GA4 bounce rate is the percentage of sessions that were not engaged. It is an analytics measure to interpret alongside the page’s purpose and measurement setup, not a published Google Search ranking cutoff. A high number does not, by itself, demonstrate a search penalty or a dissatisfied visitor.
What counts as an engaged session?
Google’s engagement and bounce-rate help defines an engaged session as meeting any of these conditions: lasting longer than 10 seconds, having a key event, or having at least two page or screen views. The time condition is “longer than,” not “at least”; the conditions are alternatives, not requirements that must all be met.
For web streams, the session settings allow the engaged-session timer to be adjusted. Check the actual setting before interpreting a property’s results. That timer and the inactivity timeout that ends a session are different settings.
Google also excludes events such as first_visit, first_open and session_start from engaged-session calculations even if marked as key events. Do not assume that marking any automatically collected event makes every visit engaged.
A short visit can still answer a user’s question. Conversely, a session meeting a time or event rule is not proof of satisfaction or a qualified lead. Use real outcome definitions rather than changing event settings merely to make a percentage look better.
Calculate it with the same population and denominator
Engagement rate = engaged sessions ÷ total sessions × 100%. Bounce rate = non-engaged sessions ÷ total sessions × 100%, or 100% minus engagement rate for the same reporting scope. Published rounded percentages may show minor rounding differences.
When combining groups, use the underlying session counts. This second fictional example assumes two non-overlapping groups that together cover the entire population under the same definition:
| Group | Sessions | Engaged | Not engaged | Bounce rate |
|---|---|---|---|---|
| A | 100 | 60 | 40 | 40% |
| B | 300 | 270 | 30 | 10% |
| Combined | 400 | 330 | 70 | 17.5% |
The combined bounce rate is 70 ÷ 400 × 100% = 17.5%, not the simple average of 40% and 10% (25%). Do not combine overlapping groups or incomplete rows as though they cover the whole property. Zero sessions gives no usable rate; missing measurement is not zero bounce.
Add the metrics to a GA4 detail report
- Open Reports and choose a detail report, such as Pages and screens.
- Choose Customize report. Google’s help requires the Editor or Administrator role for this customization.
- Under Report data, open Metrics and Add metric. Search for Engagement rate and Bounce rate; a metric may already be included.
- Apply and save. The added columns may be at the far right of the table.
If you only see Add Cards rather than Metrics, you are in an overview report; the documented metric customization applies to detail reports. These steps follow the linked public help, not an inspection of your private account.
What does this tell you about SEO?
A 2010 Google Search Central post (German translation) stated that Google Analytics was not used for ranking. That is a historical statement predating GA4, not a new GA4-specific announcement or evidence that Search is technically unable to access any Analytics data.
The current page-experience guidance says there is no single page-experience signal, identifies Core Web Vitals as used by ranking systems and does not publish a GA4 bounce-rate target. It also says good scores do not guarantee top rankings. These are different measurements with different purposes.
Investigate a change before trying to lower the number
- Measurement: check the timer, key-event definitions, duplicate events, tagging, consent and the selected report scope. A configuration change can change the metric without changing the page.
- Comparable traffic: separate page types, source/medium, device, language and periods. A new audience can change the aggregate.
- The actual task: read the answer and try the next action on a phone. Distinguish a completed short task from a broken or misleading path.
- Search observations: review Search Console queries and pages separately. It does not provide a one-to-one join between every organic query and a GA4 session.
- Business results: inspect actual orders or qualified enquiries with a consistent attribution and deduplication method.
A lower bounce percentage is not the objective if it comes from counting a trivial event as success. Improve the actual problem and report the definition, period and outcome. For page-purpose examples and practical checks, see the high-bounce diagnosis. For the wider search investigation, use the ranking diagnosis and traffic-drop guide.