GA4 Bounce Rate vs Engagement Rate: Definitions and Setup

In Google Analytics 4, bounce rate is the percentage of sessions that were not engaged, so it is exactly the inverse of engagement rate: bounce rate = 1 − engagement rate. Google's GA4 documentation defines an engaged session as one that lasts longer than 10 seconds, has a key event, or has 2 or more page or screen views. That makes the GA4 number a different metric from the Universal Analytics bounce rate, which counted single-page sessions, and the two should never be compared directly.

What is the Google Analytics 4 bounce rate definition?

Google's help article "[GA4] Engagement rate and bounce rate" defines bounce rate as the percentage of sessions that were not engaged.

Engaged session criteria

According to the same Google article, an engaged session is a session that meets any of the following criteria:

• Lasts longer than 10 seconds (the 10 seconds is the default and can be changed, see below).
• Has a key event (what GA4 used to call a conversion).
• Has 2 or more screen or page views.

The criteria are alternatives, not a checklist. A visitor who reads one article for two minutes and leaves is engaged. A visitor who opens two pages in four seconds is engaged. A visitor who lands, waits seven seconds and closes the tab, with no key event, is a bounce.

The formulas

• Engagement rate = engaged sessions ÷ sessions
• Bounce rate = (sessions − engaged sessions) ÷ sessions = 1 − engagement rate
• Average engagement time = total time the page was in focus (or the app in the foreground) ÷ active users

Google's "User engagement" help article explains that GA4 measures engagement time in milliseconds through the engagement_time_msec parameter, which is sent with events and with the automatic user_engagement event when someone leaves a page. Time with the tab in the background does not count. That is why average engagement time is usually lower than the "session duration" people remember from Universal Analytics.

Is GA4 bounce rate the inverse of engaged sessions? A worked example

Yes. Bounce rate is not computed from exits or single-page sessions; it is simply the share of sessions that failed every engagement test. Here is a worked example for one landing page over 30 days:

Metric	Value	Calculation
Sessions	2,000	From the report
Engaged sessions	1,240	From the report
Engagement rate	62.0%	1,240 ÷ 2,000
Bounced (non-engaged) sessions	760	2,000 − 1,240
Bounce rate	38.0%	760 ÷ 2,000, or 100% − 62%
Sessions with a key event	90	From the report
Session key event rate	4.5%	90 ÷ 2,000

The two rates always add up to 100%, so showing both adds no information. The key event rate (4.5%) tells a very different story: most engaged visitors still did not do what the page exists for.

How does GA4 bounce rate differ from Universal Analytics?

Google's Universal Analytics help article on bounce rate defined a bounce as a single-page session: a session that triggered only a single request to the Analytics server. GA4 replaced that with the engagement logic above. The same visit can be classified differently by the two tools:

Session	Universal Analytics	GA4
One page, read for 3 minutes, no other hits	Bounce	Engaged (longer than 10 seconds)
Two pages viewed in 5 seconds	Not a bounce	Engaged (2+ page views)
One page, 6 seconds, clicked a tracked button (interaction event, not a key event)	Not a bounce	Bounce
One page, 4 seconds, submitted a form marked as a key event	Not a bounce (if the submit was tracked)	Engaged (key event)

Content pages where readers read one article and left often look much better in GA4. That is a change of definition, not a better page. Do not put a UA bounce rate and a GA4 bounce rate on the same chart, and do not report a "drop in bounce rate" that coincides with the migration.

How to add bounce rate in GA4 reports

Bounce rate exists in GA4 but is hidden from most standard reports by default. Google's help documentation notes that you can only add metrics to a detail report (a table), not an overview report, and that you need the Editor or Administrator role.

Add bounce rate and engagement rate to a standard report

1. Open a detail report, for example Reports > Engagement > Landing page or Pages and screens.
2. Click the pencil icon, Customize report, in the top right.
3. Under Report data, click Metrics. If you only see "Add cards", you are in an overview report; go back and open a detail report.
4. Click Add metric, type "Bounce rate", and select it. Repeat for "Engagement rate", "Average engagement time per session" and "Session key event rate" if they are missing.
5. Drag the metrics into a sensible order, click Apply, then Save > Save changes to current report (or save as a new report so the default stays intact).
6. To make a new report visible in the left navigation, open Library, edit a collection, add the report to a topic, save, and use More > Publish on the collection card.

