The Call Center Doctors Logo Get Started

The questions a call center vendor would rather you did not ask

Almost every list of questions to ask an outsourced calling partner was written by somebody who wants to be hired, or by a directory paid to introduce them. So the questions that would genuinely cost a floor the deal are the ones that are missing. This is the version written by a floor that has to answer them, and every question on it arrives with the federal document or the measured call data you can check the answer against.

Take it into the meeting. Ask the question, write down what they say, then read their answer against the evidence printed underneath it. A vendor who cannot answer a question has told you something; a vendor whose answer has no denominator in it has told you more.

Download the CSV

No signup, no email wall, nothing to fill in first. The CSV opens straight into a spreadsheet with a column for their answer and a column for your note.

How to use it

Ask the questions in the order they are printed. The compliance ones are first because they are the cheapest to ask and the most expensive to skip, and because a floor that answers them crisply has almost certainly answered them before, which is itself the finding. Leave the arithmetic questions until the vendor is comfortable: they take longer and the interesting part is not the number but whether the denominator arrives with it.

There is no scoring on this sheet and that is deliberate rather than an omission. A total would let a strong answer on wages cancel a missing answer on removal requests, and those are not the same kind of risk. Read the answers, not a sum of them.

Where the answers underneath come from

Two kinds of source, and the sheet marks which is which on every line. The compliance lines describe published federal documents -- the Telemarketing Sales Rule, the FCC consent rules, the Bureau of Labor Statistics wage series -- and quote them rather than summarising them. The measurement lines come from this floor, and the two corpora they rest on are stated here before any share drawn from them appears below.

10,794 outbound roofing appointment-setting calls placed Feb-Mar 2026, every one transcribed and analysed (10,848 transcribed in total; 54 belonged to a non-roofing plumbing list and were removed)
Size of the ccdocs outbound roofing call corpus the first-party figures are measured onapps/airoofing/files/10k_agent_scorecards.md (period, per-agent totals), apps/airoofing/files/10k_winning_vs_losing.md (roofing-filtered totals), apps/airoofing/roadmap.md (unfiltered totals), 2026-Q1
89,858 inbound calls arrived at a ccdocs phone number over the 90 days from 2026-05-03 to 2026-08-01, across nine client in-groups. 86,082 (95.8%) were routed to an agent queue and 79,553 reached one; 48,761 were handled by a live agent. Of those 48,761 answered calls, 24,700 (50.7%) ended as a hangup, 9,397 (19.3%) were a wrong number, dead air, an answering machine or a prank call, 281 (0.6%) booked an appointment, and the mean call ran 23.8 seconds.
Size and composition of the inbound call corpus the ccdocs inbound figures are measured onProduction VICIdial `asterisk` schema (tables vicidial_did_log, vicidial_closer_log, vicidial_list, vicidial_list_archive), read through the SELECT-only sarah_ro proxy documented in agent-ops/packages/vici-read-proxy/README.md; every query reproduced by apps/ccdocs-astro/scripts/derive-inbound-call-facts.py, 2026-05-03/2026-08-01

Read the first-party figures as one floor over one window, published so you can see the shape of a complete answer rather than as an industry benchmark to hold a vendor to. Where our own number is unflattering it is printed anyway, because a sheet that only quoted our good numbers would be the marketing document it is meant to replace.

This is not legal advice

The compliance questions state the SCOPE of published federal documents. They do not say that any rule applies to any particular business, which turns on facts about the campaign that no web page can settle. Two of the consent rules cited here have moved more than once already, so read the current text before relying on this sheet later, and take any answer that reads like an obligation to counsel.

The rules a driveway close does not remove

Four questions, and they come first because a vendor who cannot answer them is a vendor whose exposure travels with the campaign they run for you. Every contractor sale is signed in person, so buyers are told -- and sellers repeat -- that federal telemarketing rules belong to somebody else. The exemption behind that idea is real. It is also narrow, and it withholds exactly the provisions that carry the penalties.

