What WhatsApp automation actually costs a small store

WhatsApp automation is billed twice: your platform charges a subscription, and Meta charges you directly for each delivered template message. Almost every story of a shocking WhatsApp bill comes down to a merchant who budgeted for the first and did not know about the second.

The good news is that Meta's rules, since the move to per-message pricing on 1 July 2025, make a large share of a well-designed store's messages free. Not cheap — free. Understanding which ones is worth more than shopping for a cheaper subscription.

The two bills

	Platform subscription	Meta message charges
Who charges you	Your WhatsApp platform	Meta, directly
What for	The software: inbox, builder, broadcasts	Each delivered template message
Predictable?	Yes, fixed	No — scales with sending
Capped?	Yes	Only by your own spend controls in Meta

Two consequences follow. First, a platform's price page tells you less than half of what you will pay. Second, you should check whether your platform adds a markup on top of Meta's rate or passes it through — the difference compounds with volume, and it is a fair question to ask before signing up. On ConvoFly, message charges go from Meta to you directly and we add no per-message markup; our own billing is the subscription only.

Which messages are free

This is the part worth memorising. From Meta's pricing documentation:

Message	Inside an open 24-hour window	Outside it
Anything you type freely	Free	Cannot be sent
Utility template	Free	Charged
Authentication template	Free	Charged
Marketing template	Charged	Charged

Service conversations — anything inside the window — have been free and unlimited since 1 November 2024. And there is a longer window most merchants never use: someone who reaches you by tapping a Click-to-WhatsApp ad opens a 72-hour free entry point window, three days rather than one.

The design consequence

Since the window opens when the customer messages, calls, or taps a button, you can open it deliberately. A template that asks a question rather than making a statement converts one charged message into a day of free conversation.

What you cannot budget from a blog

Meta's rates vary by the recipient's country and by template category, and they change. Rates were revised on 1 July 2026 and further adjustments are scheduled for 1 October 2026. Publishing a rate table here would be wrong within months, so we do not — check Meta's live rate card, linked from the pricing page, for your own markets.

What we can tell you is the shape: marketing costs meaningfully more than utility in every market, and the gap is the whole argument for classifying honestly rather than optimistically. See utility versus marketing templates.

Where surprise bills actually come from

1. Marketing where utility would do. A discount code in an order confirmation turns a free message into a charged one — on every order, forever.
2. Broadcasts to a whole list rather than an engaged segment. You pay per delivered message whether or not anyone reads it.
3. Flows that send outside the window when they could send inside it. A review request sent three days after delivery is charged; the same request sent while the customer is mid-conversation is not.
4. Retrying blocked sends. A message blocked with error 131049 should not be retried for at least 24 hours — retrying sooner risks further blocks.
5. Prepaid credit that expires. Check whether unused balance carries over before you top up.

A cost checklist before you switch a flow on

1. Does this flow start because the customer messaged us? If yes, it is free.
2. If it starts from an order, is the template utility rather than marketing?
3. Could the first message be a question instead of a statement, to open the window?
4. Is there a delay long enough to push a later step outside the window? Can it be shortened?
5. Are we sending to everyone, or to people who engage?

The honest summary

A store that runs order confirmations, shipping updates and support inside open windows can automate most of its customer communication for close to nothing in message fees, paying only a subscription. A store that leads with marketing broadcasts will pay per message, per recipient, forever — and, in the United States, will find that marketing templates are not delivered at all.

Which of those you become is decided by how you design flows, not by which platform you buy.

Working out your own number

You cannot get a monthly cost from a blog, because Meta's rates vary by country and category and change on a schedule. What you can do is build the estimate yourself in about ten minutes, and it is worth doing before you commit to a platform.

1. List every flow you intend to run and mark each one utility, marketing or free-text.
2. For each, write down how many messages one customer receives per month. An order confirmation is one per order, not one per month.
3. Mark which land on a cold chat. Anything triggered by a store event or your own system does; anything triggered by a customer message does not.
4. Cross out every free-text message and every utility template sent inside an open window. Those cost nothing.
5. Look up the rate for your customers' countries and the remaining categories on Meta's live rate card, linked from the pricing page.
6. Multiply and add the subscription.

Most stores are surprised by how much gets crossed out at step four. That crossing-out is the entire argument for designing flows around the conversation window rather than around templates.

A worked comparison

Two stores, same order volume, same products. The only difference is flow design.

	Store A — announces	Store B — asks
Order placed	Utility template: "Your order is confirmed"	Utility template asking Confirm / Cancel
Does the customer reply?	Rarely — nothing to reply to	Usually — it is a question
Window after that	Closed	Open for 24 hours
Shipping update	Charged template	Free if inside the window
Delivery question from customer	Free — they messaged	Free
Upsell or reorder nudge	Marketing template, charged, capped, blocked to US numbers	Free text inside the window — no template, no cap

Store B pays for one template per order and gets a day of unrestricted conversation. Store A pays per message forever and never opens a window. Neither bought a different product.

Where the money actually goes

In practice one flow usually dominates the bill, and it is almost always a marketing one. Ranked by typical spend:

• Campaign broadcasts. Charged per delivered message to every recipient, read or not. Easily the largest line.
• Abandoned checkout recovery. Marketing by classification, so charged on every send, and undeliverable to US numbers.
• Order and shipping templates on cold chats. Cheaper, and often avoidable by getting a reply earlier.
• Everything inside a window. Zero.

Five habits that keep the bill small

1. Ask, do not announce. Covered above, and worth more than every other item combined.
2. Keep utility templates purely transactional. One discount code reclassifies the template as marketing, which charges it on every order forever. See utility versus marketing.
3. Send campaigns to engaged segments, not the whole list. You pay for delivery regardless of interest.
4. Do not retry a blocked send. Error 131049 means wait at least 24 hours; retrying sooner risks further blocks and spends money achieving nothing.
5. Shorten delays that push a step outside the window. A flow waiting 25 hours turns a free message into a charged one.

The costs that are not message fees

Two more lines belong in an honest total, and neither appears in most comparisons.

• AI usage. Any step that calls a model — an AI check, or a knowledge-base answer — is metered separately from messaging. A knowledge base wired to answer every inbound message, including "thanks" and "ok", spends on nothing. Put specific keyword flows above it and restrict it to text messages. See AI answers from your own website.
• Your team's time. The largest cost in most small stores, and the one automation is actually bought to reduce. A flow that deflects order-status questions is worth more than its message spend even if the messages were not free — and they usually are.

What to measure instead of total spend

Total spend on its own tells you nothing about whether the channel is working. Cost per outcome does, and it also shows you which flow to fix first.

1. Cost per confirmed order for the confirmation flow — message spend divided by orders confirmed. In cash-on-delivery markets, compare it against what one refused delivery costs you.
2. Cost per recovered cart — and only believe it if you ran a holdout, because a share of those customers would have returned anyway.
3. Cost per deflected support conversation — usually close to zero, since these run inside open windows.
4. Cost per qualified ad lead, compared against your other paid channels rather than against other WhatsApp flows.

The method, including why the holdout matters, is in measuring whether your automation works.

One closing caution on any figure in this article or elsewhere. Meta revised rates on 1 July 2026, with further adjustments scheduled for 1 October 2026, and moved from conversation-based to per-message pricing on 1 July 2025. A great deal of content still describes the old model. Treat the pricing page as the only current source, and re-check it before you build a budget around anything.