Most people who reach checkout and leave were going to buy. Something interrupted them. A distraction, a price they wanted to think about, a form they could not complete on their phone. A recovery email brings a share of them back.
It also annoys people if done badly, and the difference is mostly timing and honesty.
Capture the address early
You cannot email someone whose address you never collected.
That means the email field should be near the start of checkout rather than the end, so a customer who abandons halfway has already given you a way to reach them.
Which raises the obvious question of consent, covered below, capturing an address for the purpose of a recovery email is a different thing from adding someone to a mailing list, and treating them as the same is where this goes wrong.
Fix the reason before automating the recovery
The order that matters.
Recovery emails recover a fraction. Removing the cause recovers everyone it would have affected.
The usual causes are the same everywhere: unexpected shipping cost revealed at the last step, forced account creation, a form asking for more than it needs, and a checkout that is awkward on a phone. For finding which one is yours, see WooCommerce Checkout Optimization.
A shop sending recovery emails for abandonment it created is paying to patch a hole it could close.
Timing: one hour, one day, and stop
The first message about an hour after abandonment, while the intent is still fresh and the reason may still be a distraction rather than a decision.
A second after roughly a day, for people who genuinely meant to come back.
A third, if any, several days later, and this is where most shops overreach. Beyond three messages you are not recovering a sale; you are training someone to ignore your mail, or to mark it as spam, which costs you delivery for every other message you send. What to Do When Your IP or Domain Is Blocklisted explains where that ends.
Do not lead with a discount
The first message should be a reminder, not an offer.
Discounting immediately teaches customers that abandoning a cart produces a code, and a portion of them will do it deliberately from then on. You have converted a full-price sale into a discounted one and taught the behaviour that repeats it.
If you use a discount at all, put it in the last message and make it time-limited. And check the arithmetic: on a thin margin, a 10% code can cost more than the sale was worth. How to Set Up WooCommerce Coupons and Discounts sets out limiting who can use one.
Make the link do the work
The message should contain a link that restores the cart exactly as it was, same items, same variations, same quantities, and ideally lands them at checkout rather than at the cart page.
A recovery email that sends someone to your homepage is asking them to do the shopping again, and most will not.
Test that link from a device that is not logged in. A restore link that only works in the browser that created the cart recovers nobody, and it is the most common way this feature is quietly broken.
Consent and the legal side
Rules differ by jurisdiction and the safe position is narrow.
Say at the point of capture what the address will be used for, and keep recovery messages separate from marketing consent. Someone who agreed to be told about their abandoned cart has not agreed to a newsletter.
Include an unsubscribe in every recovery message and honour it immediately. And do not send to someone who abandoned the same cart repeatedly; that is not a reminder, it is pursuit. For the wider obligations, see What a Website Needs for Privacy and Cookie Compliance.
These emails have to arrive
A recovery message that lands in spam is worse than none, because it is one more signal against your sending reputation with no return.
Send through authenticated SMTP instead of the server's default, with SPF and DKIM correct for the sending domain. If the volume is significant, use a sending service rather than your hosting account, which has limits and shares its reputation. How to Fix WordPress Email Not Sending walks through why mail is accepted and then discarded.
Exclude the people it should not reach
Anyone who completed the order after abandoning: an email arriving after they bought reads as incompetence.
Anyone whose cart contained something now out of stock, unless the message says so.
And your own test orders, which is how a staging copy sends recovery mail to real customers. There is more on disabling mail there in How to Set Up a Staging Environment on a VPS.
Measure recovered revenue, not open rate
The only number that matters is orders completed through the recovery link, minus any discount given.
Compare that to the cost: the sending service, the plugin, and the goodwill spent on people who did not want the message.
A recovery sequence that recovers a few percent of abandoned carts is doing well. One that recovers nothing after a month is either reaching nobody (check delivery first) or recovering people who were never going to buy.
Measure what the recovery is actually worth
Abandoned basket messages are judged by whether they produce orders, and that requires attributing the order to the message rather than assuming. Record how many are sent, how many are opened and how many result in a completed purchase. A sequence producing almost nothing is worth removing, since every message is a cost and a small risk to the sending reputation. The other figure worth reading is how many baskets are abandoned at all, because a high rate is a checkout problem that no follow up message fixes.