Add bounce rate to an Exploration

1. Go to Explore and start a Free form exploration.
2. In the Variables column, click + next to Dimensions and import Landing page + query string, Session default channel group and Device category.
3. Click + next to Metrics and import Sessions, Engaged sessions, Engagement rate, Bounce rate, Average engagement time per session and Key events.
4. Drag Landing page into Rows and Device category into Columns (or use it as a breakdown), and drag the metrics into Values.
5. Add a filter such as Sessions > 100 by using a metric filter, so low-traffic pages with noisy rates do not top the list.

The same workspace pairs well with a GA4 funnel exploration for checkout drop-off.

How to change the 10-second engaged session timer (and when you should)

The timer lives in the web data stream's tag settings:

1. Open Admin > Data collection and modification > Data streams and click your web stream.
2. Click Configure tag settings, then Show more (or "Show all") under Settings.
3. Choose Adjust session timeout.
4. Under Adjust timer for engaged sessions, pick a value. The setting accepts values between 10 and 60 seconds. The session timeout itself (30 minutes by default) is a separate field on the same screen.
5. Save. The change only applies to data collected from that point on; historical data is not reprocessed.

Decision rule for changing it:

• Raise it when 10 seconds is meaningless for your content. On long-form articles or detailed product pages, nobody gets value in 11 seconds, so a 30 or 60 second threshold separates readers from skimmers.
• Keep the default when you want comparability with past data, with other properties you manage, or with GA4's peer benchmarks, which use the standard definition.
• Never change it mid-test or mid-campaign. Annotate the date you change it, because every engagement and bounce figure shifts on that day.

An alternative is to keep 10 seconds and add a meaningful key event, such as a "read 75%" scroll event. See the guide to tracking scroll depth in GA4.

What is a good engagement rate in GA4?

There is no universal good number. Engagement rate depends on page type and traffic source:

• A store locator or contact page answers the question in a few seconds. A low engagement rate there can mean the page worked.
• A blog post gets most of its engaged sessions from the 10-second rule, so it tends to look healthy even when nobody clicks onward.
• A paid-social landing page receives cold, accidental and in-app-browser traffic, so it will usually trail an email landing page that receives warm subscribers.

How to set your own baseline

1. Pull 8 to 12 weeks of landing-page data in an Exploration, excluding pages with fewer than 100 sessions.
2. Tag each landing page by type: home, category, product, article, lead-gen landing page, utility.
3. For each type, calculate the median engagement rate and the 25th percentile. Pages below the 25th percentile of their own type are your review list.
4. Repeat the calculation by channel and device, because a page can be fine on desktop organic and broken on mobile paid.
5. Re-run monthly with the same date length, same reporting identity and same timer setting.

Worked example: your eight article pages have engagement rates of 48%, 52%, 55%, 58%, 60%, 63%, 66% and 71%. The median is (58 + 60) ÷ 2 = 59%, and the 25th percentile sits around 53.5% (between 52% and 55%). An article at 48% is the one to investigate first. A product page at 48% is a separate question and should be compared against other product pages only.

If you want an outside reference, GA4 itself offers a Benchmarking feature. Google's "[GA4] Benchmarking" help article explains that it shows the median, 25th and 75th percentiles for a peer group of properties in your industry category, for metrics including engagement rate and average engagement time per session. That is the only benchmark worth using, because it is calculated with the same definitions as your own data. The same reasoning applies to conversion rates, as covered in why generic conversion rate benchmarks mislead.

How to use engagement rate and bounce rate to judge landing pages

Engagement rate tells you where to look, not what to change.

Segment before you conclude

Break every page-level rate down by at least:

• Session source / medium or default channel group. If paid search engages at 35% and organic at 65% on the same page, the problem is probably the match between ad and page, not the page itself. See how ad scent affects landing page conversion.
• Device category. A large gap between desktop and mobile usually points to layout, speed or above-the-fold content on mobile.

Pair it with key events

Read engagement rate next to session key event rate. The combinations tell you different things:

Engagement rate	Key event rate	Likely reading	Next step
Low	Low	Wrong audience or weak first impression	Check traffic source match and the first screen
High	Low	People read but are not convinced or cannot find the action	Audit the call to action, offer and friction
Low	High	Fast, decisive visitors (often brand or returning traffic)	Leave it alone; do not add clutter
High	High	Page works	Test incremental improvements

Key events only help if they are tracked properly. If your forms, sign-ups or add-to-cart clicks are not yet events, start with the GA4 custom events tracking guide.

