Most stores send the first abandoned cart email far too soon, because the default delay in the tool was written for a fast, cheap, impulse purchase and nobody changed it. The shopper has not abandoned anything. They are comparing, checking a delivery date, waiting for payday, or asking someone else in the house.
Sending into that moment does two things, neither of them good. It reminds a person who was still deciding that they can put the decision off, and it uses up the one message they were most likely to open on a moment when there was nothing to resolve.
We start from a different question. Not how fast can we send, but what stopped this person, and when will that thing have changed. A cart is abandoned for a handful of reasons, and each reason has its own clock. Get the clock right and the sequence does more with fewer messages, which is also better for the deliverability of everything else you send.
A cart is not a decision, it is a shortlist
People use carts the way they use a shortlist. Adding something is how you save it, compare it, and see what shipping would cost. Treating every add-to-cart as a purchase that nearly happened will make your recovery numbers look impressive and your judgement worse.
The practical consequence is that the first message should not read like a rescue. A shopper who is mid-comparison and receives an urgent reminder with a countdown learns that this store panics. A shopper who receives a calm message that restates what they were considering, with the answer to the question that probably stopped them, is being helped.
It also means your recovery rate is not a clean measure of the sequence. Some of the people who come back and buy would have come back anyway, because the shortlist was doing its job. The number worth watching is total orders from the store, and whether the sequence adds to them, not the revenue the email tool attributes to itself.
So the design brief for the first message is narrow. Remove one obstacle. Do not sell.
The first message should remove one obstacle, not rescue a sale that was never lost.
Match the delay to the reason people leave
Different obstacles resolve on different timelines, so a single delay cannot suit them all. This is how we think about spacing a sequence.
Shipping cost and delivery date questions resolve immediately or not at all. The shopper wanted a number they could not find. A message that answers it plainly is useful within a few hours, and stale after a day.
Comparison shopping resolves over an evening or a weekend. The person is looking at two or three options and will decide when they have looked at all of them. A message arriving before they have finished looking is noise. One arriving after they have finished, with the specific reason your version fits, arrives at the decision.
Price hesitation resolves on a schedule that has nothing to do with your store, usually a pay cycle. That is the message that belongs several days out, and it is the only place in the sequence where an incentive earns its keep.
A cart that was a genuine accident, a mis-tap or a broken checkout, resolves within minutes and is best handled by fixing the checkout rather than by email.
What we send, and when, for a considered purchase
For a store selling something people think about, three messages spread across several days does more than five messages crammed into a day.
The first goes out after enough time has passed that the session is genuinely over, not while the person is still browsing. It shows what they were looking at, answers the shipping and returns questions in plain numbers, and links straight back to the cart with everything still in it.
The second goes out a day or two later and does a different job. It is about the product, not the cart. Sizing, materials, how long it lasts, what it works with, what people ask before they buy. This is where a store wins the comparison, and it is the message stores most often skip.
The third goes out several days later and acknowledges reality. Either an incentive, if your margin can carry it and your brand is comfortable with it, or a simple offer to answer a question with a human reply address. For a high-consideration product the human reply often outperforms the discount, because the obstacle was uncertainty rather than price.
For a low-price, repeat-purchase item, compress all of this. A shopper buying refills does not need three days to think.
The discount you add trains the next visit
A discount in the first cart email is the most expensive habit a store can build. It is paid to people who were already coming back, and it teaches a segment of your customers that the way to buy from you is to fill a cart and wait.
You can watch this happen in the data. Look at customers who have received a cart discount more than twice. If a growing share of your orders carry a recovery code, the sequence has stopped recovering anything and started running your pricing.
When we do use an incentive, it goes last, it is capped, and it is framed as a reason rather than a reflex. Free shipping on this order is often better than a percentage off, because it removes the obstacle that most commonly stopped the cart and it costs you less than a discount on the product.
The better lever is upstream. Most of what a cart sequence is trying to repair was created on the product page or at the first checkout step. Shipping cost hidden until the end, a forced account, a slow payment step. Fix those and the sequence has less to do, which is the outcome you actually want.
Judging the sequence honestly
Two numbers tell you whether the change worked, and neither of them is the revenue figure your email platform reports.
The first is checkout completion for the store as a whole, month over month. A cart sequence that is working shows up here, because the people it brings back complete checkouts they had left.
The second is the share of your list that is still opening. A sequence that fires too often on too many triggers burns attention, and the cost lands on your other email. If open rates on your regular sends fall in the months after you add a sequence, the sequence is not free.
Test the timing rather than the copy first. Changing a delay is one field, it applies to everyone, and it usually moves more than a rewritten subject line. Change one delay, leave it for a month with steady volume, and compare the same span of days before and after.
And keep the exit rules tight. Anyone who buys leaves the sequence immediately, anyone who buys a different item leaves too, and nobody receives it twice in a week. Those three rules prevent most of the complaints stores get about their own automation.