  1. Which parts of the Telemarketing Sales Rule do you treat the face-to-face exemption as removing?

    The exemption at 16 CFR 310.6(b)(3) covers a call in which the sale is not completed, and payment is not required, until after a face-to-face sales presentation -- and it expressly does not apply to the requirements of 310.4(a)(1), (a)(8), (b), and (c), which are the threats-and-intimidation, caller-ID-transmission, do-not-call and abandoned-call, and calling-hours provisions.

    What they said

  2. Whose clock decides that a number is out of hours -- the floor, or the homeowner?

    The federal residential window runs 8:00 a.m. to 9:00 p.m. local time at the called person location, not at the caller location. A floor dialling east leaves the window earlier than its own wall clock suggests.

    What they said

  3. What abandonment rate do you run, and what definition did you measure it with?

    A call is abandoned if a person answers and is not connected to a live representative within 2 seconds of their completed greeting; the safe harbor caps abandonment at 3% of calls answered by a person and requires at least 15 seconds or 4 rings before disconnecting. A rate quoted without the definition behind it is not comparable to that one.

    What they said

  4. Which established-business-relationship clock do you scrub against?

    There are two federal clocks and they are not the same. The FTC measures from a purchase, rental, lease or financial transaction within the 540 days before the call, or an inquiry or application within the 90 days before it. The FCC measures from a purchase or transaction within the eighteen months before the call, or an inquiry or application within the three months before it.

    What they said

Read this before you use the section

These lines describe what published federal documents say, and nothing else. Whether a particular floor sits inside the FTC jurisdiction turns on facts about the campaign that no web page can settle: the rule defines telemarketing as a plan or campaign involving more than one interstate telephone call, so a single-state operation may fall outside the FTC while remaining inside the FCC rules and its own state statute. Ask the question, read the answer against the citation printed under it, and take the result to counsel. Nothing here is legal advice and nothing here says a rule applies to you.

It does not. 16 CFR 310.6(b)(3) exempts calls in which the sale is not completed, and payment is not required, until after a face-to-face sales presentation -- but that exemption expressly does not apply to the requirements of 310.4(a)(1), (a)(8), (b), and (c), which are the threats-and-intimidation, caller-ID-transmission, do-not-call and abandoned-call, and calling-hours provisions.
Whether the Telemarketing Sales Rule face-to-face exemption covers do-not-call, abandoned calls, calling hours and caller IDFTC Telemarketing Sales Rule, 16 CFR 310.6(b)(3), https://www.ecfr.gov/current/title-16/chapter-I/subchapter-C/part-310/section-310.6, 2026
8:00 a.m. to 9:00 p.m. local time at the called person location
Federal permitted calling window for outbound telemarketing to a residenceFTC Telemarketing Sales Rule, 16 CFR 310.4(c), https://www.ecfr.gov/current/title-16/chapter-I/subchapter-C/part-310/section-310.4, 2026
A call is abandoned if a person answers and is not connected to a live representative within 2 seconds of their completed greeting; the safe harbor caps abandonment at 3% of calls answered by a person and requires at least 15 seconds or 4 rings before disconnecting
Federal definition of an abandoned telemarketing call, and the safe-harbor abandonment capFTC Telemarketing Sales Rule, 16 CFR 310.4(b)(1)(iv) and 310.4(b)(4), https://www.ecfr.gov/current/title-16/chapter-I/subchapter-C/part-310/section-310.4, 2026
They are not the same. The FTC Telemarketing Sales Rule (16 CFR 310.2) measures an established business relationship from a purchase, rental, lease or financial transaction within the 540 days before the call, or an inquiry or application within the 90 days before it. The FCC (47 CFR 64.1200(f)(5)) measures it from a purchase or transaction within the eighteen months before the call, or an inquiry or application within the three months before it.
The two different federal established-business-relationship clocksFTC Telemarketing Sales Rule, 16 CFR 310.2, https://www.ecfr.gov/current/title-16/chapter-I/subchapter-C/part-310/section-310.2; FCC rules, 47 CFR 64.1200(f)(5), https://www.ecfr.gov/current/title-47/chapter-I/subchapter-B/part-64/subpart-L/section-64.1200, 2026

