Bounce Rate

In Universal Analytics, the version of Google Analytics common until 2023, bounce rate was clearly defined: the share of sessions in which someone viewed exactly one page and then triggered no further action — no click, no second pageview, no form. The count depended purely on the number of hits recorded within a session, regardless of how long someone stayed on the page.

With the successor version, Google Analytics 4 (GA4), the logic was swapped out. Instead of counting drop-offs, GA4 first counts engaged sessions: a session counts as engaged if it lasts at least ten seconds, or contains a conversion event, or comprises at least two page or screen views. From that figure comes the engagement rate — the share of engaged sessions among all sessions.

This is exactly where the common confusion arises: many reports and dashboards keep talking about 'bounce rate' out of habit, even when the data comes from GA4. Google did bring the term back at users' request, but there it's simply calculated as the counterpart to the engagement rate — a session lacking the three criteria above. That's a different calculation than in Universal Analytics, not the same metric under a new name.

The reason for the switch lies in an old measurement flaw: someone who reads a long guide article for two minutes and then leaves the page without clicking counted as a bounce in Universal Analytics — even though the page fully served its purpose. GA4 correctly recognizes the same session as engaged via the ten-second rule, because time on page alone is already an engagement signal.

In practice that means: comparing 'bounce rate 2022 versus bounce rate 2024' across the switch from Universal Analytics to GA4 compares two differently defined metrics, not the same measurement at two points in time. Teams sometimes wrongly conclude that content quality changed, when in fact only the measurement method changed — Universal Analytics was discontinued for standard accounts on July 1, 2023, and for the 360 version on July 1, 2024.

That's why we look first at the engagement rate together with average engagement time and specific conversion events, rather than at the back-calculated bounce rate alone, and we always state which system — Universal Analytics or GA4 — a figure comes from. A high single-page rate is therefore not automatically a bad sign if the click-through rate beforehand was already sound and the content served its purpose.