Chatbot Channel Mix

How to Split Conversations Across Channels, Priced on a 2,000-Conversation Month

Quick answer: A channel mix is the split of your chatbot's conversations across the channels you run, plus the rules that decide which channel a given conversation travels on. Two meters price every route. The platform meters its own unit: Manychat's pricing page counts "Active Contacts per month" ("One person = one Active Contact for the month"), Tidio's counts "Billable conversations" ("Live conversations and tickets handled by your team"), SendPulse's counts subscribers. The network meters messages, but only when the business speaks first: Meta's developer page says "You are only charged when a template message is delivered" and "All non-template messages are free," and its USD rate card effective 1 July 2026 prices a North American utility template at $0.0034 and a marketing template at $0.0250, while a second Meta page, updated 17 June 2026, says WhatsApp does not currently deliver marketing template messages to US phone numbers at all; Twilio's US page prices an SMS segment at $0.0083 plus a carrier fee of $0.0035 to $0.0045; Messenger, Instagram and Telegram charge nothing in ordinary use.

On the 2,000-conversation month worked below, the network line is $4.76, the platform line is a subscription, and the handoffs to a human are $600. So the mix is a containment decision with a messaging surcharge. The guide is organized that way: who speaks first, which route each kind of message should take, the worked month, the rebalance, and how four reviewed platforms meter a mix.

This guide owns the operating question. Which channel to launch on first, and when a second one is warranted, is the channel selection guide; what a channel strategy is and why fewer channels is often better is the channel strategy entry; what it takes for channels to share context is the omnichannel chatbot entry; the all-in number this guide optimizes is the same-day cost per conversation entry.

Two meters, and who speaks first decides which one runs

Every conversation on every channel is billed by the platform you build on, in the platform's unit. That meter runs whether the customer or the business spoke first, and it is the same for a web widget conversation as for a WhatsApp one on the platforms that carry both. The second meter belongs to the channel's owner or carrier, and it runs only in one direction.

The table's left column is nearly empty and its middle column is not, and that asymmetry is the whole economics of a channel mix. A conversation the customer opens costs you the platform unit and nothing else on any channel. A conversation you open costs a template on WhatsApp, a segment on SMS, and on Messenger is not possible outside the window except through the routes Meta permits, which are few and paid. So the first sorting rule is not "which channel is cheapest" but "who is speaking first," and the chatbot funnel and retention playbook both turn on that distinction: the playbook's first three moments are replies and cost nothing on the network; its last two are business-first and cost a template.

The US order for business-first messages, from Meta's pages and Twilio's: for a transactional message, a WhatsApp utility template ($0.0034) is the cheapest and an SMS segment landed with its carrier fee ($0.0118 to $0.0128) is next; for a promotional message, WhatsApp is not a route at all, because Meta's per-user limits page states (in the English wording of our 12 September read) that WhatsApp does not currently deliver marketing template messages to users with US phone numbers, so the choice is SMS at the segment price or a paid Messenger Marketing Message to a list that opted in (a product our retention playbook records as in beta in 20 markets including the US). That surprises people who assume WhatsApp carries everything: in the US it carries service and utility, and promotion travels elsewhere. Outside the US the marketing template exists and its price is the comparison: Meta's card prints $0.0250 on the row it labels North America (the United States and, as we read the card, Canada, so in practice a Canadian rate), $0.0625 for Brazil, $0.1365 for Germany and $0.0118 for India, against SMS rates that also vary by country. Our RCS versus WhatsApp guide covers the carrier-priced alternative to both.

Six routing rules

These are ours, drawn from the vendor rules above and from the patterns in our fifteen reviews; none is a vendor's recommendation.

1. Reply where the customer wrote. A customer who messaged on Instagram gets the answer on Instagram, inside the 24-hour window, at no network cost. Moving them to another channel to reply ("please continue on WhatsApp") turns a free reply into a business-first message on a channel where it may cost a template, and loses the context the customer already gave. The unified inbox entry covers the tooling that lets one team do this across channels.

2. Send business-first messages on the cheapest permitted route the customer has opted into. Permitted comes first: a WhatsApp template needs Meta's category approval and the customer's opt-in, and a marketing template is not delivered to US numbers at all; an SMS needs consent and, in the US, the A2P 10DLC registration Twilio's page refers to; a Messenger message after 24 hours needs one of the routes Meta permits outside the window. Among the permitted routes, the order above applies: utility template, then SMS, then, where it is delivered, marketing template. A shipping update to a customer who bought on WhatsApp is a $0.0034 utility template; the same update by SMS is a two-segment text at about $0.026 with the carrier fee; a "we miss you" message cannot be a utility template at all, and to a US number it goes by SMS at the segment price or as a Messenger Marketing Message, not on WhatsApp.

