An appointment setting answering service fills events and webinars by phone: callers invite a targeted list and register people on the call. Then they confirm and remind. Plan from attendees: Goldcast and Livestorm platform data put webinar attendance at 33% to 48% of registrations.
In This Article
Phone survey companies lead with a rate. Online survey tools lead with a rate too, and the two sides are not dividing by the same thing. An online tool divides starts or finished forms by views or invitations, and a phone program divides interviews by people reached or records dialed. Both figures can be true. Our market research outsourcing service states 25 to 48% contact conversion for phone programs that work a client’s contact list. Its email figure is 5 to 15%. The gap between those two denominators comes first, and the twelve terms below carry the rest of the post.
- Study: the whole project, from objective to report.
- Sample frame: everyone who could be asked.
- Sample: the frame cases a study attempts.
- Completes (n): finished interviews or finished forms.
- Instrument: the questionnaire, in whichever mode it runs.
- Fielding: the days the instrument is live.
- Response rate: completes divided by every eligible case in the sample, reached or not, under the AAPOR Standard Definitions.
- Contact rate: the share of sampled cases with someone reached, counting every case the study attempted, reached or not.
- Cooperation rate: completes divided by the people reached.
- Refusal rate: the share of sampled cases that declined, counted against every case the study attempted.
- Completion rate: in online tools, the share of people who started a form and finished it.
- Contact conversion rate: our label for its phone figure, with the base counts pending confirmation.
Each mode reports the rate its tools count. That makes the headline comparison half the story; the other half, who answered and what they said, decides whether a study carries a decision.
What response rate do phone survey companies get compared with online surveys?
Phone survey companies working a named B2B list publish 25 to 48% contact conversion for those programs. Online surveys posted a 9.98% median in 2025.
The online figure comes from Survicate’s 2025 benchmark report, which analyzed 4,332 surveys from 460 companies across the 2025 calendar year. A response is a start divided by a view, and link surveys are excluded because their audience selects itself. The B2B median is 8.18%. The middle half of those B2B surveys runs from 3.20% to 17.87%, and surveys of seven or more questions drop to a 6.87% median. Once a person starts, 75.20% finish. The median finish takes 42 seconds, and these are vendor figures drawn from one platform, so the post uses them as the online benchmark.
The exclusion carries the post’s argument. Survicate dropped link surveys because their audience selects itself and would push the median higher, which is the sender relationship at work. A list that opted in writes back. A cold frame does not, and the same instrument will post two different rates on the two frames before a word of it changes.
Phone has the same split. Pew Research Center’s 2025 reference survey gave households online, paper and a live-interviewer phone option, and 5,022 completed at a 29% AAPOR response rate. Every sampled address counts, reached or not. Only 342 of those completes came by phone, so the figure describes a public-opinion frame, not a phone program. A named B2B list is a different frame. The interviewer knows who to ask for, calls back on a schedule, and works the list to exhaustion rather than waiting for volunteers. The market research outsourcing page publishes 25 to 48% contact conversion for list programs, and one B2B construction study completed 137 interviews at 48.58%. Automated phone surveys are the third service. So a phone survey and IVR company’s rate depends on which service it is quoting.
Each row states its own denominator. The rates are not comparable across rows, and that incomparability is the point of the table.
Mode and sample | Published rate | What it is divided by |
|---|---|---|
Online survey, all formats | 9.98% median (middle half 3.75% to 21.69%) | Survey starts divided by views; link surveys excluded |
Online survey, B2B audience | 8.18% median (middle half 3.20% to 17.87%) | Same basis |
General public, mail invitation with online, paper and live-interviewer phone options | 29% (AAPOR Response Rate 1) | Completes divided by every sampled address, reached or not |
Live interviewer, named B2B list | 25 to 48%+; 48.58% in one 137-interview program | Labelled contact conversion; base counts pending |
Read the table left to right. A rate without its denominator is a number without a unit, and the next two sections take the two halves the rate leaves out.
Why is a low NPS survey response rate a sample problem, not a volume problem?
A low NPS survey response rate is a sample problem, because the people who skip a survey differ from the people who take it. Sending more invitations does not change who answers.
Pew’s 2017 report on low response rates is the benchmark study, and it compared respondents with non-respondents across random-digit-dial polls. The rate proved an unreliable guide to bias. Put plainly: more answers are not better answers.
For your study, the silence is the question. If satisfaction drives the decision to answer, promoters and detractors reply and the lukewarm middle deletes the email. Churn risk and expansion live in that middle. Take an illustrative list with a 9% response rate: that is not a 9% sample of your customers. It samples your customers with strong feelings. The score it produces averages the two ends, and nothing in the rate says how far those ends sit from the middle.
Pew’s published data shows the failure. The measures it tested held up, except those tied to the decision to participate, and civic engagement was the one that broke. People who volunteer also answer polls, so the respondents overstated volunteering and every measure tied to willingness drifted the same way. The correction still runs. Pew’s 2026 panel methodology weights its 9,750 respondents to a volunteerism benchmark from the Census Bureau’s Current Population Survey. The lesson transfers to B2B satisfaction. Satisfaction drives the willingness to answer in both directions, so the silent group is not a smaller copy of the vocal one.
