Cart abandonment vs. checkout abandonment

A cart contains an item the shopper has selected. Checkout starts when the shopper begins placing the order. The completed-order event needs a consistent definition too. Google Analytics has separate begin_checkout and purchase events, but your implementation still determines when they are sent.

Some reports call every cart without an order “cart abandonment,” including carts that reached checkout. Others use the term only for shoppers who left before checkout. Label that narrower measure “left before checkout” so readers know which group you mean.

For the simple funnel below, count each journey once at each stage: cart, checkout and completed order. Extra item additions, page refreshes and repeated order events should not increase those counts.

In this example, follow the same 100 shopping journeys over one observation period. Of those journeys, 60 start checkout and 45 complete an order. Assume every completed order passed through the tracked cart and checkout steps.

The counts are 40 journeys lost before checkout and 15 during checkout: 55 unfinished journeys in total. Do not add 40% and 25%. The first divides by all 100 cart journeys; the second divides by the 60 that started checkout.

Choose one counting unit and use it throughout. Journeys, users, sessions and orders are not interchangeable. Document how returning shoppers are handled, and keep the group and observation period consistent. Today’s cart starts and today’s purchases may belong to different journeys.

Test a normal order, a refresh, a retry and a return visit. Look for missing or repeated completion events. Separate known test traffic, and note gaps caused by consent choices or journeys that continue on another device.

Keep the event definitions, period and known gaps beside the report. If the numbers change after analytics code changes, first check whether you are measuring the same thing.

For the 40 journeys that stopped before checkout, test the route out of the cart: does the button work, is the total clear, and can the shopper review the items? For the 15 that stopped during checkout, inspect the steps reached and any recorded errors.

The rate tells you where to look, not why someone left. Shoppers may compare options, encounter a technical problem or return later. Use reproducible test journeys and relevant customer feedback to identify the problem.

If a test reveals a delivery-charge error that blocks progress, fix it and repeat the test. Measure the change using the same definitions. Do not treat every unfinished checkout as a payment decline or assume that shortening the form will solve it.