Seven out of ten carts die before checkout. Here's how to actually get some of them back, without training your customers to wait for a coupon.
Abandoned cart recovery is not one email, it’s a decision tree based on intent. A browse abandoner, a cart abandoner, and someone who started checkout and hit a shipping cost surprise are three different problems, and AI segmentation tools (Klaviyo, Omnisend) can now split them automatically instead of dumping everyone into one flow. Use AI for dynamic product blocks pulling the actual abandoned items, send-time optimization based on real open behavior, and subject line generation, but never skip the human edit pass. And be honest with yourself about discounts: a flat 10-15-20% escalating sequence is the fastest way to teach repeat customers that abandoning is a negotiating tactic. Segment discounts by cart value and purchase history instead of blasting them to everyone who leaves a tab open.
Picture the moment: someone adds a jacket to their cart, gets to the shipping page, sees $14.99 tacked on for standard delivery, and closes the tab. That single moment, multiplied across every ecommerce store on earth, is why the Baymard Institute’s tracked average across 50 separate studies sits at 70.22% cart abandonment. Nearly three out of four carts never become an order.
Before anyone panics, some of that number is just how shopping works. People compare prices across three tabs, save things for payday, or add an item to a cart the way they’d toss something in a physical basket while still deciding. Baymard’s own research on abandonment reasons backs this up: 42% of US shoppers say they abandoned a cart because they were “just browsing, not ready to buy.” That’s not a recovery problem, that’s window shopping, and no AI tool on earth is going to convert someone who never intended to buy today.
The abandonment worth chasing is the other half: the 40% who cite extra costs (shipping, tax, fees) as the reason they bailed, the 20% who found the checkout too slow, the 18% who didn’t want to create an account. Those are fixable, and they’re exactly the moments a well-built recovery email or text can catch, because the intent was real. The person got close. Something specific stopped them. Your job is figuring out what that something was for each segment, and that’s where the segmentation work in this piece starts to matter more than the copywriting.
This is also why abandoned cart flows consistently outperform every other automated email you send. Klaviyo’s benchmark data across more than 143,000 abandoned cart flows shows an average conversion rate of 3.33%, nearly double the placed-order rate of a welcome series, because the customer already told you what they want. You’re not generating demand from scratch. You’re just removing whatever friction got in the way.
If you’re on Shopify, you already have a baseline recovery flow whether you’ve touched it or not. Shopify’s native abandoned checkout automation, found under Apps > Messaging > Automations in the admin, sends a recovery email automatically once a checkout is abandoned, and you can enable an automatic discount that applies directly to the customer’s pre-filled cart so they only need to re-enter payment details. It’s not AI-driven, and it’s not going to win any awards for personalization, but it’s a real safety net, and I’ve seen too many stores go live without even turning it on.
For anything beyond that baseline, Klaviyo is the default for a reason. Klaviyo AI generates subject line variants, drafts flow copy, and powers the predictive analytics behind segments like “likely to abandon” or “predicted next order date.” Its abandoned cart flow templates already pull dynamic product blocks from the actual cart contents (not a generic best-sellers list), and its Smart Sending feature caps how many messages a single customer gets across flows in a given window, which matters more than people expect once your welcome series, cart flow, and a campaign all want to fire on the same person the same week. If you’re trying to decide between Klaviyo and the other major email platforms for this kind of work, we’ve laid out an honest comparison of Klaviyo, Mailchimp, and ChatGPT-assisted email workflows that’s worth reading before you commit budget.
Omnisend plays in the same space with a lower price point and a genuinely useful AI layer of its own: an AI assistant that builds audience segments by spotting behavioral trends, and an AI-assisted form builder for capturing the email or phone number in the first place, which matters because you cannot recover a cart you have no way to contact. Omnisend’s own platform data shows automated cart recovery emails driving 22 times more orders than a standard promotional send, and its cart tracking works off both entered email addresses and browser cookies, so you can catch abandoners who never got as far as typing anything in.
One thing worth being blunt about: this piece is about the pre-purchase recovery moment specifically, the cart or checkout someone walked away from before ever becoming a customer. If you’re looking for the AI logic behind post-purchase emails, the “you might also like this too” messages sent after someone has already bought, that’s a different problem with different timing and psychology, and I’ve covered how to build those upsell and cross-sell campaigns separately. Don’t run the same playbook for both. A customer who hasn’t paid you yet needs a different message than one who already has.
