Definition:
The abandonment rate is the percentage of initiated processes that do not reach a defined completion point. It is used to measure interruptions in purchases, forms, registrations, calls, media plays, and other journeys with a recognizable start and expected outcome.
There is no single abandonment rate that applies to every activity. Its meaning depends on what counts as a start, which event represents completion, what units are counted, and over what period. Abandoning a cart, canceling a subscription, and ending a call before an agent answers describe different behaviors and require their own calculations.
The metric complements completion or conversion rate. When both use exactly the same set of processes, their percentages add up to 100%. This relationship no longer holds when users are compared with sessions, periods differ, or stages do not share the same denominator.
How abandonment rate is calculated
For a process with an unambiguous start and completion, the general formula is:
Abandonment rate = (processes started minus processes completed) / processes started × 100
It can also be expressed as 100% minus completion rate, provided that both metrics use the same definition and population. If 1,000 processes begin and 650 finish, there are 350 abandonments and the resulting rate is 35%.
The calculation is comparable only when measurement criteria remain consistent. Counting carts, sessions, users, or orders produces different results. One person may start the same process several times, complete it on another device, or return during a later period. The measurement system must determine whether those actions are joined or treated as independent attempts.
Processes that remain open when a reporting period closes also need a rule. Treating them immediately as abandoned can inflate the rate; waiting indefinitely prevents the calculation from closing. A time window consistent with the normal duration of the journey is therefore usually applied.
Types of abandonment
The metric takes different definitions according to the process being observed. Common uses include the following:
- Cart abandonment: the proportion of electronic carts created, or sessions with products added, that do not generate a purchase. The definition changes depending on whether the start occurs when the first product is added or when the cart is opened.
- Checkout abandonment: initiated checkouts that do not end in a transaction. It is a later stage than the cart and normally has a smaller denominator.
- Form or registration abandonment: forms that are started but not submitted, or registration processes that do not reach confirmation. It requires recording the initial interaction, not merely the page load.
- Call abandonment: in a contact center, incoming calls that end before the caller is assisted. Some measurements exclude calls disconnected within the first few seconds.
- Task or content abandonment: this can apply to configurations, courses, videos, or multistep journeys. Completion must be defined according to the task rather than an arbitrary duration.
- Customer cancellation: churn measures customers or subscriptions lost during a period. Although it is sometimes called abandonment, it uses a customer base and cancellations rather than started and completed processes.
These rates should not be combined as if they represented the same behavior. A person may abandon a cart, return later, and buy; a customer who cancels ends an existing relationship. Their unit, time window, and objective differ.
Instrumentation and measurement quality
In a digital environment, the rate depends on the events that identify each stage. An ecommerce implementation may record add_to_cart, begin_checkout, and purchase; a form needs start and submission signals; a contact center obtains call states from its telephone system.
If the starting event is missing, only completions will be known and abandonment cannot be calculated. If an event fires twice, a confirmation page fails to load, or blockers prevent data from being sent, the rate can change even when actual behavior has not.
Identification also matters. Cookies, consent, sessions, accounts, customer management systems, and payment platforms may count differently. A purchase started on mobile and completed on a computer may appear as a separate abandonment and conversion when the sources do not share an identifier.
When analyzing a conversion funnel, abandonment can be calculated between consecutive stages rather than only between the first and final step. This distinguishes a loss while entering details from one that occurs when choosing delivery or authorizing payment.
How the metric is interpreted
A high rate shows that many processes do not reach the defined completion, but it does not explain the cause by itself. Abandonment may relate to technical errors, waiting times, costs revealed late, unexpected requirements, unavailable payment methods, or privacy concerns. It can also reflect normal behavior such as comparing prices, saving products, or starting a task to continue later.
Context determines whether the result is problematic. Comparisons should use equivalent periods and segments, such as device, channel, country, product, customer type, or stage. A change in the offer, instrumentation, or traffic composition can alter the metric without the process having deteriorated.
Conversion rate measures the proportion that reaches a target action. It can complement abandonment when both calculations share the same start and finish, but not when conversion uses all visitors while abandonment includes only those who started a task.
Bounce rate is not a general abandonment metric either. In Google Analytics 4, it is the percentage of sessions that were not engaged sessions. This definition does not automatically represent abandonment of a cart, form, or purchase.
As a KPI, abandonment rate needs its formula, source, window, and unit preserved alongside the result. Only then can it be compared with previous periods and linked to actual changes in the measured process.