Choosing an abandonment window: why we start at 30 minutes
Our view
The abandonment window is how long a checkout has to sit untouched before ACR treats it as abandoned. On a campaign’s Lead Source step it appears as “Wait before a checkout counts as abandoned”, in minutes from the shopper’s last activity. New campaigns start at 30, and the field takes anything from 5 minutes to 7 days.
We think 30 minutes suits most stores, and that the good reasons to move it come from your own checkout, your sequence and your other tools. We can’t point you to results that prove a best number, because ACR has none published, and a number found on someone else’s store may not hold on yours. What follows is reasoning you can check against your own data.
What the window starts
When the window ends, ACR starts a run, and the first step’s delay counts from that moment. In the recommended sequence the first step is a call with no delay, so on default settings the window is the time between the shopper leaving and their phone ringing.
Three rules decide when first contact really happens:
- Every update your store sends about the checkout restarts the clock. A shopper who comes back and pays inside the window never starts a run.
- First contact is the window plus the first step’s delay.
- If that moment falls outside the shopper’s allowed hours, the step waits for the next allowed minute, and every later step moves back with it. The recommended schedule allows 8 am to 9 pm in the shopper’s local time, narrowed further by stricter state and country rules and by public holidays.
Why not shorter
Five minutes is allowed, and we would not use it. People stop partway through checkout for ordinary reasons: fetching a card, checking a delivery date, asking someone at home. A very short window can ring their phone while they are still on the payment page, about an order they were about to place. We expect many shoppers to find that intrusive. It also uses one of your plan’s runs, because a run counts once its first contact has gone out.
Shopify separates two decisions that ACR’s window can merge. Its help center lists a checkout as abandoned once it has stayed incomplete for more than ten minutes, and then its recovery email waits for a delay the store picks under “Send after”. Labeling a checkout abandoned is one choice and contacting the shopper is another. In ACR the window makes both at once, unless you give the first step a delay of its own.
Why not longer
A call or a text is worth most while the shopper still remembers the cart, and each extra hour of window gives some of that up. We have no measurement of how quickly that fades for any particular store. It is a judgment, and it is the main reason we prefer a short wait to a long one.
A long window also pushes more shoppers past the end of their allowed hours. Here is when the first call of the recommended sequence would go out for two shoppers, one leaving checkout at 11:00 am and one at 7:10 pm:
| Window | Shopper who left at 11:00 am | Shopper who left at 7:10 pm |
|---|---|---|
| 10 minutes | 11:10 am | 7:20 pm |
| 30 minutes (the default) | 11:30 am | 7:40 pm |
| 2 hours | 1:00 pm | 8:00 am the next day |
| 12 hours | 8:00 am the next day | 8:00 am the next day |
For the evening shopper, a 2-hour window and a 12-hour window give the same answer. For the late-morning shopper, 12 hours turns a same-day call into a next-day one.
A window of several hours also runs into your email flows, which may already be sending by then. Klaviyo’s pre-built abandoned cart flows, for instance, wait 4 hours before their first message, according to its help center. Running recovery next to your email flows looks at that overlap.
Window or first-step delay?
Say you want first contact about an hour after the shopper leaves. You can set a 60-minute window with the call at no delay, or keep the window at 30 minutes and make the call wait 30 more. The shopper hears from you at the same moment either way. We prefer the second, because the window belongs to the lead source and the delay belongs to the campaign.
Live campaigns share one window
ACR gives each checkout the shortest window among the published campaigns whose trigger is on for that lead source. If one campaign is set to 30 minutes and another to 60, every checkout from that source is abandoned at 30, and whichever campaign takes the shopper starts then.
A/B tests work the same way, since both versions listen to the same source. A test of two windows would run both at the shorter one, so give one version a delay on its first step instead.
When we would move it
- Longer, if your checkout asks for something shoppers have to look up, such as a measurement or a business tax number. That pause is part of buying, and a call in the middle of it interrupts.
- Shorter, if many of your shoppers leave in the last half hour before 9 pm their time and you would rather reach them that evening than the next morning. Check when your abandoned checkouts actually happen before you decide.
- Neither, if only one campaign needs a later start. Leave the window at 30 and give that campaign’s first step a delay.
Whichever way you move it, change that one setting and leave it for a few weeks, so the result belongs to the change. Our post on holdout groups explains how to tell whether recovery added orders at all.
Sources
More in Checkout recovery, or back to all posts.
Read next
- Checkout recovery
Why shoppers leave checkout, and which reasons a conversation can fix
Baymard Institute's abandonment research, split into what a conversation can help with and what only a change to your site will solve.
November 7, 2026Read the post
- Checkout recovery
Discount codes in recovery conversations: our case for offering fewer
How ACR's standard and escalation codes work, the margin a code has to earn back, how codes can teach shoppers to leave checkout, and why a code also decides which campaign gets the credit.
November 7, 2026Read the post