Your Abandoned Cart Email Got the Click. Where Did It Go?

An abandoned cart email usually sends its click to one of three places: /cart, a product page or a checkout link. On the device that built the cart, most of these work fine. On a different device, several of them lose the cart, and your flow report looks the same either way.

Here's the version of this that plays out on a normal weeknight. A shopper fills a cart on a laptop and gets pulled away. Your abandoned cart flow fires on time. They open the email on a phone in bed and tap the button. They find an empty cart or a single product page, so they try to rebuild the order from memory on a small screen. Some of them finish. Some close the tab.

Every part of the flow did its job. Klaviyo logs the delivery, the open and the click. The failure happens one step after the last thing the report can see.

Why does this flow deserve the attention?

Abandoned cart is one of the highest-earning automations most brands run, so small problems in it are worth finding. Klaviyo's 2026 Abandoned Cart Benchmark Report, drawn from more than 110,000 customers, puts attributed revenue per recipient at $6.77 on average and $13.70 for the top 10%. The average click rate is 6.0%, and the top 10% reach 11.3%.

Those are attributed figures. They tell you how much revenue the flow gets credit for, and some of that revenue may have arrived without the email. They also tell you something useful about the click: it's rare, and each one comes from a person who took the trouble to come back. It makes sense to care what they land on.

Why does the cart disappear on another device?

A Shopify cart normally lives with the browser session that built it. A phone has no way to know about the laptop's cart, and a cookie can't help because a cookie never crosses devices. For the cart to appear on the phone, two things have to be true. Something has to have saved the cart on a server, and something has to tell that server who just clicked.

The email click can supply the second part, since the person tapped a link sent to their own address. A store login can supply it too. The first part is where Added to Cart flows run short. Klaviyo's Added to Cart event has no native saved-cart link, while Started Checkout has checkout_url, which restores across devices. That difference explains most of what follows. A 2022 Klaviyo community thread describes the same gap, and its accepted fix is a custom permalink needing "a fair bit of custom coding".

What does each destination show the shopper?

Destination	Same device	Different device	What to know
/cart	Their cart	Whatever that device holds, often nothing	Simple to set up. Reads as broken when empty.
Product page	One product, cart intact	One product, cart gone	Fine for a one-item cart. Loses everything else in a larger one.
checkout_url	Their checkout	Their checkout	Started Checkout flows only. Lands on a payment form.
Cart recovery page	Their cart, on a page built around it	Their cart, when the cart is held and the clicker is known	Needs an app holding the cart. Consent limits apply.

/cart

This is the most common choice, and it's a reasonable default for shoppers on the same device. On another device the shopper reaches a cart page with little or nothing in it. They've just clicked an email promising their cart, so an empty page looks like a mistake.

There's one more option here. A cart permalink can rebuild the items on another device. Most brands don't set one up, since it takes custom template work, and by default it sends the shopper straight to checkout with limited room to edit the cart. Shopify calls permalinks a legacy feature and says selling plans aren't supported. Cart permalinks vs cart recovery pages has the detail.

A product page

This gives the shopper something to look at. For a three-item cart, though, it shows one of the three. The other two depend on the shopper's memory.

The checkout link

Here Shopify has already solved the problem. If someone reached checkout, the checkout_url link in a Started Checkout flow restores across devices. So abandoned checkout doesn't need fixing. The link lands on a payment form, which suits someone who was ready to pay and only available once they reached that step.

Good, better, best for an Added to Cart button

Checkout flows already have their link, so the decision that's left is the button in your Added to Cart flow. Here's how we'd rank the options.

Rank	Link to	What the shopper gets	Where it falls short
Good	The product page	The product they added, on any device	One product, so the rest of a larger cart is lost on another device
Better	/cart	The whole cart, ready to edit, on the same device	On another device it can be empty, which looks worse than a product page
Best	A cart recovery page	The latest saved cart on whichever device they use, editable, with subscribe & save and room to keep shopping	Needs an app holding the cart, and works when the cart is held and the clicker is known

A cart rebuild link can rebuild the items on another device. Most brands don't set one up, since it takes custom template work, and by default it sends the shopper straight to checkout with limited room to edit the cart.

How can you see this on your own store?

Take the test yourself, because it takes less effort than reading more about it. Add a few things to a cart on your laptop, let your flow send, then open the email on your phone with your phone logged out of your store. Our second-device test has the full steps and what to look for at each one.

If the phone shows the cart, you're in good shape. If it doesn't, you've seen what your shoppers see, and you can decide how much it matters for your store. A brand whose carts usually hold one item has less at stake than a brand with mixed carts of several items.

Is this the same problem as Klaviyo missing shoppers?

No, and it helps to keep the two apart. One problem is the email never sending, because Klaviyo didn't know who was shopping. A lot of tools and writing focus there, and Safari's cookie limits are part of the story. Durable cookies and server-set identifiers can help with it on the same device.

The other problem starts after the email has been sent and clicked. A better cookie can't fix it, because the phone never held the laptop's cookie to begin with. A brand can solve the first problem completely and still send every cross-device click to an empty cart.

What should the click land on?

The destination should meet three tests. It shows the cart when the cart can be found. It lets the shopper change their mind about quantities and items, since they stepped away for a reason. And it behaves sensibly when the cart can't be found.

This is the job we built reOtter's cart recovery page to do. You keep your Klaviyo Added to Cart flow and change the button link. When reOtter holds that shopper's cart and knows who clicked, the cart goes back into the real Shopify cart on whatever device they're on, on a page inside your own theme and domain. They can change quantities, remove items and choose subscribe & save on any line where the product offers it. Nothing already on the device gets wiped, and a same-device click doesn't double the quantities.

The limits matter as much. Carts are held only with analytics consent from your store's banner, and we never work around it. Held carts last 30 days. reOtter doesn't identify strangers and uses no shared identity network. If the cart isn't held or the clicker isn't known, the shopper sees the page with whatever cart that device has. A cart they've already bought doesn't get put back.

How would you know whether a new destination is better?

Run a holdout. reOtter's built-in holdout sends a share of shoppers to your current destination and compares orders. It measures whether the page beats that destination, and it doesn't tell you whether the flow itself adds revenue. No brand has run it to a result yet, so we have no lift figure to offer you.

You can read the full picture of what the page does and doesn't do on our abandoned cart recovery page. Whether or not reOtter is the answer for you, test the click on a second device this week. You'll learn something your flow report can't show you.