Here’s the mistake I see on almost every recovery flow audit: one segment, “abandoned checkout in the last 24 hours,” getting the exact same three emails regardless of how far the customer actually got. That’s treating a window shopper and someone who entered their credit card number and got a payment error as the same problem, and they are not even close to the same problem.
Cart value is the second axis, and it’s the one teams skip because it takes more setup. A $40 cart and a $400 cart should not get identical recovery treatment. Klaviyo and Omnisend can both build live segments on predicted order value and purchase history, which means your highest-value abandoners can get a more personal touch, maybe a different sender, a different subject line, a customer service check-in instead of a coupon, while lower-value browse abandoners get the lighter-touch automated version. Recognizing which customer you’re looking at costs you nothing once the segmentation is built. Blasting everyone the same offer costs you margin on the segment that would have paid full price anyway.
The AI genuinely earns its keep here, because building these segments by hand across a large customer base was never realistically happening. Reviewing checkout-stage data, cart value, and purchase history for every abandoner manually doesn’t scale past a few hundred customers a month. Klaviyo’s and Omnisend’s segmentation engines can do it continuously, updating as behavior changes instead of relying on a static list someone built in a spreadsheet three months ago.
Once you know who’s getting the email, the content itself is where AI tools save the most real time, if you use them for the right job. The job is not “write me an abandoned cart email.” The job is “populate this email with the exact three items sitting in this exact cart, ranked by likelihood to convert, and give me three subject line options to test.”
Both Klaviyo and Omnisend can pull live product data, image, price, current stock status, directly into the recovery email instead of a static screenshot taken when the flow was built. That matters more than it sounds: if a product goes out of stock or the price changes after someone abandons their cart, a static email keeps advertising something that no longer exists or costs a different amount, which is its own small trust problem. Dynamic blocks update at send time, not build time.
Instead of a flat “send 4 hours after abandonment” rule for your entire list, AI-driven send-time features (built into Klaviyo’s and Omnisend’s automation engines) can shift the send window based on each customer’s own historical open behavior. Someone who reliably opens email at 7am gets the recovery email closer to 7am, not at 2pm because that’s when the flow happened to fire for everyone else. It’s a small lift on open rate individually, but it compounds across an entire list.
This is the part everyone reaches for first, and the part I’d rank as least important of the three. Klaviyo AI and Omnisend’s AI assistant will both generate subject line variants and full flow copy from a short brief in seconds. Use it. It’s faster than staring at a blank subject line field. But treat what comes back as a first draft every time, not a finished send.
The tell that an AI-drafted recovery email hasn’t been edited by a human: it reads like every other AI-drafted recovery email. Generic urgency (“don’t miss out!”), no actual reference to what’s in the cart, a tone that doesn’t match anything else your brand has ever sent. Run a real edit pass, every time, especially on the subject line.
Now the part almost nobody wants to say out loud. The default abandoned cart sequence copied across a huge share of Shopify stores looks like this: email one at 2-4 hours with no discount, email two at 24 hours with 10% off, email three at 48-72 hours with 15-20% off and a countdown timer. It’s a template everyone has seen, because it’s a template almost everyone copied from somewhere else.
Here’s the problem with running that sequence forever without thinking about who’s actually receiving it: repeat customers learn it fast. If a customer figures out that abandoning their cart reliably gets them a coupon code 48 hours later, some meaningful share of them will start doing it on purpose. You haven’t recovered a sale you were going to lose anyway, you’ve trained your best, most loyal customers to never pay full price again. That’s not a hypothetical. It’s one of the most consistent patterns I’ve seen when auditing flows for brands with a real repeat-purchase habit, subscription boxes, consumables, anything people buy more than once.
Even Klaviyo’s own benchmark research flags this directly: discounts aren’t always necessary in a recovery flow, and they can work against you by training shoppers to wait for one before buying at all. Postscript, an SMS platform built specifically around cart recovery, puts it just as bluntly in its own guidance: constantly offering discounts in recovery messages can encourage future cart abandonment for the sake of another deal, and it quietly erodes the value of your brand along with your margin.
My rule of thumb on this one: if you wouldn’t feel comfortable explaining your discount logic to a customer’s face (“we give everyone who leaves a full cart a coupon two days later”), don’t automate it blindly across your whole list either. AI makes it cheap to send the same escalating discount to everyone. Cheap doesn’t mean it’s the right call.