Two safeguards follow from that finding. Work the list rather than the volunteers. A phone program reaches the people who never volunteered, because the interviewer calls back and reaches the lukewarm account on the third attempt. That call records a rating email never produced. Contact rate is effort and schedule. Cooperation rate is the ask, and a live person asking a named contact for twelve minutes gets a different answer than a subject line. Then ask any vendor, phone or online, for the disposition counts behind its rate before you accept it. The AAPOR Standard Definitions supply the vocabulary. They specify how to count the cases sampled, reached, refused and completed, and which of the six response-rate formulas a reported figure uses. A provider who produces that table measures the study, and one who can only produce the percentage is measuring the tool.
Two related subjects live elsewhere. Post 6 on NPS and CSAT by phone covers the follow-up once the lukewarm middle is on the line. Recurring tracking of those scores, quarter over quarter, belongs to the customer insight surveys service page rather than to this post.
Tell us your research scope
Sample size, instrument length and timeline make up the whole research scope at this stage. A short note on what you need to learn gets a plain yes or no on phone, online or both.
What does a phone survey company capture that an online form cannot?
A phone survey company captures two data sets from one conversation: the structured rating for the dashboard and the reason behind it. The interviewer asks the follow-up in the moment.
The rating is the easy part. A 1 to 5 scale reads the same whether it is clicked or spoken, and a phone program captures it into the same column. The difference starts at three out of five. On a form, an optional text box follows. The median online form in Survicate’s 2025 benchmark is finished in 42 seconds, text box included. On a call, a question follows. What would have changed the rating?
The strategy team reads the second data set. It does not come from a longer form; it comes from a person who heard an unexpected answer and knew the follow-up.
Interviewers also catch the instrument’s mistakes during fielding. A question that confuses three respondents in a row gets flagged on the first morning, not discovered in the open-ended column after the send. The person who probes the three also hears that question seven reads two ways, and flags it. Scripts are conversations, not forms.
That second data set has a cost. Each phone complete carries interviewer time a form does not, and that cost decides the cases where the trade goes the other way.
What do the TCPA rules allow a phone survey company to do?
The TCPA rules let a phone survey company hand-dial any number without the called party’s consent. Autodialing or playing a recording to a cell phone needs that party’s prior express consent.
The rule is 47 CFR 64.1200 on eCFR. Paragraph (a)(1) bars autodialed or prerecorded-voice calls to a cell phone unless the called party gave prior express consent. Paragraph (a)(2) adds a stricter tier. Telemarketing calls need prior express written consent. Paragraph (f)(13) defines telemarketing as a call made to encourage the purchase, rental or investment in property, goods or services. Pure survey calls are not telemarketing. The FCC said as much in its 2012 TCPA order, which named research and survey calls among those outside the written-consent requirement. The plain-English TCPA guide walks through the consent tiers for every other call type a program runs.
Put plainly: survey calls are not sales calls.
For your study, this sets four operating facts. A call recorded for quality review follows state consent law, which the call recording laws by state guide lays out for the two-party states. State telemarketing statutes come second. The state mini-TCPA reference guide lists twelve states with their own consent rules, calling windows and do-not-call lists. Wireless numbers are dialed by hand unless the provider holds consent for each one, so research floors keep a manual-dial queue. The registry comes fourth. The National Do Not Call Registry governs telephone solicitations as defined in paragraph (f)(15), and a pure survey call sits outside that definition.
The failure case is Caribbean Cruise Line. October 2011 through July 2012, consumers heard a recording inviting them to a 30-second research survey, then a cruise offer on pressing one. Calls ran 12 to 15 million daily. The Federal Trade Commission and ten states charged the company and its partners under the Telemarketing Sales Rule in 2015. The FTC’s 2015 release records a stipulated settlement in which Caribbean Cruise Line agreed to pay $500,000 against a $7.73 million judgment. The company said no wrongdoing was found. Its 2017 release closes the case with the remaining defendants barred from robocalling under a $1.35 million judgment, suspended on payment of $2,500. The exemption did not survive the pitch. In FCC 23-44, a May 2023 forfeiture order, the Commission fined two political operatives $5,134,500 for 1,141 prerecorded calls to wireless numbers without consent. The order applies whatever the call’s content.
Three safeguards follow from the rule. Keep research and sales in separate programs with separate scripts, as the Insights Association Code of Standards, updated September 2026, requires of its members. The Code also requires that researchers identify themselves and keep participation voluntary, so a survey program carries both obligations into every call. Hand-dial every wireless number, or hold documented consent. Name the sponsor at the start or the end of the call, and read the recording notice wherever the state requires one.