Watch the sample size

On a page with 40 sessions, two extra bounced sessions move the bounce rate by 5 percentage points (2 ÷ 40). Do not react to weekly swings on low-traffic pages; widen the date range until each segment has at least a few hundred sessions.

Turn the list into fixes

GA4 tells you which pages underperform, not why. Once you have a short list, use qualitative tools: session recordings alongside GA4 to watch real visits, or an attention heatmap to check whether the first screen draws the eye to the right element. EyeCaptain is not an analytics tool and does not replace GA4; it audits a page from its URL on desktop and mobile and returns findings ranked by impact, which is a quick way to get hypotheses for a low-engagement page. You can try it with a free one-page CRO audit. Then rank the fixes with a CRO prioritization framework like ICE or PXL before changing anything.

Common mistakes with GA4 engagement rate and bounce rate

Marking page_view (or another automatic event) as a key event

Because "has a key event" makes a session engaged, marking page_view, session_start or scroll as a key event pushes engagement rate towards 100% and bounce rate towards 0%. Only mark events that represent real business outcomes.

Misreading single-page sites and single-page apps

On a one-page site, the 2-page-view rule can never fire, so engagement relies on the timer and key events. Add key events for the actions that matter (form submit, click-to-call, booking start). On single-page apps, the opposite can happen: enhanced measurement can record a page view on every browser history change, so virtual page views inflate engagement. Check in DebugView how many page_view events one real visit produces.

Ignoring consent mode

With basic consent mode, GA4 receives nothing from visitors who decline analytics cookies, so your engagement rate describes only consenting visitors. With advanced consent mode, Google's "Behavioral modeling for consent mode" documentation explains that GA4 can model the behaviour of users who decline cookies from similar users who accept, once a property meets the eligibility thresholds, and that the modeled data appears when the reporting identity is set to Blended. Compare periods only when the consent banner, consent mode type and reporting identity were the same, and expect the BigQuery export (which has no modeled data) to differ from the interface. The consent mode v2 guide covers the setup.

Assuming bots are fully filtered

GA4 automatically excludes traffic from known bots and spiders, and there is no setting to turn that off. Unknown bots and scrapers still get through, and they typically create one-page, zero-second, non-engaged sessions that drag engagement rate down. Look for spikes from a single city, a "(not set)" source or an odd screen resolution, and exclude them with a segment. Also define internal traffic and activate the internal traffic data filter.

Optimizing the metric instead of the outcome

Auto-playing video or splitting articles across pages raise engagement rate without helping anyone. Use engagement rate to find problems and key events to judge the fix, ideally with a test as described in the A/B testing guide.

Frequently asked questions

What is bounce rate in GA4?

Bounce rate in GA4 is the percentage of sessions that were not engaged. A session is engaged if it lasts longer than 10 seconds, has a key event, or has 2 or more page or screen views. Bounce rate therefore equals 100% minus engagement rate.

Why is my GA4 bounce rate lower than it was in Universal Analytics?

Universal Analytics counted any single-page session as a bounce, even if the visitor read for several minutes. GA4 counts that visit as engaged once it passes 10 seconds. The lower number reflects the new definition, not a better page.

What are engaged sessions in GA4?

Engaged sessions are sessions that meet at least one of three criteria: longer than 10 seconds, at least one key event, or at least two page or screen views. Engagement rate is engaged sessions divided by all sessions.

What is a good engagement rate in GA4?

It depends on page type, traffic source and device, so there is no universal target. Build a baseline from your own pages of the same type and review those below the 25th percentile. For an external reference, use GA4's built-in Benchmarking feature, which compares you with peers using the same definitions.

What does average engagement time measure?

Average engagement time is the total time your site was in focus in the browser (or your app was in the foreground), divided by active users. Time spent in a background tab is not counted, so it is usually shorter than session duration.

Should I change the engaged session timer from 10 seconds?

Raise it, up to 60 seconds, only if 10 seconds is too short to mean anything for your content, such as long articles. Keep the default if you need comparability with past data or GA4 benchmarks, and annotate the date of any change because it is not applied retroactively.