If you run an online store, here's a low-effort way to check something that's easy to miss: pull up your analytics and see how many of your "abandoned" sessions are from returning, logged-in customers who show up on a different device type than their previous session, phone then desktop, for example. If that number is a meaningful chunk of your total abandonment, some of what you're calling lost sales might just be customers who never finished switching screens, not customers who changed their mind.
Why this blind spot exists in the first place
Most ecommerce analytics track sessions, not people. A session is tied to a browser instance and a device, and it closes out after a period of inactivity or when someone switches device or browser entirely. So when a shopper who added items on their phone later logs in and finishes on a desktop, the platform does not retroactively connect the two. It closes the phone session as abandoned and opens the desktop session as new, because that is what a session-based system is built to do: describe visits, not follow one customer's intent across every device they used to get there. This is not a bug in your analytics setup. It is how session boundaries work almost everywhere, and it means the number you see as "abandonment" is really "sessions that ended without a purchase," which is not the same thing as "customers who changed their mind."
How to actually run the check
The check only works for customers you can identify across devices, which in practice means logged-in accounts. Most platforms only link a phone session and a desktop session to the same person once there is a login, an account, or some other identifier tying them together, an anonymous guest checkout on two different devices cannot be connected this way, so treat the result as a floor, not the full picture. With that limitation in mind: pull the list of abandoned checkouts, keep the ones attached to a known customer account, then check whether that same customer completed an order on a different device type within a short window afterward. A day or two is a reasonable place to start. Every match on that list is a cart that was never actually lost. It just finished somewhere else.
What the result tells you
If that overlap turns out to be a meaningful share of your total abandonment, it is worth treating device continuity as its own problem, separate from the shoppers who genuinely did not come back. If it is closer to negligible, the standard recovery-email approach is probably already pointed at the right problem for your store. How large the overlap is will depend heavily on your own audience and how people actually shop with you: a mobile-heavy niche where research happens on the go will likely see more of it than a desktop-first B2B catalog, so treat this as a check to run on your own numbers, not a result to assume in advance.
Worth separating the two before you spend time or budget trying to win back people who were never actually gone.
It matters because the fix is completely different: a recovery email works on someone who actually cooled off, but it does nothing for someone whose cart simply did not follow them to their next device.