Bounce rate is a session metric. It needs a definition of "a session" and a definition of "they did something else." A visitor map that counts page loads and places does not, by itself, produce this number. MapMyVisitors reports visits, countries, pages, and referrers. It does not report bounce rate. The definition still comes up the moment you read any other analytics tool, so it is worth having straight.
The formula
In Universal Analytics, the version of Google Analytics that popularized the term, a bounce was a session with a single pageview. No second page. The rate was bounces divided by sessions.
| Piece | Meaning |
|---|---|
| Session | A group of hits from one browser, ended by inactivity (classically 30 minutes) |
| Bounce | A session whose only hit was the landing page |
| Bounce rate | Bounces ÷ sessions |
A blog post that answers the question on one screen can have an 80% bounce rate and be a success. The reader arrived, read, and left satisfied. A checkout page with an 80% bounce rate is a different sentence: people who needed a second step did not take it.
So the number is not "good" or "bad" until you know what a second hit was supposed to be.
What GA4 changed
GA4 stopped leading with bounce rate and started leading with engagement. An engaged session, roughly, lasts longer than 10 seconds, or includes a conversion, or includes two pageviews. Bounce rate in GA4 came back later as the share of sessions that were not engaged. That is a different cut than "only one page."
If you are comparing a number from an old Universal Analytics export to a GA4 property, do not plot them on one line. You would be graphing two definitions. Name the tool next to the number every time you repeat it.
Other products borrowed the old definition, the new one, or a simpler "single page view" rule of their own. When a dashboard says bounce rate and does not define it, ask before you optimize anything.
How people misuse it
- Averaging the whole site. The homepage, a 4,000-word guide, and a "thank you" page should not share one target. Break the rate out by landing page.
- Chasing a lower number with tricks. A popup that fires a second event will move the metric and teach you nothing about the article.
- Ignoring traffic quality. A paid campaign pointed at the wrong query bounces because the page was not what they wanted. The fix is the campaign, which is a UTM and keyword problem, not a font problem.
- Confusing it with exit rate. Exit rate is the share of views of a page that were the last view in the session. A page can be a common exit and a rare bounce, if people read three articles and leave on that one.
Geography does not explain bounce on its own. A country with a high bounce might be hitting one untranslated page. Pair the rate with the country table if you have both. Tracking visitors by country is the other half of that check. The map will show you the country either way. It will not show you the bounce.
A rate worth acting on
Look at bounce rate, or engagement rate, for one landing page, for one source, over a few weeks. Compare it to the same page's other sources. If email readers continue and paid readers do not, the page is fine and the ad is not. If every source dies on the same page, read the page on a phone and see what is missing.
If you do not have sessions at all, do not invent a bounce rate from a counter. Count the pages and the countries you do have, and use a suite when the session question becomes the one you need. Google Analytics alternatives is the comparison for that choice.