What happens when a homeowner says stop

Four questions about the least glamorous part of the work, which is also the part a buyer inherits. A removal request is the one instruction on the call that outlives the campaign: it follows the list, not the vendor, and a floor with no answer here is a floor whose list you would rather not adopt.

  1. How does a spoken stop request reach the list, and how long does that take?

    The core federal duty is live: a called party may revoke consent using any reasonable method to clearly express a desire not to receive further calls or text messages, and the revocation must be honoured within ten business days from receipt. The broader revoke-all component is currently waived and under reconsideration rather than in force.

    What they said

  2. Who buys the Do Not Call registry access, and for how many area codes of data?

    The published annual rate is $82 for each area code of data accessed, up to a maximum of $22,626, with no charge for the first five area codes. The FTC resets the fee each fiscal year, so read the current rule rather than a quoted figure.

    What they said

  3. What share of the homeowners you dial ask to be taken off the list?

    On our own outbound roofing corpus, 439 of 10,794 calls (4.1%) ended in a do-not-call request. Those six disposition codes cover 5,058 of the 10,794 calls, or 46.9%, so they do not partition the corpus and must not be read as a whole.

    What they said

  4. What do you assume the homeowner already believes about an unknown number?

    About 258.5 million registrations were active on the federal Do Not Call registry as of 2025-09-30, up roughly 1.9% over FY2024. Overall complaints rose in FY2025, while unwanted-call reports remain about 48% below FY2021, when the FTC received approximately five million reports about unwanted calls.

    What they said

Read this before you use the section

The regulatory lines describe documents, not postures, and the status of both federal consent rules has moved more than once -- re-read them before relying on this sheet a year from now. The first-party line is our own outbound cold dialling and describes nothing about how an inbound caller behaves. Ask the question, keep the answer in writing, and send anything that reads like an obligation to counsel.

Split, and the two halves must not be stated as one. The core duty is in force: a called party may revoke consent using any reasonable method to clearly express a desire not to receive further calls or text messages, and the revocation must be honoured within ten business days from receipt. The broader "revoke-all" component -- which would make a revocation given on one topic apply to unrelated future calls and texts from that caller -- is WAIVED and not yet effective; the FCC extended its effective date to January 31, 2027 while it decides whether to change the rule.
Current federal status of the FCC consent-revocation rule, 47 CFR 64.1200(a)(10)FCC rules, 47 CFR 64.1200(a)(10), https://www.ecfr.gov/current/title-47/chapter-I/subchapter-B/part-64/subpart-L/section-64.1200; FCC waiver order released 2026-01-06, https://docs.fcc.gov/public/attachments/DA-26-12A1.pdf, 2026
The annual fee is $82 for each area code of data accessed, up to a maximum of $22,626, and there is no charge for accessing the first five area codes of data.
The annual fee to access the US National Do Not Call RegistryFTC Telemarketing Sales Rule, 16 CFR 310.8(c), https://www.ecfr.gov/current/title-16/chapter-I/subchapter-C/part-310/section-310.8, 2026
Of 10,794 outbound roofing calls: 2,827 (26.2%) ended in a hang-up, 1,173 (10.9%) in an explicit "not interested", 439 (4.1%) in a do-not-call request, 329 (3.0%) did not qualify, 212 (2.0%) booked an inspection and 78 (0.7%) asked for a callback
How outbound roofing calls actually endapps/airoofing/files/10k_winning_vs_losing.md (dataset header) and apps/airoofing/roadmap.md ("The Numbers" table); percentages computed against the 10,794 denominator, 2026-Q1
About 258.5 million active registrations as of 2025-09-30, up roughly 1.9% over FY2024. Overall complaints ROSE in FY2025, while unwanted-call reports remain about 48% below FY2021, when the FTC received approximately five million reports about unwanted calls.
Active registrations on the US federal Do Not Call registryUS Federal Trade Commission, National Do Not Call Registry Data Book FY2025 press release (2025-12-11), https://www.ftc.gov/news-events/news/press-releases/2025/12/ftc-releases-annual-do-not-call-registry-data-book, 2025

