What Is Bounce Rate? GA4 Formula, Exits and Benchmark Pitfalls

April 12, 2026

Author: Shusaku Yosa

In GA4, bounce rate is the percentage of sessions that were not engaged. Calculate it as non-engaged sessions divided by all sessions, multiplied by 100. For the same reporting scope, bounce rate and engagement rate sum to 100%.

A high bounce rate does not automatically mean a bad page, and a low one does not prove revenue contribution. Align the definition with the page’s intended task before judging the number.

What qualifies as engagement?

The standard criteria are a session lasting longer than ten seconds, a key event, or at least two page or screen views. The time setting can be adjusted. first_visit, first_open and session_start are exceptions excluded from this calculation even if marked as key events.

A session that does not meet an engagement criterion contributes to bounce rate. A one-page visit can still be engaged.

Example	Interpretation
One page read long enough	May meet the duration criterion
Short visit with an eligible key event	May be engaged
Two page views	Meets the view-count criterion
Brief single-page visit without another criterion	Non-engaged

Calculate a fictional example

Suppose there are 80 sessions: 60 engaged and 20 not engaged. Bounce rate is 20 ÷ 80 × 100 = 25%; engagement rate is 75%.

Do not average page percentages without their denominators. If page A has 10 sessions and an 80% bounce rate, and page B has 90 sessions and a 20% bounce rate, the total is 8 + 18 = 26 non-engaged sessions out of 100. The combined rate is 26%, not the simple average of 50%.

Separate bounces from exits and older definitions

Concept	What it describes	Key distinction
GA4 bounce rate	Non-engaged sessions	Uses session engagement criteria
Exit	The final page or screen in a journey	A successful journey also ends somewhere
Single-page visit	Number of pages viewed	Can be engaged in GA4
Universal Analytics bounce	A single-page session without another interaction hit	Not the same definition as GA4

Bounce rate cannot reconstruct the number who abandoned the first screen or read nothing. Do not compare an old UA benchmark directly with current GA4 reporting.

Evaluate a benchmark before using it

Check the year, country, channel, platform, sample and measurement definition. A benchmark without those conditions is not a sound target.

A contact-details page, article, booking page and login screen have different purposes. Prefer comparable pages and acquisition segments within your own site, or the same page over consistent periods, rather than inventing one pass/fail percentage.

Interpret high and low rates

Pattern	Possible explanation	Evidence to inspect
High rate and few outcomes	Intent mismatch or friction	Queries, ads, real-device tasks and forms
High rate on a quick-answer page	The task may finish quickly	Page purpose and reader feedback
Low rate without useful outcomes	Intermediate activity or measurement effects	Key-event selection and duplicates
Sudden movement	Acquisition, tagging or configuration change	Change log and segment trends

An unusually low rate is a reason to investigate, not proof of an error. Likewise, a high rate is not a diagnosis by itself.

For reporting steps, use how to find bounce rate in GA4. When evidence identifies a problem, use the improvement checklist and track business outcomes alongside engagement.

Consistent definitions and a reliable change log come before searching for an industry average.