The rule’s reading is less settled now. The Supreme Court’s McLaughlin Chiropractic decision, issued 20 June 2025, held that district courts must independently determine whether the FCC’s interpretation is correct. The 2012 order is one such interpretation. The exemption therefore rests on the statute’s words in paragraph (f)(13), not on the order. The call has to be research in fact. The FCC’s own rules keep moving as well. FCC 26-67, published 26 October 2026 and effective 25 November 2026, lets a caller designate one exclusive, disclosed method for revoking consent. A revocation sent after an informational message can then be read as covering that category alone. Hand-dialed survey calls hold no consent to revoke. A program that autodials wireless numbers under consent does, and its opt-out handling follows that order from its effective date.
States add their own layer. Florida’s Telephone Solicitation Act defines a telephonic sales call as a call, text or voicemail to solicit a sale of consumer goods or services. A pure survey program’s calls sit outside that definition, on the same boundary the federal rule draws. A survey ending in an offer crosses both.
This section describes operating practice, not legal advice. A research program that touches wireless numbers, recordings or consent records should be reviewed by counsel who practices TCPA law.
When is an online form the right CSAT survey methodology?
A form is the right CSAT survey methodology when the population is large, the questions are short and closed, and the list opted in. Speed matters more than the reason.
Transactional CSAT is the clearest case. A support desk closing 2,000 tickets a month wants a one-click rating on each one, at a cost per response in cents. No phone program competes on that job. Survicate’s 2025 data shows the format working: two-to-three-question surveys posted a 15.97% median, the highest of any length in its dataset.
Online also wins on anonymity, consistency and reach. A form asks sensitive questions with no one listening, and every respondent sees the same wording in the same order. A form reaches shifts a dialer cannot.
The lines cross for small populations. Picture a list of 400 plant managers, a 12-minute instrument, and an open-ended section on why they chose a competitor. The online send yields a few dozen self-selected forms, and the phone program yields a reading of the list. Win/loss interviews sit on that side. So do competitive intelligence and executive-level satisfaction studies, and so does any study where the client’s next question will be why.
Mixed mode closes the gap. Send the form, then put interviewers on the non-responders, so the final sample is the list rather than the volunteers. A recurring program that tracks satisfaction quarter over quarter is a different service, and this post is about the one-time study.
How should you decide between phone survey companies and an online tool for your study?
Decide between phone survey companies and an online tool with four questions about the study itself. Who is listed, how long the instrument runs, what the open ends are worth, and what a missing segment costs.
- Who is on the list. Named senior professionals who ignore unsolicited email are a phone list, and a large opted-in customer base that writes back is a send.
- How long the instrument is. Under three closed questions, the form wins. Past ten minutes with skip logic and open ends, the interviewer holds attention that a scroll bar does not.
- What the open-ended answers are worth. If the next decision depends on why, budget for the person who can ask it; a trend line can be bought in volume.
- What a missing segment would cost. If the silent middle is the segment the study exists to measure, a self-selected sample answers the wrong question however large it grows.
Run the arithmetic on a worked example. A 400-contact list at the 2025 B2B online median starts 33 forms, and the 33 chose themselves. The phone side reads the list rather than its volunteers, and the completes follow two numbers: the contact rate and the cooperation rate. Neither mode fits every study. The phone program fits the study whose client will ask why, and the form fits the study whose client will ask how many. Post 2 on designing a survey for phone administration takes the instrument from that decision onward.
Frequently asked questions
Is a phone survey worth it over an online survey for a B2B product, and what response rate should I expect from each?
A phone survey is worth it for a B2B product when buyers are named professionals and the study needs the reason behind the answers. Expect two different rates. An online send to a B2B list starts near 8% of views, the 2025 benchmark median. A live phone program working the same named list reports 25 to 48% contact conversion in this publisher’s published programs.
Why is my NPS survey response rate so low, and does a higher response rate make the score more reliable?
An NPS survey response rate is low because the typical customer has no strong feeling to report and no reason to open the email. Promoters and detractors answer. The middle stays silent. A higher rate makes the score more reliable only when the added respondents resemble the silent group; more of the same volunteers does not.
How do phone survey companies reach the people who ignore email surveys, and is the sample more representative?
Phone survey companies reach the people who ignore email by working a named list to exhaustion. That means scheduled callbacks, a live person asking for a specific contact, and a short direct request for twelve minutes of time. The sample is more representative of the list. Interviewers choose who to pursue rather than waiting for volunteers, which a self-selected email sample cannot offer.
What is a typical CSAT survey response rate online, and should a phone survey and IVR company be expected to beat it?
Is a phone call inviting someone to a free webinar telemarketing under the TCPA?
A CSAT survey response rate online runs near a 10% median across formats in 2025 benchmark data, counting starts against views. B2B audiences sit nearer 8%. A live-interviewer phone program working a named list reports 25 to 48% contact conversion on this publisher’s service page. An IVR survey runs with no interviewer. So the answer depends on which service the phone survey and IVR company is quoting.
Figure out if phone research fits your study
A 20-minute conversation covers sample, instrument and timeline. It ends with a plain answer on whether phone, online or a mix fits what you need to learn.