The record of the call

Two questions, and they are the ones a buyer most often skips because the answers sound like plumbing. They are not. A floor that cannot produce the call it made a year ago cannot defend it either, and a floor that will only ever play you calls it chose has told you which calls exist.

  1. How long do you keep the call record, and what is in it?

    The federal retention period is five years from the date the record is produced, under 16 CFR 310.5(a). The retained set includes a record of each telemarketing call covering the calling number, called number, date, time and duration, plus advertising material, scripts, prerecorded messages, consent records and do-not-call requests.

    What they said

  2. Can I hear calls I select from a date I select, rather than calls you selected?

    We publish 20 consent-cleared roofing appointment call recordings on this site, and it is a curated booked-only reel: the honest use of a published set like that one is to hear how the conversation goes, never to infer a rate from it.

    What they said

Read this before you use the section

The retention line describes the rule as written and licenses nothing about what storage costs, how many recordings that amounts to, or whether any given dialler can actually produce them. The recording line is a count of what this site publishes; that set is curated and booked-only, so it can never carry a rate of its own.

Five years from the date the record is produced, under 16 CFR 310.5(a). The retained set includes a record of each telemarketing call covering the calling number, called number, date, time and duration, plus advertising and promotional material, telemarketing scripts, prerecorded messages, consent records and do-not-call requests; scripts and advertising are kept for five years from the date they are no longer used.
How long a telemarketer must retain telemarketing records under the Telemarketing Sales RuleFTC Telemarketing Sales Rule, 16 CFR 310.5(a), https://www.ecfr.gov/current/title-16/chapter-I/subchapter-C/part-310/section-310.5, 2026
20
Consent-cleared roofing appointment call recordings published on ccdocs.comsrc/data/storm-calls.json, rendered at https://ccdocs.com/roofing-storm-appointments/live-calls/, 2026

The denominator under the number

Five questions about arithmetic. A vendor quoting a rate is quoting a fraction, and almost every disagreement between a buyer and a floor is a disagreement about the bottom of it. These are the five places the bottom usually goes missing, each shown with the answer our own floor gives so the shape of a complete answer is visible.

  1. Your booking rate is a fraction. What is on the bottom of it?

    Both denominators, or neither. On our own outbound roofing corpus 3.6% of contacted homeowners booked an inspection (210 of 5,772 contact calls) and 2.0% of all dials did (212 of 10,794 calls). The first flatters, because it hides that most dials never reach a person; the second damns, because it charges the caller for disconnected numbers.

    What they said

  2. How many calls is that share measured on, and over what window?

    Our own shares rest on 10,794 outbound roofing appointment-setting calls placed Feb-Mar 2026, every one transcribed and analysed. A percentage off a call floor means nothing without the n it was measured on.

    What they said

  3. Is that answer rate the pooled floor, or what the median client on it actually gets?

    Ours are different numbers and we publish both: 54.3% of inbound calls reached a live agent (48,761 of 89,858 that arrived) and 61.3% of the calls that reached a queue did, while across the five staffed in-groups the median is 49.2%. The two largest groups are 84.9% of the volume, so anyone quoting the pooled figure as what a client gets is quoting those two.

    What they said

  4. When nobody is free, does the caller wait, or does the call end?

    On our own floor the median wait once a call reaches the switch is 0 seconds and 43,926 of 48,761 answered calls had no measurable hold at all -- but that describes the answered share and nothing else. Of the unanswered calls carrying a recorded wait, 50.1% ended within five seconds and only 38 in the quarter waited 30 seconds or more: when nobody is there the caller is not held, the call is refused.

