An engaged session is a visit that GA4 treats as meaningful because it met at least one of three conditions: it lasted longer than 10 seconds, it included a key event, or it included two or more page or screen views. Any session that meets none of them is counted as not engaged.
How an engaged session works
GA4 checks each session against the three conditions as events arrive, and flags it as engaged as soon as one is met. The 10 seconds are measured as engagement time, which only runs while the page is in the foreground and the browser tab is active. A tab left open behind other windows does not count.
Several metrics are built on that flag:
- Engaged sessions: the number of sessions that qualified.
- Engagement rate: engaged sessions divided by all sessions.
- Bounce rate: in GA4, simply the share of sessions that were not engaged, so engagement rate and bounce rate always add up to 100%.
- Engaged sessions per user: how often the average visitor came and engaged during the period.
The 10-second threshold can be raised, up to 60 seconds, in the web data stream’s tag settings, under the session timeout options. A change only applies from that day forward.
Why it matters
For a UK service business, the engaged session is a fairer first filter than the old Universal Analytics bounce rate, which counted a visitor who read one page for five minutes and then phoned you as a bounce. Under GA4’s definition, that visitor is engaged, because they stayed longer than 10 seconds.
It is most useful for comparison. Set engagement rate side by side across channels and you can spot traffic that never had a chance: a display placement where almost every visit lasts a few seconds, or a social campaign whose visitors leave before the page has loaded. Compare it across landing pages and you can see which ones hold the people your ads send. A paid search landing page with an engagement rate far below the rest of the site deserves a look before its budget goes up.
Common mistakes
- Treating engaged as qualified. Ten seconds is a low bar. An engaged session tells you the visit was not a dead end, not that it had value. Judge quality on key events such as enquiries and sales.
- A key event that fires on every visit. If page_view or session_start has been marked as a key event, every session counts as engaged and the metric means nothing.
- Double page views. A site that sends two page_view events per page, often from a plugin plus Tag Manager, makes every one-page visit look like two pages, and therefore engaged.
- Changing the timer mid-year. Moving from 10 to 30 seconds lowers the engagement rate overnight, and comparisons across that date become meaningless.
- Comparing with Universal Analytics. The definitions differ, so an old bounce rate and a new engagement rate cannot be compared directly.
How to act on it
- Make sure engagement rate and engaged sessions appear in your traffic acquisition and landing page reports; add them through report customisation if not.
- Check the basics in DebugView: one page_view per page, and no key event that fires automatically on load.
- Look for channels or campaigns with an engagement rate well below the rest, and landing pages people leave within seconds. Check load speed on mobile and whether the page matches what the ad promised.
- Leave the timer at 10 seconds unless you have a clear reason, such as a site where visitors routinely spend 20 seconds reading a single answer.
I use engagement rate to sanity-check paid traffic in performance marketing, alongside the enquiries and sales that actually pay for it.