How it works
Bounce rate depends on the analytics system’s definition. In GA4 it is the share of sessions that were not engaged, the inverse of engagement rate. Comparing it with an older system without checking definitions can lead to wrong conclusions. Event and consent configuration also influence what is measured.
A realistic example
A visitor may read a short definition, understand it and leave. That can be a successful informational visit even if it looks weak in one metric. By contrast, someone who immediately abandons a broken form has not completed the intended task.
What to check
Check which definition and reporting period are used. Segment by page purpose, device and source, and inspect the actual interface. Verify event collection before interpreting a change; an event added to the page can alter the rate without improving the experience.
Limits and next steps
A lower bounce rate is not automatically better. Avoid artificial engagement events that merely improve the reported number. Judge the metric alongside useful task completion and qualitative checks, and acknowledge visits that are not measured because of privacy choices.