3. Spend the free windows deliberately. Meta's pricing page opens a 72-hour free entry point window, in which "All messages are free … including template messages," when a customer messages you from a Click-to-WhatsApp ad or a Facebook Page call-to-action button and "you respond within 24 hours using any type of message"; the Click-to-WhatsApp ads entry covers the entry points. Meta's Messenger policy page lists the user actions that open its 24-hour window, and a reaction to a message is one of them. A flow that asks one useful question inside a window, so the customer replies, keeps the window open and the network meter at zero; a flow that says its piece and stops closes it.

4. Give each channel a job, not a copy of the others. The web widget handles the on-site question before the sale; WhatsApp carries the post-purchase utility stream because that is where the cheapest business-first route lives; Instagram handles comment-to-DM and story replies; Telegram serves the audience that is there, at no network cost, with the Telegram bot entry's caveats on reach; SMS is the route for a time-critical message to a customer with no WhatsApp opt-in, because it reaches every phone. The channel selection guide makes the case for designing each channel to its own constraints rather than mirroring one flow across all of them.

5. Keep one human exit per channel and count it per channel. The handoff is where a conversation goes from cents to dollars, and the count of handoffs by channel is the single most useful number in the rebalance below. The human handoff entry and the handoff design guide cover the design; the rule here is only that the exit exists on every channel and is tagged with the channel it came from.

6. Cap business-first frequency per person, per channel. A business-first message that is ignored still costs its rate, and a customer who blocks the number costs every future utility message on that channel. Meta enforces per-user marketing limits of its own on WhatsApp, which the retention playbook records; the broadcast campaign guide has the sending discipline. A two-attempt stop rule per message purpose is our default.

The same 2,000-conversation month on four channels

Assumptions, stated once: a US online shop, 2,000 conversations in the month, US recipients throughout (Meta's "North America" row), a Twilio long code for SMS with T-Mobile and Verizon's $0.0045 outbound and T-Mobile's $0.0025 inbound carrier fee used as the working figures, 15 percent of conversations handed to a human, six agent minutes each at $20 an hour fully loaded. Rates as read the night of 20–21 September 2026.

Three readings of that table, all ours.

The network line is $4.76 for 2,000 conversations, about a quarter of a cent each. SMS carries 5 percent of the conversations and 89 percent of the network bill, which makes it look like the thing to fix, and in the US it cannot be fixed by moving to WhatsApp, because those 100 texts are promotional and WhatsApp marketing templates are not delivered to US numbers; if they were transactional they would qualify as utility templates at $0.34 for the hundred and the SMS line, number included, would go. The most that move could save is $3.91 a month.

The platform line is a subscription, and the mix decides which subscriptions are possible. Manychat's pricing page lists seven channels (Instagram, TikTok, Facebook Messenger, WhatsApp, Telegram, SMS and Email), sells three of them on Pro and all of them on Business, and a website widget is not among the channels it lists; this month's 1,000 web conversations would need a second tool. Tidio's page checks Messenger, Instagram and WhatsApp integrations on every plan alongside its live chat, and SMS is not among the integration rows we read; its unit counts only the conversations a human handled, so this month's 300 handoffs are 300 billable conversations, above the Growth plan's floor of "From 250 Billable conversations" at $49.17 a month on annual billing; the quota selector's next stop is 400, and we did not record that stop's price, so $49.17 is a lower bound for this month, not its price. SendPulse's page prices Pro by subscribers with "Unlimited chatbots and live chats," so the mix does not change the price and the audience size does. None of the three meters the same thing, and the cost per conversation entry has the conversion.

The human line is $600: 300 handoffs, 1,800 agent minutes, 30 hours at $20. That is more than a hundred times the network line and, on the Tidio path, at least twelve times the Growth floor of $49.17 (the 400-conversation stop this month needs will cost more). Raise containment by five points, from 85 to 90 percent, and 100 handoffs disappear: $200 a month, or forty times the entire messaging bill. The containment rate entry and the resolution rate guide are where that lever lives; on this month, a channel mix that moves conversations toward the channel where the bot contains best is worth more than any messaging arbitrage, and the rebalance below is built on that.

