Short answer: Price the completed interview, not the attempt. A WhatsApp survey at the utility rate is cheapest per response — but Meta’s rules say a general opinion survey will not be approved as utility, which moves it to a rate 7.52 times higher. IVR is cheap per dial, expensive per completion. CATI costs most and reaches people the others miss.
The three methods, and what you are actually buying
An IVR survey is a recorded call collecting keypress or spoken answers with no human on the line. CATI — computer-assisted telephone interviewing — is a person with a script dialling a sample. A WhatsApp survey is a template message with reply buttons or a link, sent to numbers that have opted in.
The three are priced on completely different units. IVR and CATI are billed by connected call time, in pulses; WhatsApp is billed per delivered message; and CATI’s dominant cost is not telecom at all, but interviewer hours, which nobody publishes a rate card for. That is why per-unit comparisons mislead — a pulse, a message and an interviewer-minute are not comparable quantities. The only comparable unit is one completed interview, and reaching it means knowing both the per-unit rate and the share of contacts who finish.
Layer one: what each channel charges
Published rates, each read on 12 October 2026 from the publisher’s own page:
| Channel and unit | Published rate | Billing unit | Source read 12 Oct 2026 |
|---|---|---|---|
| Outbound domestic voice (IVR or CATI leg) | ₹0.38 per minute | “Pulse is 30 seconds” | Plivo India voice pricing |
| Outbound voice, alternative provider | $0.0496 per minute to mobile | Increment not stated on the page | Twilio India voice pricing |
| WhatsApp utility template | ₹0.145 per delivered message | Per delivered message | AiSensy published India card, effective 1 Oct 2026 |
| WhatsApp marketing template | ₹1.09 per delivered message | Per delivered message | AiSensy published India card, effective 1 Oct 2026 |
| CATI interviewer time | No published rate card exists | Per interviewer-hour | — |
Two gaps in that table are worth stating rather than papering over. Meta’s own India rate card did not verify today. Its pricing calculator loads rates from an endpoint that returned HTTP 502 on every attempt this morning, for both marketing and utility, with the market and currency correctly set to India and INR. So no Meta list rate appears in this post, and nothing was carried over from an earlier one. The WhatsApp figures here are a provider’s published rates, which sit above Meta’s card — our pillar on WhatsApp pricing in India covers that gap.
The second gap is CATI. No Indian provider publishes an interviewer rate, and the Chief Labour Commissioner’s minimum-wage page is closed to automated fetching, so we could not read a primary wage figure either. Every CATI figure below is therefore set against a loaded cost per interviewer-hour that you supply; the ₹150, ₹250 and ₹400 shown are illustrations, not claims about the market.
The rule that decides your WhatsApp bill
This is the finding that matters most, and almost nobody mentions it. Meta’s template categorisation documentation, updated 15 September 2026 and read today, says of feedback-survey templates:
Specificity of the order or interaction to which these relate is necessary. A general/generic survey or request for feedback will not be approved as utility.
The same page classes templates with mixed content as marketing, giving as its own example “a feedback survey with promotional content”.
So WhatsApp has two prices for the word “survey”. Ask a customer about the order they received last Tuesday and you have a utility template at ₹0.145. Ask a general-population sample what they think about anything and you are running a marketing template at ₹1.09 — a ratio of 7.52, a 651.72% premium, for the same number of messages. Market research, brand tracking, political polling and employee surveys all sit on the expensive side of that line; transactional feedback sits on the cheap side.
Layer five: the completion gap
Layers one to four give a cost per message or per minute. What converts that into a cost per completed interview is the share of contacts who finish — and that is where the ordering changes.
We publish no completion-rate figure of our own, because we have no verified India-wide source for one. What exists is published research, worth citing precisely. Lau, Cronberg, Marks and Amaya, in Survey Research Methods in December 2019, ran a four-mode comparison in Nigeria and report: “FTF had the highest response rate (99%), followed by CATI (15%), IVR (3%) and SMS (0.2%).” For a 3,000-complete survey they found “IVR is 43% the cost of CATI, and SMS is 24% the cost of CATI”.
That is Nigeria in 2019, not India in 2026, and should not be read as an Indian figure. For India there is Nagpal, Mathur, Biswas and Fraker in BMJ Global Health, August 2021, on CATI across nine Indian states. Their pilot without a structured protocol achieved a 30% response rate; with randomised call-back slots and up to seven attempts it reached 70%, against 55% for a morning-and-evening-only protocol. Completion also tracked income: 59.3% in the top income quartile against 49.9% in the bottom.
