How many WhatsApp automations should a store actually run?

A frequency budget is a deliberate cap on how many messages one customer receives from you across all your flows in a given period. Most stores do not have one, which is how a single customer ends up receiving an order confirmation, a shipping update, a review request and a cart reminder inside one week.

The failure that follows is worse than it looks. A customer who mutes or blocks you does not lose only your marketing — they lose the delivery notification too, and you lose the channel entirely for that person.

Start with five flows, not fifteen

For a store under roughly 500 orders a month, this set covers the ground:

Flow	Messages per customer	Category	Cost
Stop the bot when a human replies	0	—	Free
After-hours reply	0–1 per visit	Service	Free
Order confirmation with buttons	1 per order	Utility	Charged once, then free
Shipping and tracking	1 per order	Utility	Free inside the window
"Where is my order?"	Only when asked	Service	Free

Four of the five cost nothing to run, because they happen inside an open conversation window — which Meta's pricing documentation confirms has been free and unlimited for service conversations since 1 November 2024.

Then add carefully

Each addition should have to justify itself against a specific number: how many extra messages does this put in front of one person?

• Review request after delivery — 1 per order. Worth it.
• Abandoned checkout recovery — 1 or 2 per abandoned cart, and these are marketing: charged, capped, and undeliverable to US numbers.
• Reorder reminder — 1 per cycle, and only for consumables.
• Campaign broadcasts — this is where budgets are actually blown.

The arithmetic that should worry you

A customer who orders twice a month from a store running everything above receives, in a month: 2 order confirmations, 2 shipping updates, 2 review requests, possibly 2 cart reminders, and however many campaigns you sent. That is comfortably into double figures for someone who simply likes your products.

Now consider that Meta's per-user marketing cap is shared across all businesses and tightens for recipients who appear unreceptive — with no published number. You are not budgeting against your own sends. You are budgeting against everyone's.

Three rules that keep it sane

1. One marketing message per customer per week, maximum. The reasoning behind that figure, and how to budget the other message types around it, is in how often you can message a customer. Set it as a per-contact frequency cap under Broadcasts → Opt-outs so it is enforced rather than remembered.
2. Never let two flows fire for the same event. Two order-placed automations mean two confirmations. Before publishing, check what already triggers on that event.
3. Prefer one flow with branches to three flows with conditions. Three separate keyword automations for returns, refunds and exchanges can all match one message. One flow with three branches cannot.

Prefer utility over marketing wherever possible

This is the strategic version of the frequency argument. Utility messages are free inside the window, exempt from per-user marketing caps, deliverable to US numbers, and welcomed rather than tolerated. Marketing is charged, capped, regionally blocked and resented.

So the highest-return move is usually not adding a marketing flow — it is making an existing utility flow earn a reply, which opens the window and lets you say anything for free. That argument is developed in the 24-hour window.

How to tell you have gone too far

• Your quality rating moved off green.
• Opt-outs are rising faster than your list.
• Reply rates on transactional messages are falling — the clearest sign that people have stopped reading you.
• You are seeing error 131049 on campaigns.

The first and third are leading indicators; by the time you are seeing 131049 you have been over-sending for a while.

An honest default

If you want a rule of thumb rather than a framework: run every utility flow you can, and one marketing flow at a time. Utility messages buy goodwill and cost nothing. Marketing spends both. A store that gets this ratio right can automate almost all its customer communication and still be welcome in the chat.

A month in one customer's inbox

The clearest way to check your own frequency is to write out what a single regular customer receives. For a store running everything, ordering twice a month:

Week	Messages that customer gets	Category
1	Order confirmation, shipping update	Utility
1	Review request after delivery	Utility
2	Campaign broadcast	Marketing
3	Order confirmation, shipping update, review request	Utility
3	Abandoned cart reminder from a browse	Marketing
4	Campaign broadcast, reorder nudge	Marketing

That is ten messages in a month to someone whose only crime is liking your products. Six are welcome; four are you asking for something. The utility ones are free and the marketing ones are charged, so the half that irritates is also the half you are paying for.

Where to cut first

If that exercise made you uncomfortable, the order to trim in:

1. The second campaign in a month. Almost never worth its cost in goodwill.
2. The reorder nudge, unless you sell genuine consumables with a predictable cycle.
3. The cart reminder for existing customers — someone who orders twice a month browsing without buying is not a lost sale, they are shopping.
4. The review request on small or repeat orders. Ask on first orders and high-value ones.

What should survive every cut is the transactional set. Those are the messages customers would complain about not receiving.

One flow per job, and one job per flow

Two failure modes, in opposite directions, and both cost you.

Too many flows for one job means separate automations for returns, refunds and exchanges — all of which match the message "I want to return this for a refund", so the customer gets three replies. Merge them into one flow with three branches.

Too few flows for different jobs means bolting an FAQ responder onto your order-confirmation flow, so you cannot fix a bad FAQ answer without touching the thing that confirms orders, and one on/off switch controls both. Separate rules all still run on the same customer message, so splitting costs nothing.

A quarterly audit

1. List every published automation and what triggers it.
2. Flag any event with more than one flow on it.
3. Check each flow's activity view — one with zero runs is either broken or unnecessary, and both are worth knowing.
4. Write out the month-in-one-inbox table above with your real flows.
5. Retire the worst performer rather than adding another.

Step three catches more than you would expect. A keyword flow that stopped matching because customers changed how they phrase things does not error — it silently stops running, and nobody notices for months.

Why the platform enforces this whether you do or not

Frequency discipline is not only a matter of taste. Meta operates an adaptive cap on how many marketing template messages one person receives from all businesses combined, tightening for recipients who appear unreceptive, and its per-user marketing template limits documentation publishes no numeric threshold for it.

Two consequences follow. First, you are not budgeting against your own calendar — you are competing for an allowance other shops are also spending. Second, the only lever on your side is being a business that recipient wants to hear from, which is exactly what a lower frequency produces.

The enforcement is real at the account level too. Blocks and reports move your quality rating, and quality gates increases to your messaging tier — which starts at 250 unique recipients per rolling 24 hours and runs 250 → 2,000 → 10,000 → 100,000 → unlimited. Over-messaging therefore caps how many people you can reach at all. The store that sends less ends up able to send to more.

The exception worth knowing

Everything above concerns messages you start. Replies inside an open 24-hour conversation window are a different category entirely: uncapped, free, and not subject to per-user marketing limits.

So the frequency budget really only constrains outbound campaigns and cold templates. A store whose customers message it — from an ad, a website button, a QR code on the packaging — can have far more conversation with them than the numbers above imply, because none of that traffic counts. Designing to be messaged is the way around the whole constraint, and it is covered in the 24-hour window.

The question to ask before adding one

Not "would this be useful?" — almost anything would. Ask instead: which existing flow would I switch off to make room for this one? If the answer is none, you are adding to the frequency budget rather than spending it better.