Three numbers get mixed up constantly, and they come from different places. Bounce rate and engagement time are GA4 metrics with exact definitions that Google publishes. Dwell time is not a GA4 metric at all: it is the time between a click on a search result and the return to the results page, and only the search engine sees it. Most bad advice about these metrics starts by treating the three as one.
Engaged sessions: the definition everything else rests on
GA4 builds both bounce rate and engagement rate on one idea, the engaged session. According to Google's documentation, a session is engaged when it does at least one of three things:
- lasts longer than 10 seconds;
- has a key event (what GA4 used to call a conversion);
- has 2 or more page or screen views.
The 10 seconds is a property setting, not a constant. It can be changed under Admin, Data Streams, your web stream, Configure tag settings, Adjust session timeout. If two properties disagree on engagement, check that setting before anything else.
Bounce rate
In GA4, "the bounce rate is the percentage of sessions that were not engaged" (source). It is the exact complement of engagement rate: a property with a 62% engagement rate has a 38% bounce rate, by definition.
That makes it a different metric from the bounce rate of Universal Analytics, which counted single-page sessions regardless of how long they lasted. Under the old definition, a visitor who read one article for ten minutes and left was a bounce. Under GA4's, that visit is engaged. Standard Universal Analytics properties stopped processing data on 1 July 2023, so any comparison across that date compares two definitions, not two periods.
There is no benchmark worth quoting here. A "good" bounce rate depends on what the page is for, and the published tables by industry rarely say how they were measured or which definition they used. Compare a page with itself over time, and with pages of the same kind on your own site.
Engagement time
GA4 counts engagement time while your page is in focus or your app is in the foreground. The time is sent in an engagement_time_msec parameter when the page loses focus, the visitor navigates away or the app goes to the background (source). A tab left open behind other windows does not accumulate time.
Two averages are built from it, and they are easy to confuse:
- Average engagement time is total engagement time divided by active users (source). It appears per page in Reports, Engagement, Pages and screens.
- Average engagement time per session is the same total divided by sessions (source), shown in Reports, Engagement, Engagement overview.
A visitor who comes back three times raises the first number more than the second. If one moved and the other did not, returning visitors are the likely reason.
This replaced the old "time on page", which Universal Analytics calculated from the gap between two page views and so could never measure the last page of a visit. Numbers from the two systems are not comparable either.
Dwell time
Dwell time is a search engine's view of a visit: a searcher clicks your result, and either comes back to the results page after some interval or does not come back at all. The click and the return both happen on Google's pages, which is why Google can see the interval and your analytics cannot. No GA4 report contains it, and no tool that reads your analytics can show it to you.
Whether and how Google uses it is a separate question, covered in do clicks affect Google rankings. The short version: click data is used, the 2024 documentation leak names attributes for long clicks, and nobody outside Google knows the thresholds or the weight. Tables that classify clicks as "short" under 10 seconds and "long" over 30 are invented.
Side by side
| Bounce rate | Engagement time | Dwell time | |
|---|---|---|---|
| Who measures it | GA4, in your property | GA4, in your property | The search engine |
| Visible to you | Yes | Yes | No |
| Definition published | Yes, by Google | Yes, by Google | No |
| Which visits it covers | All sessions | All sessions | Visits from search results only |
| Changes when you change the 10-second setting | Yes | No | No |
Reading them without fooling yourself
Segment before you judge. Engagement metrics averaged across every source describe your traffic mix as much as your pages. A campaign of paid social visits, or automated traffic from a tool like ours, moves the averages without saying anything about how readers respond to the text. Look at organic search sessions on their own when the question is about content.
Judge a page against its job. A page that answers a quick question well can have short engagement time and a high bounce rate and still be the best result. A long guide with the same numbers has a problem.
Do not optimize GA4 metrics to influence rankings. Your analytics data is not what Google ranks on. If engagement in GA4 improves because the page got better, the search side is likely to follow, because it is reacting to the same visitors. If it improves because the measurement changed, nothing else does.