When does the network line matter? At volume and abroad. Fifty thousand marketing templates a month is $3,125 to Brazilian numbers on Meta's card and $6,825 to German ones; a business sending that many is choosing between a WhatsApp marketing template and an SMS segment on cost alone and should. Below roughly ten thousand business-first messages a month, our reading is that the network line is a rounding error next to handoffs and the platform fee.

The monthly rebalance

Four numbers per channel, once a month: conversations held, containment (the share the bot finished), handoffs (and therefore agent minutes), and the network cost from the table above. From those, two derived figures: all-in cost per conversation by channel (that channel's network cost, plus its share of the platform fee by conversation count, plus its agent minutes at your rate, divided by its conversations), and the business-first share of its conversations.

Then three moves, in order. First, if one channel's containment is materially lower than the others' on the same kinds of questions, fix the flow there before touching the mix; a channel that hands off twice as often is twice as expensive whatever its messaging rate, and the fallback rate guide is usually the place to start. Second, if a business-first message type is running on a costlier route than a permitted cheaper one (marketing template where utility qualifies, SMS where the customer has a WhatsApp opt-in and the message is transactional), move it. Third, if a channel's all-in cost per conversation is highest and its conversations are the kind another channel already handles, stop advertising the entry point rather than shutting the channel: customers who arrive there still get an answer, and the volume drifts to the cheaper route on its own. The chatbot metrics guide has the dashboard; the chatbot analytics entry covers per-channel reporting on the reviewed platforms.

How four reviewed platforms meter a mix

Manychat sells the mix by channel count and by people. The pricing page's Free tier connects "any 2 channels" from Instagram, TikTok, Messenger and Telegram, Pro connects "any 3" from all seven including WhatsApp, SMS and Email, and Business connects "Unlimited channels"; our Manychat review records the same two-channel, non-WhatsApp restriction on Essential and all seven channels on Advanced. The unit is the active contact, 2,500 on Pro and 7,500 on Business, with overage at $0.05 and $0.025 a contact on monthly billing. The same review records the monthly-billed prices ($39 Pro, $99 Business) against the annual-equivalent ones the page prints ($29, $69), and dates the restriction of Free and Essential to non-WhatsApp channels to March 2026. A WhatsApp-plus-SMS-plus-Instagram mix needs Pro; adding a fourth channel needs Business.

Tidio sells the mix by work. Its unit ignores the channel and counts "Live conversations and tickets handled by your team," so a mix that contains well is cheap on Tidio regardless of how many channels feed it, and one that hands off often is not. Messenger, Instagram and WhatsApp integrations are on every plan column of the pricing page we read. Our Tidio review, last updated 29 May 2026, recorded the Plus tier at $749 a month; the page we read prints Plus at "Starts at $300/mo … + monthly usage," which is a change worth knowing about and one the review does not yet reflect.

SendPulse sells the mix by audience. Pro is priced on a subscriber slider ($12 a month monthly-billed at 500 subscribers, $9.60 on annual billing) with "Unlimited chatbots and live chats," and the Free plan carries "3 chatbots … Up to 500 subscribers … Up to 10,000 messages per month across all of your chatbots." Our SendPulse review exercised the free plan across Telegram, Messenger, WhatsApp and live chat on one account, the one free plan in our corpus that a review exercised across four channels on a single account.

Wati is a WhatsApp-first platform and says so. Our Wati review records a markup of roughly 20 percent on Meta's per-message rates (its worked line: 3,000 marketing messages at $0.025 times 1.20) and names "Multi-channel-first operators where WhatsApp is one of several roughly equal channels" as a weak fit, because its adjacent-channel coverage is a shared inbox rather than native automation. For a mix where WhatsApp carries most of the business-first volume, the markup is the number to model; for a balanced mix, the review's own verdict points elsewhere.

What our fifteen reviews record

Searched 21 September 2026, case-insensitively, across the fifteen files matched by sample-reviews/*-review.md: channel mix appears in 3 (Blip, Landbot, Wati), in each case in passing, as a variable in a bespoke quote or a plan comparison rather than as a subject; cost per conversation appears in 0. No review protocol ran the same conversations across two channels on one platform and compared the all-in cost, and none read a network invoice; the channel facts in the reviews are pricing-page and configuration-surface reads, which the Manychat review records screen by screen. The protocol has not yet priced a channel against a channel; that is on us, not on any vendor. The single cheapest experiment that would close it is one platform, two channels, fifty identical business-first messages each, and the platform's own billing export at month end. Corrections to editorial@chatbotscape.com.

