Should a cart reminder include a discount?
The discount works. That is exactly why it is a trap. Four alternatives and a rule for when to use it anyway.
A discount in a cart reminder will recover more carts and will teach some of your customers that abandoning a checkout produces a discount. Both things are true, and the decision is about which cost you would rather pay.
Most advice skips the second half. This page does not, because the second half is what shows up in your margin six months later.
The mechanism nobody mentions
Customers learn. A store that reliably sends 10% off an hour after an abandoned checkout is a store where the informed customer abandons deliberately. It does not require cynicism, only one experience of it working.
The effect is invisible in your recovery numbers, which look excellent, and visible in your average discount rate, which nobody attributes to the cart flow.
Four things to try before a discount
| Instead of a discount | Why it works |
|---|---|
| Free or clarified delivery | Unexpected delivery cost is a leading reason for abandoning; naming it removes the surprise |
| Answer the likely objection | Returns policy, sizing, authenticity — the thing they went to check |
| Reassurance about stock | "Still available" is information, not pressure, when it is true |
| An offer to help | A question invites a reply, which opens a conversation window and costs nothing after that |
The last row is underrated and it is the one WhatsApp is uniquely good at. A reminder that ends with a question can turn into a conversation, and once the customer replies you are inside the 24-hour window, where messages are free and unlimited under Meta's pricing documentation.
When a discount is genuinely the right answer
- Second contact, not first. Reserve it for someone who did not respond to a plain reminder.
- Slow-moving stock. A discount on something that is not selling anyway is a different decision from a discount on your bestseller.
- A genuinely time-boxed promotion that exists independently of the cart flow.
- High-margin categories where the discount does not eat the order.
Notice that none of those is "always". A discount that fires on every abandoned checkout is not a tactic; it is a price cut you have not admitted to.
If you do use one, protect it
- Make it expire, so it is an event rather than a standing policy.
- Make it single use and tied to the customer, so it does not end up on a coupon site.
- Do not send it to the same person twice, which is what a tag on the contact is for.
- Keep it out of the first message, so the customers who would have bought anyway buy at full price.
The tagging half of that is in tagging and segmenting customers automatically — a "had a recovery discount" tag is one of the more useful labels a store can keep.
The template problem
A discount changes what the message is. A plain reminder can sometimes be written as a utility-style message about the customer's own order; a discount cannot. It is marketing, plainly, under Meta's template documentation, which means opt-in and the marketing category.
It also means a separate approved template. If you want to test with and without, you need both approved before you start, and approval is not instant.
A structure that tests it properly
- Run the plain reminder alone for a fortnight and record how many recovered.
- Add a discounted second message for non-responders only, and run another fortnight.
- Compare total recovered revenue, not recovery rate — a higher rate at a lower margin can be worse.
- Check the discount is not being claimed by people who never abandoned anything.
- Look at repeat abandonment among customers who received one. That is the training effect, and it takes a couple of months to appear.
Step five is the one that distinguishes a real evaluation from a flattering one.
The wording matters more than the number
Given two messages of equal value, the one that explains itself outperforms the one that just offers money. "We noticed you did not finish — delivery to your city is free this week" is a reason. "Here is 10% off" is a bribe, and customers price it accordingly.
Which is also the argument for the objection-answering version: it makes the customer's actual hesitation the subject, and hesitation is rarely about ten percent.
Frequently asked
What discount size works?
There is no honest general answer, and any specific figure you have read was somebody's store rather than yours. Test yours.
Can I put the code in the template?
Yes, as a variable, so each customer gets their own. A single shared code in an approved template is a code that will circulate.
Does a discount need a marketing template?
Yes, with opt-in. See utility versus marketing.
What if my competitors all discount?
Then the informed customer in your category already abandons deliberately, and the marginal effect of joining in is smaller than it looks. Competing on the objection-answering version is the less crowded option — see how long to wait before a cart reminder for the timing side of the same decision.
The arithmetic, done properly
The reason this decision gets made badly is that the good outcome is visible and the cost is not. Put both in the same sum.
Suppose 100 customers abandon. A plain reminder recovers some number of them at full margin. A discounted reminder recovers more, at a lower margin, and gives the discount to every one of the customers who would have come back anyway.
That last clause is the one people forget. If 10 of those 100 were going to return regardless, a blanket discount has just paid 10 unnecessary discounts to win however many extra orders it won. Whether that trade is good depends entirely on your margin, which is why no universal answer exists.
| What to measure | Why |
|---|---|
| Recovered revenue, not recovery rate | A higher rate at a lower price can be a worse month |
| Margin on recovered orders | The discount comes out of exactly this |
| Repeat abandonment by discount recipients | The training effect, visible after a couple of months |
| Full-price recovery in the control period | Your baseline — without it the test proves nothing |
The version most stores should build first
One reminder, no discount, with a line that answers the most common reason people abandon in your category. For a lot of stores that is delivery cost or delivery time; for others it is whether the item is genuine, or what happens if it does not fit.
Build that, measure it for a month, and you have both a working flow and the baseline you need to evaluate anything more aggressive. Starting with the discount means you will never know what the plain version would have done.
What to do with the customers it does not recover
Not another discount. A customer who ignored a reminder and an offer is telling you something, and the useful response is to stop rather than escalate.
Where they are worth revisiting is later, as part of an ordinary campaign to lapsed customers, at a point where the message is about something new rather than about a basket they abandoned six weeks ago. That is a different flow with a different audience, and keeping the two separate is what stops your cart reminders turning into a nagging sequence.
Free delivery is usually the better lever
If you are going to give something away, delivery is frequently the more efficient thing to give. It is often the specific reason the customer stopped, it costs you less than an equivalent percentage off the order, and it does not set a price expectation for the product itself.
A customer who learns that abandoning gets them free delivery has learned something far less damaging than a customer who learns that abandoning gets them 15 percent off your catalogue.
Making it a real decision
- Write down your margin on the products people abandon most.
- Decide the maximum you would pay to recover an order, as a figure rather than a feeling.
- Check what free delivery costs you against that figure — for many stores it is well inside it.
- Only then consider a percentage discount, and only on the second contact.
- Put an expiry on whatever you offer, so it is an event rather than a policy.
Four of those five steps happen in a spreadsheet rather than in the flow builder, which is the point: this is a pricing decision that happens to be delivered by an automation.
A note on honesty
Whatever you offer, it should be real. A "24-hour" discount that still works next week, a "last few remaining" line on a product with plenty of stock, a countdown that resets — customers notice, and what they take from it is not about the discount.
The message you can defend is the one worth automating, because it will run thousands of times without anyone reading it again.
The short version
Start without one. Answer the objection instead. Keep the discount for a second contact, for slow stock, and for customers who have shown you they will not come back otherwise — and always with an expiry and a code that belongs to one person.