    What they said

  5. What is the gap between your best and your worst caller on the same list?

    Across the 11 agents with a published scorecard on our own floor, the appointment rate per contacted homeowner runs from 1.2% (2 of 168 contacts) to 5.3% (43 of 806), a 4.5x spread against a blended 3.6%. A blend hides which half of that range a new account would land in.

    What they said

Read this before you use the section

Everything below is our own measurement, published so a reader can judge it rather than accept it, and it describes one floor over one window rather than an industry benchmark. The outbound figures come from a fixed two-month slice that nothing refreshes. The inbound figures come from a separate and later window, they are pooled across every client on the floor, and that is exactly why the median sits printed beside every pooled rate here.

3.6% of contacted homeowners booked an inspection (210 of 5,772 contact calls); 2.0% of all dials did (212 of 10,794 calls)
ccdocs appointment-set rate on outbound roofing appointment-setting callsapps/airoofing/files/10k_agent_scorecards.md (per-agent contact and appointment counts, summed over the 11 published scorecards) plus apps/airoofing/files/10k_winning_vs_losing.md and apps/airoofing/roadmap.md (corpus totals: 10,794 roofing calls, 212 booked), 2026-Q1
10,794 outbound roofing appointment-setting calls placed Feb-Mar 2026, every one transcribed and analysed (10,848 transcribed in total; 54 belonged to a non-roofing plumbing list and were removed)
Size of the ccdocs outbound roofing call corpus the first-party figures are measured onapps/airoofing/files/10k_agent_scorecards.md (period, per-agent totals), apps/airoofing/files/10k_winning_vs_losing.md (roofing-filtered totals), apps/airoofing/roadmap.md (unfiltered totals), 2026-Q1
54.3% of inbound calls reached a live agent (48,761 of 89,858 that arrived); 61.3% of the calls that reached an agent queue did (48,761 of 79,553), over the 90 days from 2026-05-03 to 2026-08-01
Share of inbound calls to ccdocs that reach a live agent, across every client on the floorProduction VICIdial `asterisk` schema (vicidial_did_log joined to vicidial_closer_log on uniqueid), read through the SELECT-only sarah_ro proxy; query and both denominators reproduced by apps/ccdocs-astro/scripts/derive-inbound-call-facts.py, 2026-05-03/2026-08-01
Median 0 seconds. 43,926 of 48,761 answered inbound calls (90.1%) were connected to an agent with no measurable hold at all; the 95th percentile waited 1 second, the 99th waited 4 seconds and the longest wait in 90 days was 53 seconds (2026-05-03 to 2026-08-01)
Seconds a caller waits for a live agent once an inbound call reaches ccdocs, when it is answered at allProduction VICIdial `asterisk` schema (vicidial_closer_log.queue_seconds over agent-handled calls, plus the vicidial_did_log arrival-to-queue leg), read through the SELECT-only sarah_ro proxy; distribution reproduced by apps/ccdocs-astro/scripts/derive-inbound-call-facts.py, 2026-05-03/2026-08-01
Across the 11 agents with a published scorecard, the appointment rate per contacted homeowner ranges from 1.2% (2 of 168 contacts) to 5.3% (43 of 806), a 4.5x spread; the blended rate is 3.6% (210 of 5,772)
How much the individual agent moves the outcome on identical campaigns and identical listsapps/airoofing/files/10k_agent_scorecards.md, the 11 per-agent scorecards, 2026-Q1

The floor behind the price

Three questions about what a seat actually costs, which is the fastest way to tell a priced service from a bid nobody can staff. A quote well under the published wage for the job is not a bargain; it is a different job, or a different country, or a shorter tenure than the ramp needs, and a buyer is entitled to know which.