FAQ

What is a chatbot channel mix?

The split of a chatbot's conversations across the channels it runs on (website widget, WhatsApp, Instagram, Messenger, Telegram, SMS), together with the rules that decide which channel a particular conversation or message travels on. Choosing which channels to run is channel strategy; the mix is how you operate them once they are live.

Is WhatsApp or SMS cheaper for business-initiated messages?

To a US number, it depends on what the message is. A transactional update is cheapest as a WhatsApp utility template, $0.0034 on Meta's USD rate card effective 1 July 2026, against an SMS segment at $0.0083 plus a carrier fee of $0.0035 to $0.0045 on Twilio's US page. A promotional message cannot travel on WhatsApp to a US number at present, because Meta does not deliver marketing templates there, so it goes by SMS or as a Messenger Marketing Message. Where marketing templates are delivered, their price varies by country ($0.0250 on the North America row, $0.0625 for Brazil, $0.1365 for Germany) and so does the SMS rate.

Do Messenger and Instagram charge per message?

Not in ordinary use. Meta's policy page gives businesses 24 hours to respond to a user and allows promotional content in that window; outside it, business-first messages need one of the routes Meta permits, which the page lists as message tags for approved update types and sponsored messages, which are ads, and which have been changing through 2026; on Instagram, our retention playbook records, only a human agent may write outside the window, for up to 7 days. There is no per-message rate on the page.

Which channel should carry post-purchase messages?

WhatsApp, where the customer has opted in, because order and shipping updates qualify as utility templates at $0.0034 to US numbers and are free inside an open service window; SMS where they have not, at the segment rate; and Messenger only if the customer wrote within 24 hours or the update fits a route Meta permits outside the window. That is rule two above, applied to the highest-volume business-first stream most shops have.

How often should the mix be rebalanced?

Monthly, with four numbers per channel: conversations, containment, handoffs and network cost. Fix a low-containment channel's flow first, move business-first messages to a cheaper permitted route second, and only then reduce a channel's entry points.

Related guides

• Cost per conversation — the four meters and the formula this guide optimizes, with the per-channel arithmetic.
• How to choose chatbot channels — the decision upstream of this one: which channel first, and when a second is warranted.
• Channel strategy — the definition, and the case that fewer channels is often better.
• Omnichannel chatbot — what it takes for channels to share context, and when you need it.
• Chatbot retention playbook — the five moments, split by who spoke last, that this guide's first rule comes from.
• WhatsApp broadcast campaign guide — sending discipline for the business-first messages that carry the network cost.
• How to reduce chatbot costs — the levers on the platform and AI lines, which this guide treats as given.
• WhatsApp chatbots: the complete channel guide — platform choice, pricing structure and compliance for the channel that carries most business-first volume.