SMS earns a real spot in cart recovery for one reason: speed. A text sent within the first 15-30 minutes of an abandoned checkout catches someone while they’re still mentally in the purchase, often still holding their phone, in a way a 2-hour-delayed email doesn’t. Postscript, one of the platforms built specifically for ecommerce SMS, recommends that first message land within the initial 1-2 hour window, with a softer follow-up a couple of days later for anyone still deciding.
The honest caveat: SMS only reaches customers who opted in, and opt-in rates are meaningfully lower than email capture rates on most stores, so it’s a smaller, higher-intent audience by nature rather than a replacement for your email flow. Treat it as a second channel for the segment that gave you a phone number, not a swap-out. If you’re building this out for the first time, we’ve written a fuller breakdown of how to run AI-assisted SMS marketing that covers the compliance and timing details this piece doesn’t have room for.
Klaviyo and Postscript both let you build cart recovery texts with the same dynamic product data used in email (product image, price, a direct link back to the cart), and both cap send frequency the same way email platforms do, since nothing kills an SMS opt-in list faster than three texts in one day about the same abandoned hoodie. If you’re running both channels for the same abandonment event, coordinate them. A customer who gets an email and a text within the same ten minutes about the same cart doesn’t feel personalized, they feel targeted, and that’s a fast way to lose the SMS opt-in entirely.
Cart recovery is one of the few flows where testing pays off fast, because volume is usually high enough to reach statistical significance within a couple of weeks instead of a couple of quarters. Klaviyo’s own guidance on this, echoed by agency partners who build these flows for a living, points at the same handful of variables worth testing systematically: message frequency, send-time delay, subject lines, copy tone, and whether a discount appears at all.
Run one variable at a time, not five at once, or you won’t know what actually moved the number. Start with timing, since it’s the cheapest test to run and often has the biggest effect: try a 1-hour delay against a 4-hour delay on the first email for two to four weeks, holding everything else constant. Then test discount presence: run one segment through a no-discount sequence and a matched segment through your current discount sequence, and measure total revenue recovered, not just the raw conversion rate. This is the step people skip, and it matters. A sequence with a discount often converts a slightly higher percentage of carts, but at a lower average order value and a lower margin per order. Sometimes the “worse converting” sequence is the more profitable one.
Let the test run long enough to matter. Two or three days of data on a mid-size list is noise, not a result. Give it the two to four weeks most platforms recommend before declaring a winner and moving on to the next variable.
For most Shopify or BigCommerce stores, Klaviyo AI is the most built-out option, with dynamic product blocks, predictive segmentation, and Smart Sending frequency caps included in its abandoned cart flow templates. Omnisend AI is a strong lower-cost alternative with its own AI segmentation and form-building tools. If you’re only using Shopify’s native automation, it works as a baseline but has none of the AI-driven personalization or send-time optimization the dedicated platforms offer.
Three emails is the most common structure (an initial reminder within a few hours, a follow-up around 24 hours, and a final message at 48-72 hours), and Klaviyo’s own data shows three-email sequences significantly outperforming single emails on total revenue. That said, don’t assume three is automatically right for your list. Test a two-email version against your current three-email sequence before locking in a length.
No, and defaulting to one is a common mistake. Discounts can train repeat customers to abandon their cart on purpose, expecting a coupon to show up within a day or two. Segment by cart value and purchase history first, try a non-discount incentive like free shipping, and reserve percentage discounts for the segments where the math (higher-value cart, first-time customer sitting on the fence) actually supports the margin hit.
For email, most platforms recommend the first message land within 1 to 4 hours of abandonment, while purchase intent is still fresh. For SMS, the window is tighter, ideally within 15 to 30 minutes, since text is a more immediate channel and works best while the customer is still mentally in the purchase.
SMS typically sees faster response times and higher open rates than email, simply because texts get read almost immediately and most cart recovery texts land within a much shorter window after abandonment. But SMS only reaches customers who’ve opted in via phone number, which is usually a smaller list than your email base. Treat SMS as a high-intent second channel layered onto your email flow, not a full replacement for it.
This guide draws on the Baymard Institute’s tracked cart abandonment statistics, Klaviyo’s Abandoned Cart Benchmark Report (analyzing over 143,000 flows), Omnisend’s abandoned cart automation platform data, Shopify’s official Help Center documentation on checkout recovery, and Postscript’s published guidance on SMS cart recovery, including their own warning against over-reliance on discounts. All statistics and platform features were verified against live source pages before publishing. Sources are linked below.
Weekly tips. Real examples. Practical help for busy professionals.
We care about your data, check out our privacy policy.