The lesson in that second paper is not a number at all: call protocol moved the response rate by 15 percentage points on the same sample. A completion rate is managed, not inherited.
What a completed interview actually costs
Four things follow.
- WhatsApp at the utility rate is cheapest at every completion rate, by a wide margin.
- WhatsApp at the marketing rate costs more per completed interview than IVR at the same completion rate — the category, not the channel, is doing the work.
- IVR’s cheap pulse does not survive a low completion rate. At the 3% response rate published for Nigeria it costs ₹7.28 per completed interview; at 20% it costs ₹1.90.
- CATI stays an order of magnitude above everything else throughout, and no plausible wage assumption changes that.
A worked example: 1,000 completed interviews
One thousand completions, Indian mobile numbers, at the published rates above. The CBIC’s note on telecommunication services puts such services at 18% GST, but your provider’s classification and your own invoice govern what you are charged.
| Method | Contacts needed | Ex-GST | Inc. 18% GST | Per completed interview |
|---|---|---|---|---|
| WhatsApp utility, 20% completion | 5,000 delivered | ₹725.00 | ₹855.50 | ₹0.73 |
| WhatsApp utility, 10% completion | 10,000 delivered | ₹1,450.00 | ₹1,711.00 | ₹1.45 |
| WhatsApp marketing, 10% completion | 10,000 delivered | ₹10,900.00 | ₹12,862.00 | ₹10.90 |
| IVR at a 3% completion rate | 33,333 answered calls | ₹7,283.33 | ₹8,594.33 | ₹7.28 |
| CATI at 15%, ₹150/interviewer-hour | 6,667 answered calls | ₹39,606.67 | ₹46,735.87 | ₹39.61 |
| CATI at 15%, ₹250/interviewer-hour | 6,667 answered calls | ₹63,773.33 | ₹75,252.53 | ₹63.77 |
| CATI at 15%, ₹400/interviewer-hour | 6,667 answered calls | ₹1,00,023.33 | ₹1,18,027.53 | ₹100.02 |
The IVR line assumes a completed interview runs three minutes (six pulses) and an abandoned call thirty seconds. The CATI lines assume six interviewer-minutes on a completion and ninety seconds on an unproductive attempt — 14.5 interviewer-minutes per completion, or about 242 hours for the full thousand. Those are campaign parameters, not published figures: change them and the arithmetic changes, but the ordering does not.
At ₹250 an hour, CATI costs about 44 times what WhatsApp utility costs per completed interview at a 10% completion rate. The real decision is therefore not a pricing one — it is whether the cheap methods can reach your sample at all. Our IVR survey cost breakdown and bulk voice call pricing work the voice side through.
The compliance layer
Survey calls are commercial communication, and the timing rules are less settled than they are usually made to sound. A 9:00 AM to 9:00 PM IST window is widely repeated, and this post cited it on 12 October 2026 on the strength of Exotel’s developer documentation. That citation no longer holds. Re-read on 14 October 2026, the page states no such window — only a four-hour call time limit, a DND lookup field and a transactional call type. We have found no primary TRAI instrument setting a sender-side calling window, so we no longer state one.
What the regulations do set out is a recipient control. Schedule-II of the TCCCPR provides customer preference options by time band and day type, under which transactional and service communications are not blocked. That is the customer choosing when they will accept commercial calls — not a blanket window on senders. The distinction matters for a survey campaign, because it decides whether your calling hours are a legal limit or a reachability problem. Confirm your own position with your access provider in writing.
TRAI’s own regulation PDF would not fetch today, so treat that as a provider’s reading of the rules rather than the rules themselves, and check it before building a dialler schedule around it. The broader headers-and-consent regime is in our guide to the 140 and 1600 series. The consequence for method choice is that a nine-to-nine window compresses IVR and CATI into twelve hours a day — the hours in which working respondents are least available, which is what the BMJ paper’s call-back slots are addressing. WhatsApp has no equivalent published clock.