Sources

• Meta for Developers, Pricing on the WhatsApp Business Platform (developers.facebook.com/documentation/business-messaging/whatsapp/pricing), read the night of 20–21 September 2026: "You are only charged when a template message is delivered"; "All non-template messages are free"; "Utility templates delivered within an open customer service window are free"; "All messages are free for 72 hours, including template messages, if sent within an open free entry point window"; the free entry point rule ("If you respond within 24 hours using any type of message, the message will be free, and a Free Entry Point ('FEP') window will be opened"); "Meta may update pricing only on the 1st day of each quarter"; "Rate card updates effective October 1, 2026."
• Meta, Cost per message in USD on the WhatsApp Business Platform, effective July 1, 2026 (the CSV linked from the page above as "USD rates"), read 21 September 2026: rows "North America, USD, 0.0250, 0.0034, 0.0034"; "Brazil, USD, 0.0625, 0.0068, 0.0068"; "Germany, USD, 0.1365, 0.0550, 0.0550"; "India, USD, 0.0118, 0.0014, 0.0014" (columns Marketing, Utility, Authentication).
• Meta for Developers, WhatsApp Business Platform, Per-user marketing template message limits (developers.facebook.com/documentation/business-messaging/whatsapp/templates/marketing-templates/per-user-limits, page dated "Updated: 17 June 2026"), read 21 September 2026: WhatsApp "does not currently deliver marketing template messages to WhatsApp users with United States phone numbers," and an attempt returns an error. (Our reader received the page in Russian; the quotation follows the English wording our retention playbook recorded from the same page on 12 September 2026.)
• Meta for Developers, Messenger Platform and IG Messaging API policy (developers.facebook.com/documentation/business-messaging/messenger-platform/policy), read the night of 20–21 September 2026: "Businesses have up to 24 hours to respond to a user. Messages sent within the 24-hour window may contain promotional content"; the list of user actions that open the window, including "User reacts to a message"; message tags "for a set of approved use cases"; sponsored messages "can be sent outside the 24-hour standard messaging window."
• Twilio, SMS Pricing in United States for Text Messaging (twilio.com/en-us/sms/pricing/us), read the night of 20–21 September 2026: long codes "$0.0083" outbound and inbound SMS; "Text messages are charged per segment"; SMS carrier fees, outbound AT&T "$0.0035," T-Mobile "$0.0045," Verizon "$0.0045," inbound AT&T "$0.0035," T-Mobile "$0.0025"; long code number "$1.15" a month; the page's references to A2P 10DLC registration.
• Telegram, Bots FAQ (core.telegram.org/bots/faq), read the night of 20–21 September 2026: "By default, bots are able to message their users at no cost"; "not able to broadcast more than about 30 messages per second, unless they enable paid broadcasts"; "0.1 Stars per message"; "at least 100,000 Stars on its balance and at least 100,000 monthly active users."
• Manychat, Pricing (manychat.com/pricing), read the night of 20–21 September 2026: Free "Connect any 2 channels and choose from: Instagram, TikTok, Facebook Messenger or Telegram"; Pro "$29/mo … $348 billed annually (25% off) … 2500 Active Contacts per month … Connect any 3 channels, select from: Instagram, TikTok, Facebook Messenger, Whatsapp, Telegram, SMS and Email"; Business "$69/mo … $828 billed annually … 7500 … Unlimited channels"; overage per additional Active Contact (monthly) Pro "$0.05," Business "$0.025"; "One person = one Active Contact for the month."
• Tidio, Pricing (tidio.com/pricing), read the night of 20–21 September 2026, annual toggle: Starter "$24.17/mo … 100 Billable conversations"; Growth "Starts at $49.17/mo … From 250 Billable conversations"; Plus "Starts at $300/mo … + monthly usage"; "Live conversations and tickets handled by your team make up your monthly billable conversations limit"; quota stops 50 through 2,000; Messenger, Instagram and WhatsApp integration rows checked on every plan column.
• SendPulse, Chatbot pricing (sendpulse.com/pricing/messengers), read the night of 20–21 September 2026: Free "3 chatbots … Up to 500 subscribers … Up to 10,000 messages per month across all of your chatbots"; Pro "$ 9.60 /month" yearly and "$ 12 /month" at 500 subscribers, "Unlimited chatbots and live chats."
• Chatbotscape review corpus (the fifteen platform reviews at /reviews), searched 21 September 2026 from the repository root. Denominator: ls sample-reviews/*-review.md | wc -l returns 15. grep -liE 'channel mix' sample-reviews/*-review.md returns 3 (blip, landbot, wati); grep -liE 'cost per conversation' sample-reviews/*-review.md returns 0. Passages cited: manychat-review.md lines 351–359 (the two-price tier table and the channel restriction); wati-review.md lines 148 and 278 (the weak-fit line and the "× 1.20 markup" line); tidio-review.md line 313 (Plus at $749, last updated 29 May 2026); sendpulse-review.md line 48 (the free-plan channel note).
• Chatbotscape evaluation methodology. /methodology (continuously updated).

About this guide

Chatbotscape launched in 2026 as an independent review site for chatbot platforms. This guide is part of our SMB chatbot Academy and is written for the owner or operations lead of a small business that already runs a chatbot on more than one channel and wants to know which conversations belong where and what each route costs. It reads Meta's, Twilio's, Telegram's, Manychat's, Tidio's and SendPulse's published pricing and policy pages as of the night of 20–21 September 2026, prices one stated month on them, and draws on our fifteen platform reviews for how each platform meters a mix.

Methodology

Every rate, limit and quotation was read on the vendor page named in Sources and is quoted with the page's own wording; Meta's per-message figures are from the USD rate card CSV its pricing page links, headed "effective July 1, 2026." The two-meter framing, the six routing rules, the worked month and the rebalance procedure are editorial working recommendations derived from those documented rules and from the deployment patterns recorded across our reviews; they are not the output of a controlled study, and no message on this page was sent or billed by us. The corpus counts are reproducible with the commands printed in Sources. Manychat, Tidio, SendPulse and Wati carry affiliate links elsewhere on this site; the arithmetic here does not depend on which platform a reader chooses.

Last updated

21 September 2026 — first published.