Definition
A cart is created when a shopper adds the first item to it. An abandoned cart is one that was created and not converted into an order within the window the business chooses, such as the same session or a set number of days. The rate divides abandoned carts by all carts created.
(Carts created − Orders completed) ÷ Carts created × 100(%)
- Carts created are distinct carts that received at least one item in the period. Counting by cart rather than by add-to-cart event keeps a shopper who adds five items from counting five times.
- Orders completed are orders placed from those carts. An order placed from a cart created in an earlier period should be attributed to its cart, or the rate will swing at period boundaries.
- The window decides when a cart counts as abandoned. State it, because a same-session rule produces a higher rate than a rule that waits several days.
Checkout abandonment
Checkout abandonment is a narrower measure: of the shoppers who began checkout, how many did not finish. It reduces, though it does not remove, the browsers who never intended to buy that day, and it can be split step by step (delivery details, delivery options, payment, review). Most of the problems a retailer can fix show up here rather than in the cart-level rate. The checkout optimisation guide works through those steps one at a time.
How to interpret it
Lower is better, but not all abandonment is failure. Shoppers use carts as wish lists, check the total including delivery, compare with other shops and come back later. The useful readings are:
- Cart stage versus checkout steps. A high cart-stage rate with a low checkout drop-off suggests browsing behaviour or price comparison; a sharp drop at one checkout step points to a specific obstacle.
- By device. Mobile shoppers often abandon at steps that demand typing, such as address and card details, where desktop shoppers do not.
- By traffic source and landing page. Shoppers arriving on a deal-focused Landing PageThe page a visitor arrives on from a campaign, built to turn that visit into one action. may add items to see the price and leave; comparing sources shows whether the rate reflects intent or the checkout.
- Over time for one step. A step whose drop-off jumps after a site change is the first place to look for a broken field or payment method.
The CRO funnel section places cart and checkout among the other stages where shoppers can be lost.
A candle shop moves the delivery cost forward
Tamsin Lane Candles creates 5,000 carts in a month and takes 1,250 orders, a cart abandonment rate of 75%. Split by checkout step, mobile shoppers drop out most at the delivery-options step, where a £4.95 delivery charge first appears: 38% of mobile checkouts end there, against 14% on desktop. The shop shows the delivery cost and the free-delivery threshold on every product page. The next month, 5,100 carts produce 1,530 orders and the rate falls to 70%. Part of that fall may be seasonal browsing rather than the change, so the shop checks whether drop-off at the delivery-options step, the step the change targeted, fell by more than at the other steps before crediting it. Its abandoned-cart email is credited only with the 60 orders it recovered above a holdout group that received no email, not with the 210 orders its recipients placed.
What moves it
- Surprise costs. Delivery charges, taxes or fees revealed only at checkout are a common reason to leave; showing them early lowers abandonment. The free shipping threshold experiment tests one version of that remedy.
- Forced account creation. Requiring a password before purchase loses shoppers who would have checked out as guests.
- Payment options. Missing the method a shopper expected, or a method that fails on one browser or device, ends checkouts.
- Speed and form length. Slow pages and long forms add friction, especially on mobile.
- Trust signals. Unclear returns policies and unfamiliar payment pages make shoppers hesitate.
- Shopping intent. Seasonal browsing, gift research and price comparison raise the rate without any fault in the checkout.
Recovery and how to credit it
Abandoned-cart emails and SMS messages remind a shopper of what they left and sometimes add an incentive. The abandoned cart flows guide covers timing and content, and the ecommerce email and SMS flows guide sets them among the other automated messages a shop runs; the wider email automation section explains how such flows are triggered. The crediting question is the one that is most often got wrong: many shoppers who receive the reminder would have returned anyway. Counting every order placed after the email overstates its effect. Holding back a random share of abandoned carts from the flow, and crediting the flow only with the difference in orders, measures what it actually recovered.
Related metrics
Add-to-Cart RateShare of sessions in which a product was added to the cart. is the step before: it measures how many sessions create carts at all, so the two together split a weak conversion rate into a product-page problem and a checkout problem. CVRShare of visits or clicks that convert. is the overall outcome both feed. AOVAverage revenue per order. matters because abandoned carts often differ in value from completed ones; a fix that recovers small carts and loses large ones can lower revenue while improving the rate.
What good looks like
There is no reliable universal benchmark, and the site does not cite one. The rate depends on the category, the price point, the share of mobile traffic, the window used and how much browsing the shop attracts, and published figures rarely state any of these. Judge the rate against your own history, step by step and by device, and judge each fix by the orders it adds rather than by the percentage it removes.
Common mistakes
- Treating all abandonment as lost sales, when much of it is browsing.
- Reading only the cart-level rate instead of drop-off at each checkout step.
- Counting add-to-cart events instead of carts, which inflates the base.
- Crediting recovery flows with every later order, including those that would have happened anyway.
- Not stating the window, so same-session and multi-day rates are compared.
- Discounting in every reminder, which teaches shoppers to abandon carts to wait for a code.
