earliest real arrival, the disclosure belongs at the top of the call, not two minutes in.
- Deduplicate the list before dialling. Twenty rows carried eighteen names, and one business appeared twice under two names. Both of those listings were dialled (calls 1 and 2), so one of the ten calls spent itself on a company the job had already reached.
- Ask for a price in the same breath as the time. The four providers who gave a window were never successfully asked what the visit would cost, which is the second half of the question a homeowner is really asking.
Why publish the whole funnel
A tool that reported “one plumber can come today” would be hiding the most useful thing it learned: what happened to the other nine calls, and how little separated the line that said yes from the eight that did not. The value of calling ten at once is not the single appointment at the end. It is the shape of the morning: how many lines answered, how few would commit, and how easily the one same-day promise could have been missed if the call had stopped a minute earlier.