Reorder reminders for things people run out of

A reorder reminder messages a customer around the time they are likely to run out of something they bought before. For consumables with a predictable cycle — skincare, supplements, coffee, pet food, contact lenses — it is one of the few marketing messages customers genuinely welcome.

For everything else it is an offer wearing a helpful disguise, and customers can tell.

Is it worth building for you?

Product	Predictable cycle?	Build it?
Supplements, skincare, coffee, pet food	Yes, fairly	Yes
Contact lenses, razor blades, filters	Very	Yes
Clothing, homeware, electronics	No	No — this is just a campaign
Gifts and seasonal items	Annual at best	Use a date-based campaign instead

The test is whether you can name the interval. If you cannot say roughly how many days a customer's purchase lasts, you are not sending a reminder — you are sending a promotion and calling it one.

The part nothing does for you

Be clear about this before you plan around it: nothing tracks consumption, and no lifetime-order fields exist. There is no built-in "last order date" or "total orders" attribute, so the reminder has to be built from a date you stamp yourself.

The mechanism has two halves:

1. Stamp the date on every order. A flow triggered by your store's order event, with an Update Customer Attribute step writing today's date to a field such as last_order_date.
2. Build the audience from that field using a relative date filter — "more than N days ago" — and send a broadcast to it.

Two honest caveats. The field must be created first under Settings → Contact Attributes or the save does nothing. And the audience only knows about orders placed after you switch the stamping on, so build it early even if you will not use it for months.

Choosing the interval

Work it out from your own repeat data rather than guessing, and then send slightly before you think they run out. A reminder that arrives after the customer has already bought elsewhere is worse than useless — it tells them you were too late.

• A 30-day supply — remind at around 25 days.
• A 60-day supply — remind at around 50.
• Unsure — pick the shorter end, since early is recoverable and late is not.

It is a marketing message

This matters for cost and reach. A reorder reminder promotes a new purchase, so it is a marketing template: charged on every delivered message, subject to Meta's adaptive per-user marketing limits — for which no numeric threshold is published — and undeliverable to US phone numbers entirely. Those limits do not apply in the EEA, the UK, Japan or South Korea.

So it needs marketing consent, collected separately from order-update consent, and it counts against the frequency budget you spend on campaigns. See how often you can message a customer.

Write it as a reminder, not an offer

The whole appeal is usefulness. Three rules:

1. Name what they bought. "Running low on the [product]?" is a reminder. "Check out our range" is a campaign.
2. Make reordering one step. A link straight to the item, not to your homepage.
3. No discount by default. A customer who was going to reorder anyway does not need one, and offering it teaches them to wait.

Send it once

One reminder per cycle. A customer who did not reorder either bought elsewhere, still has some, or has stopped using the product — and none of those is improved by a second message. A follow-up converts very little and produces the opt-outs that damage your quality rating.

The version that costs nothing

There is a cheaper approach worth trying first. Instead of a scheduled broadcast, ask at the moment of delivery: "most people finish this in about six weeks — want us to check in then?" A yes gives you explicit consent and a much warmer reminder later.

That question is free, because it happens inside an open conversation window under Meta's pricing rules, and it converts better than an unsolicited reminder because the customer asked for it.

Measuring it

• Reorder rate among reminded customers versus a holdout. Without a holdout you are counting people who would have reordered anyway.
• Opt-outs following the reminder. A rise means your interval is wrong or your list is too broad.
• Revenue per message sent, since every one is charged.

Method is in measuring whether your automation works.

Where the date comes from

Because nothing tracks consumption, the whole flow rests on one field you maintain yourself. Three ways to fill it, in order of reliability:

Method	How	Trade-off
Stamp on order	A flow on the store's order event writes today's date	Best — automatic, but only from the day you switch it on
Stamp on delivery	Same, on the delivered event	More accurate, but the delivered event is unreliable
Ask the customer	"Want a nudge when you're running low?"	Explicit consent, best conversion, needs them to answer

Most stores should do the first and the third together: stamp automatically so the audience builds itself, and ask at delivery so the customers who opt in get a warmer, better-converting version.

Building the audience

1. Create a date field under Settings → Contact Attributes — for example last_order_date. Without this the save silently does nothing.
2. Build a flow on your store's order event with an Update Customer Attribute step writing today's date to it.
3. When you want to send, build a segment filtered on that field with a relative date condition — "more than N days ago".
4. Exclude anyone who has ordered since, or you will remind people who already bought.

Step four is the one that produces complaints when skipped. A reminder to reorder something the customer bought last week reads as carelessness, and it is the commonest defect in a first attempt at this flow.

What it costs

Every reminder is a marketing template on a cold chat, so it is charged per delivered message under Meta's pricing documentation — unlike the utility flows around it, which are free inside an open conversation window.

That makes the economics worth checking rather than assuming. Revenue from reorders above your holdout baseline, minus message cost, is the number. If it is not clearly positive, a smaller, better-targeted list will fix it faster than a better message will.

Who not to send it to

• Anyone who ordered since. Obvious, and easy to forget.
• One-off gift buyers. Someone who bought a present is not on a consumption cycle.
• Customers who never opened the last two. Sending to people who ignore you is what tightens Meta's adaptive per-user cap.
• US numbers, which cannot receive marketing templates at all.

Filtering those out typically halves the send and improves every metric that matters, because the people removed were the ones producing cost without revenue.

Timing by category

The interval is the whole design, and getting it wrong in either direction wastes the message. Some rough starting points, to be replaced by your own repeat data as soon as you have it:

Product	Typical supply	Send at around
Daily supplements	30 days	Day 25
Skincare, one item	45–60 days	Day 45
Coffee, regular drinker	21–30 days	Day 20
Pet food, one bag	30 days	Day 25
Contact lenses, monthly	30 days	Day 25

Send early rather than late. A reminder that arrives after the customer has bought elsewhere does not just fail — it tells them you were slow, which is worse than saying nothing.

The version worth testing first

Before building the scheduled version, try the free one. At delivery, ask whether they would like a nudge when they are running low, and save the answer.

That question costs nothing, because it happens inside an open conversation window under Meta's pricing documentation — service conversations have been free and unlimited since 1 November 2024. It also gives you explicit, recorded consent, and a reminder someone asked for converts far better than one they did not.

If that version works, the scheduled broadcast becomes a way to scale it rather than a bet you are placing blind.

Getting the message right

A reorder reminder lives or dies on sounding useful rather than promotional. Compare:

Reads as a campaign	Reads as a reminder
"Time to restock! Shop our bestsellers now."	"Running low on the [product]? It's about six weeks since your last one."
"20% off everything this week only."	"Same one again? Here's a link straight to it."
"We miss you — come back and save."	"Just a nudge in case you're nearly out."

The right-hand column names what they bought, references the interval, and makes reordering one tap. The left-hand column could have been sent to anyone, which is exactly what makes it feel like marketing.

When to stop sending them

After one unanswered cycle, pause that customer. Someone who ignored a well-timed, well-written reminder has either stopped using the product or bought elsewhere, and a second cycle of reminders converts almost nothing while producing the opt-outs that damage everything else you send.