Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Metric

Bounce Rate

Share of sessions that were not engaged (GA4); historically single-page sessions.
AnalyticsmetricUpdated

Formula

Sessions that were not engaged ÷ Sessions × 100

Unit: % · read in context

Definition

In GA4, a Session that is not engaged is a bounce. A session is engaged if it lasted beyond the documented engagement threshold, included a key event, or included at least two page or screen views; a bounce meets none of those conditions. Bounce rate is the share of sessions that were bounces, which makes it the exact complement of engagement rate.

Sessions that were not engaged ÷ Sessions × 100(%)

  • Sessions that were not engaged are visits that were short, contained a single page or screen view and fired no key event.
  • Sessions are all visits in the same period and segment.

The older single-page definition

Earlier analytics tools defined a bounce as a single-page session, regardless of how long the visitor spent on that page. Under that rule a person who read a long article for several minutes and then left was a bounce; under the current rule the same visit is engaged because of its length. Figures from the two definitions are not comparable, which matters whenever a report sets a current rate against a historical one or against a tool that still uses the old rule. The GA4 reports and dimensions guide explains how the current engagement conditions are defined.

Not to be confused with email bounces

In email, a bounce is a message the receiving server would not accept. A hard bounce is a permanent failure, such as an address that does not exist; a soft bounce is a temporary one, such as a full mailbox or a server that is briefly unavailable. Email bounce rate is a list-quality and deliverability measure and has nothing to do with what people did on a web page. The email deliverability guide covers why bounced addresses must be suppressed to protect sender reputation.

How to interpret it

Direction is context: the purpose of the page decides whether a high rate is bad.

  • Pages meant to lead somewhere. A Landing Page built to start a purchase, booking or enquiry should move people on or convert them. A high bounce rate there is a warning that the page is not keeping the promise that brought people to it.
  • Pages meant to answer a question. A page of opening hours, a contact number or a short definition can satisfy the visitor in seconds. A high rate there may simply mean the answer was easy to find.
  • By source. The same page can bounce differently for visitors from search, paid social and email, because they arrive with different intent.

Bounce rate is best used as a guardrail rather than a goal: a check that a change has not driven people away, read beside the outcome the page exists for. The landing page headline test uses it in exactly that role.

Illustrative example — not a real result

An eyewear shop reads bounce by page purpose

Pellucid Eyewear sees a site-wide bounce rate of 47% and is told to bring it down. Split by page, the store-hours page bounces at 78%, and 85% of those sessions come from people searching for the shop's opening times; walk-in visits on those days are steady. The frame-finder landing page also bounces at 78%, but it exists to start an online fitting, and only 1.2% of its sessions do. The team leaves the store-hours page alone and rewrites the frame-finder page so the fitting tool appears above the fold. Its bounce rate falls to 52% and fittings started rise to 3.0% of sessions.

What moves it

  • Match between the link and the page. Visitors who find what the ad, result or email promised stay; those who do not leave at once.
  • Page speed. A slow page loses people before the engagement threshold is reached.
  • Clarity of the next step. Pages with an obvious next action keep more visitors moving.
  • Traffic mix. Sources with weak intent, such as broad display or accidental clicks on mobile, raise the rate without any change to the page.
  • Tracking configuration. Key events, duplicate tags and single-page app routing all change whether a session is recorded as engaged, a set of errors the GA4 common mistakes guide lists.

Engagement Rate is the same measurement seen from the other side; in current Google Analytics the two always sum to the whole, so reporting both adds nothing. Session is the unit being classified, and Landing Page is the dimension that makes bounce rate useful, because the page's purpose sets the direction. The landing pages section covers the page-level work that changes how many visitors move on, and the GA4 section shows where the metric sits in the reports.

What good looks like

There is no reliable universal benchmark, and the site does not cite one. A good rate for a booking page and a good rate for an opening-hours page are different numbers, and published averages blend sites, page types and definitions that are not alike. Judge each page against its own history and against its purpose: if the page exists to lead somewhere, pair the rate with the action it should produce; if it exists to answer, check that the answer is there and accept that people leave once they have it.

Common mistakes

  • Comparing current bounce rates with single-page figures from an older tool or an earlier period.
  • Setting a site-wide target when the right direction differs by page.
  • Treating every bounce as a failure on pages that answer a question at once.
  • Reporting bounce rate and engagement rate together as if they were two findings.
  • Confusing web bounces with email bounces, which measure undeliverable messages.
  • Chasing the rate down with tricks, such as forcing a second page view, that change the number but not the outcome.

More terms to look up

Where this term is used

Sources

  1. Google Analytics Help — Google Analytics Help

Related