Which method when
| If you need… | Use | Why |
|---|---|---|
| Feedback on a specific order or interaction | WhatsApp utility template | Cheapest per completion by an order of magnitude, and Meta’s rules allow it as utility |
| A general or opinion survey, at volume | IVR | A general survey cannot be a utility template, and IVR’s per-pulse rate undercuts the marketing rate |
| Open-ended answers, or a respondent who needs help | CATI | Nothing else handles probing, clarification or a confused respondent |
| Feature-phone or low-literacy respondents | IVR or CATI | No smartphone or app required |
| Reach into lower-income groups | CATI, with a call-back protocol | The one published India study we cite found completion falls with income, and that protocol recovers much of it |
| A brand-tracking study on WhatsApp | Reconsider | It is a marketing template at 7.52 times the utility rate |
Where we stand
XReach51 sells WhatsApp, SMS and voice messaging in India, so AiSensy, Plivo, Twilio and Exotel are all competitors or substitutes for parts of what we do. We publish no XReach51 rates here. Credit where it is due: Plivo and Twilio publish per-minute India voice rates openly and without a sales call, which most Indian voice providers do not; AiSensy publishes a complete India card by category with an effective date; and Exotel’s calling-hours documentation is clearer than anything we found from a competitor. We used their pages because they were there to be read. Our plans are on the pricing page, and our channel cost comparison sets the channels against each other on reach.
FAQ
Is an IVR survey always cheaper than CATI?
Per completed interview, on every assumption we ran, yes — by roughly an order of magnitude. Per usable interview depends on your questionnaire: IVR cannot probe an unclear answer, and a 2013 India study by Chakraborty and colleagues measured a 4.0% error rate across 500 IVR questions and found IVR took 14 minutes 12 seconds against 6 minutes 3 seconds for the same survey with a live operator.
Can I run a market research survey on WhatsApp at the utility rate?
Meta’s documentation says no. A general or generic survey “will not be approved as utility”. Transaction-specific feedback is the only survey use case Meta lists as utility.
What completion rate should I assume?
We will not give you one, because we have no verified Indian figure to give. Run a pilot of a few hundred contacts and measure it. The published research we cite shows the number moves by tens of percentage points with call protocol alone.
Do undelivered WhatsApp messages cost anything?
No. Meta charges on delivery, so an undelivered template is free. That means your cost per reached contact is exactly the list rate whatever your delivery rate — the delivery gap moves your total bill, not your cost per reach. Completion, not delivery, is the lever on WhatsApp.
Does the nine-to-nine window apply to WhatsApp surveys too?
The calling-hours documentation we read today is about voice calls and SMS. We found no equivalent published clock for WhatsApp template messages, and we are not going to infer one.
Which single number should I track?
Cost per completed interview, computed after the fact from your own invoice and completion count. Every per-unit rate here is an input to that number, not a substitute for it.
How we verified this
Every figure here was read from a primary source on 12 October 2026. We date every price because prices move, and we do not estimate: where a number could not be verified on the day, we say so rather than filling the gap. All arithmetic was run rather than eyeballed.
What we could not verify today. Meta’s own India INR rate card — its pricing endpoint returned HTTP 502 throughout, so no Meta list rate appears above. A primary minimum-wage or interviewer-cost figure for India: none is published in a form we could read, so all CATI labour figures are illustrations at a cost you supply. TRAI’s own regulation text on calling hours, which would not fetch; the nine-to-nine window above is a provider’s statement of the rule. And any completion rate for Indian survey work, for which we cite published research from Nigeria in 2019 and India in 2021 rather than offering a figure of our own.
Sources, all read 12 October 2026
- Meta, Template categorization, developers.facebook.com (page updated 15 September 2026) — utility and marketing definitions, the feedback-survey specificity rule, mixed-content classification.
- Meta, Pricing on the WhatsApp Business Platform, developers.facebook.com — charging on delivery, volume tiers for utility and authentication.
- Plivo, India voice pricing — ₹0.38 per minute outbound domestic, “Pulse is 30 seconds”.
- Twilio, India voice pricing — $0.0496 per minute to mobile.
- AiSensy, pricing page — published India per-message card by category, effective 1 October 2026.
- Exotel, developer documentation on calling-hour restrictions — the 9:00 AM to 9:00 PM IST window and the list of time-restricted categories.
- Lau, Cronberg, Marks and Amaya, “In Search of the Optimal Mode for Mobile Phone Surveys in Developing Countries”, Survey Research Methods 13(3), 2019 — response rates and relative costs by mode, Nigeria.
- Nagpal, Mathur, Biswas and Fraker, “Who do phone surveys miss, and how to reduce exclusion”, BMJ Global Health, August 2021 — CATI response and completion rates across nine Indian states.
- Chakraborty, Medhi, Cutrell and Thies, “Man versus Machine: Evaluating IVR versus a Live Operator for Phone Surveys in India”, DEV’13 — IVR error rate and interview duration.
- CBIC, GST on telecom services — the 18% rate for telecommunication services.