How many emails, how far apart, and what each one is allowed to say. The published data from Klaviyo and Omnisend agrees on the shape more than most marketers expect.
The abandoned cart flow is a sequencing problem before it is a copywriting problem. The same three emails in the wrong order, or at the wrong intervals, recover less than a mediocre set of emails timed correctly, because the value of each message depends on how fresh the shopping session still is. The good news is that the two largest public datasets on this flow, from Klaviyo and Omnisend, describe almost the same shape.
Automated flows are a small fraction of sending volume and a large fraction of results. Klaviyo’s 2026 benchmark across more than 183,000 brands found that flow emails generated roughly 41% of email revenue from just 5.3% of sends, with flows converting at a 2.11% placed order rate against 0.16% for regular campaigns. Within flows, cart recovery is the strongest performer: Klaviyo’s dedicated cart benchmark report, covering 143K+ flows sent in 2023, puts the average at a 50.5% open rate, a 6.25% click rate, a 3.33% placed order rate, and $3.65 in revenue per recipient. There is no other automation where three emails do this much work.
Omnisend’s guidance is that “a 30, 45, or 60-minute delay is best for your first cart abandonment email,” specifically so the customer gets a window to return on their own before the reminder lands. Klaviyo’s suggested sequence runs slightly cooler on the first touch and describes a common three-email pattern: a simple reminder after 2-4 hours, a follow-up with a discount after 24 hours, and a final message after 48 hours with alternative recommendations or a feedback ask. Between the two you get a defensible corridor rather than one magic number: first email within the first hours while the session is warm, second email the next day, third email a day or two after that, then stop.
Email one is a reminder and nothing more. It reproduces the cart, links back to checkout, and skips the discount entirely. The shoppers it recovers are the ones who got interrupted, and they cost you nothing to close. Baymard’s survey work is a useful reality check on what this email is up against: 43% of US shoppers abandon because they were just browsing, and no email timing fixes that, which is why the realistic goal is recovering a slice of the remainder rather than the whole 70.22% documented abandonment rate.
Email two does the persuading. A day has passed, the interruption excuse is gone, so something stopped this person. This is where you answer the documented blockers, shipping cost, delivery speed, returns, payment trust, and where Klaviyo’s suggested pattern places the discount if your margins allow one at all. Putting the offer here instead of in email one protects you from paying for recoveries that the plain reminder would have gotten for free.
Email three closes the loop and lets go. Klaviyo’s suggested final message includes alternative product recommendations or asks for feedback, which is the honest admission that the original item may simply be wrong for this person. If a genuine expiry applies to the offer from email two, this is where you say so, once, without a fake timer.
Three mechanics matter more than shaving minutes off any delay. First, the purchase filter: anyone who buys exits the flow instantly, at every step. Second, frequency collision: a shopper in the cart flow should not also receive that day’s campaign blast, so use your platform’s smart sending or flow priority settings. Third, the stop rule: when email three is done, the flow is done, and the person returns to normal sending. Omnisend’s 2025 user data shows why restraint costs so little, the average cart email already converts 1.51% of sends with a 39.46% click-to-conversion rate, meaning nearly four in ten people who click go on to buy. The flow works. Extending it past the third message mostly buys unsubscribes.
Set the corridor, assign each email its single job, wire the exits, and then leave it alone. This is one of the few places in retention where the published data hands you the architecture.
Within the first few hours, while the shopping session is still fresh. Omnisend's guidance is a 30 to 60 minute delay so the customer has a chance to come back on their own first, and Klaviyo's suggested window is 2 to 4 hours. Anything past a day stops being a reminder.
One is far better than zero, but both Klaviyo and Omnisend describe three-message sequences as the standard pattern because each position does a different job. If you only send one, you are betting the entire recovery on the reminder and skipping the objection and closing steps.
Not before the second email. Klaviyo's suggested sequence places the discount at the 24 hour follow-up, after the plain reminder has had its chance. Holding it back protects margin on the shoppers who only needed the nudge.
Yes, immediately, and every serious platform does this by default through a purchase filter on the flow. Sending a reminder for a cart someone already bought is the fastest way to look like a robot.