  1. What are the people on the phone paid, and which job are you paying them for?

    Federal survey medians differ by occupation and the difference is the whole question. Telemarketers (SOC 41-9041) sit at a $17.04 hourly and $35,450 annual median across 58,430 jobs. An outbound appointment setter is not a customer service representative, and pricing one at the other job is the most common error in a build-versus-buy comparison.

    What they said

  2. Who supervises the seats working my account, and at what ratio?

    There is no published call center manager occupation to price this against. The closest federal series is first-line supervisors of office and administrative support workers (SOC 43-1011), at a $33.41 hourly and $69,500 annual median across 1,436,680 jobs -- a floor is not N seats, it is N seats plus somebody to run them, and that person costs more than any of them.

    What they said

  3. How many dials does the floor place, and over what period did you count?

    A volume claim without its window is not a rate. Ours is one measured day: 137,748 outbound call records on 2026-07-24, between 00:02 and 20:36 America/New_York, and the export was cut while dialling was still at full rate -- so it is a floor on that day rather than the day, it covers every campaign rather than one vertical, and it must never be multiplied out.

    What they said

Read this before you use the section

The wage lines are federal survey medians for named occupation codes and are a floor on labour alone: no loaded-cost multiplier, no supervision ratio and no margin is sourced here, and none may be attributed to the survey. The volume line is one measured day on one floor, published with its own truncation so nobody multiplies it.

$17.04 per hour / $35,450 per year (median), across 58,430 jobs
US median wage for telemarketers (SOC 41-9041)US Bureau of Labor Statistics, Occupational Employment and Wage Statistics, https://www.bls.gov/oes/current/oes419041.htm, 2025
$33.41 per hour / $69,500 per year (median), across 1,436,680 jobs
US median wage for first-line supervisors of office and administrative support workers (SOC 43-1011)US Bureau of Labor Statistics, Occupational Employment and Wage Statistics, https://www.bls.gov/oes/current/oes431011.htm, 2025
137,748 outbound call records on 2026-07-24, between 00:02 and 20:36 America/New_York
Outbound call volume across the ccdocs floor in one measured dayVICIdial call-log export at /opt/workspace/amd-corpora/data/vlog_today.tsv on dev-9555p -- 137,748 data rows, columns uniqueid/status/lead_id/start_epoch/end_epoch/length_in_sec/campaign_id/phone_number; re-derivable from vicidial_log for call_date 2026-07-24, 2026-Q3

What this sheet cannot do for you

It cannot tell you whether a floor is good. It can tell you whether a floor is measured, which is a different and much cheaper thing to establish, and the two are correlated only loosely: a shop with immaculate answers may still put the wrong people on your list, and a shop that fumbles the arithmetic questions may be excellent at the work. What the sheet reliably separates is the vendors who know their own numbers from the vendors who have never had to produce them.

It is also silent about the half of the job that happens after the appointment is set. Nothing on it asks about a homeowner being there when the truck arrives, because that outcome sits on your side of the handoff and no vendor answer settles it. If you have room for one metric outside this sheet, make it the share of set appointments that turned into an inspection that happened, and agree how it will be measured before anybody dials.

And it is a snapshot. Rules move, wage surveys are re-released, and our own figures come from windows that nothing refreshes. Every line carries the document or the corpus it came from, and the date, so you can go and check whether it still holds rather than trusting a sheet somebody handed you.

Want us to answer these on the record?

Send the sheet back with the questions you care about and we will answer them in writing, including the ones where our own number is the unflattering one.

Give us either an email or a phone number.

If you are still sizing the problem rather than choosing between quotes, the call center cost estimator and the speed-to-lead curve answer the arithmetic that comes before vendor selection.

Ask us these questions first

We wrote the sheet, so it is only fair that we go first. Bring it to a call and we will work down it in order, including the lines where the honest answer is that the measurement does not exist yet.

Book a Free Consultation
Call Now Book a Call