Bounce Rate vs Exit Rate: The Difference (and What GA4 Changed)

The confusion is understandable: both are percentages, both describe people leaving, and both sat side by side in Universal Analytics reports for years. But the 2 metrics answer different questions. Bounce rate judges arrivals: of the people who started their visit here, how many gave up immediately? Exit rate judges departures: of everyone who viewed this page at any point in their visit, how many left the site from it? A page can carry a poor bounce rate and a healthy exit rate at the same time, because the 2 numbers are counting different visitors.

A worked example makes the split concrete. Imagine your pricing page recorded 1,000 views last month. 200 of those views opened a session, and 120 of those arrivals left without engaging: a 60% bounce rate. Meanwhile 300 of the 1,000 views turned out to be the last page of a visit: a 30% exit rate. Same page, same month, 2 different stories. The bounce rate says arrivals from ads and search are not being caught. The exit rate says the page holds people reasonably well once they are already inside the site. Fixing the first screen helps the first number; fixing what the page leads to helps the 2nd.

The 1-line version worth memorizing: every bounce is an exit, but not every exit is a bounce. A bounce is a visit that never got going. An exit is simply the place where a normal visit ended, and every visit ends somewhere.

In Universal Analytics, a bounce was a single-page session: one page viewed, nothing clicked, session over. Time on the page did not matter. Someone who read your entire article for 4 minutes and left counted exactly the same as someone who lasted 2 seconds. GA4 threw that definition out.

Google’s official definition now reads: bounce rate is “the percentage of sessions that were not engaged.” An engaged session is one that lasts longer than 10 seconds, has a key event, or has 2 or more page views. Fail all 3 tests and the session counts as a bounce. A key event, in GA4 vocabulary, is simply an action you have told the tool matters, like a submitted form, a signup, or a purchase.

That redefinition changes what the number means. Reading now counts as engagement even when nothing gets clicked, so a long visit to a single article no longer counts against you, and content pages can report lower bounce rates in GA4 than the same pages did in Universal Analytics. It also makes bounce rate the exact mirror image of GA4’s engagement rate: an engagement rate of 60% means a bounce rate of 40%, always.

The change was not cosmetic. The old definition punished pages for doing their job: a visitor who found the answer, read it, and left registered as a failure. It also broke on modern sites, where a visit can scroll, watch, and interact for minutes without ever loading a 2nd page. Anchoring the metric to engagement instead of page count made it mean something closer to what site owners assumed it meant all along: did this visit get going, or not?

No. As of July 2026, GA4’s official dimensions and metrics reference lists no exit rate and no exits metric of any kind, and Google’s help article defining bounce rate never mentions exit rate either. The metric was not renamed or hidden. It was simply not brought into GA4.

The closest built-in view is a path exploration in the Explore section, which can be read backwards to show the pages where sessions ended. That answers the same underlying question, where do people leave, but it is a report you assemble yourself, not a percentage GA4 calculates for you.

The practical consequence: if a dashboard, an agency report, or a blog post quotes your exit rate today, that number came from Universal Analytics history, from a different analytics tool, or from someone’s own math on exported GA4 data. None of those is wrong, but it is worth knowing the figure no longer comes from Google.

If you need the insight the metric used to give, the workable habit is a monthly check: look at where sessions end for your most important pages in a path exploration, and watch your funnel reports for the specific step where completions drop. Both views name the page that loses people, which is the decision the exit rate was feeding all along.

Benchmarks vary by industry and page type, but the published bands are consistent enough to be useful. Fullstory’s benchmark roundup draws the lines like this:

The spread makes sense once you remember what different visits are for. B2B visitors research and compare across many tabs, blog readers often want a single answer, and B2C stores pull people deeper by design. A blog post at 65% is doing its job; a product landing page at 65% is leaking money.

One honest caveat before you compare yourself to any of these: benchmark tables were largely built on years of Universal Analytics data, and GA4’s redefinition moved the numbers. The same page can report a noticeably lower bounce rate in GA4 than it did in Universal Analytics without a single visitor behaving differently, because long reads now count as engaged. Treat old benchmarks as rough guides and your own trend line as the real signal.

And resist reading one sitewide average as a verdict. A single number blends your blog, your landing pages, and your checkout into a figure that describes none of them. Segment before judging: bounce rate by landing page and by traffic source is where the decisions live. A 70% bounce from one ad campaign next to 35% from organic search is a finding you can act on. A 52% sitewide average is not.

It depends entirely on the job of the page. Shopify’s guide to the 2 metrics puts normal exit rates for content pages, meaning blogs and articles, at 70–80%, and the ideal range for funnel pages, like product and checkout pages, at 20–40%. The same guide cites a 2022 survey of 3,244 Shopify sites that found a 41.1% average bounce rate for visitors arriving from desktop Google search, a useful anchor for how ordinary search traffic behaves.

The clearest way to see why context decides everything: a thank-you page should have a very high exit rate. The visitor bought, the page confirmed it, and they left satisfied. That exit is the success the whole site exists to produce. The same number on a checkout page would be a disaster.

When you review your own pages, write down the job of each page first, then look at the number. The metric never carries the verdict on its own; the page’s job does.

3 patterns cover most real cases. Match your numbers to the pattern before changing anything, because the fix is different in each one.

High bounce rate on a landing page

This is a first-impression problem. Nielsen Norman Group’s research found most visitors leave within 10–20 seconds unless the page communicates value in the first 10, and GA4’s 10-second engagement line sits right on top of that window. If arrivals are bouncing, the first screen is failing to answer what this is, who it is for, and what to do next. The full pattern list lives in why visitors leave a website in seconds.

Good traffic numbers, high bounce rate

Usually message mismatch. The ad, the search snippet, or the social post promised one thing and the page opened with another, so the visitor concluded they were in the wrong place and left. The traffic source is doing its job; the page is dropping the handoff. That gap between visits and buyers has its own playbook in traffic but no customers.

High exit rates in the middle of a funnel

A conversion problem. People who arrived, engaged, and moved 2 or 3 steps toward buying are abandoning at one specific page, which usually means a surprise on that page: shipping costs appearing late, a form demanding too much, a trust wobble at the moment money enters the conversation. Walk each step the way a first-time buyer would, using the full website conversion audit guide, and fix the page where the numbers break.

One discipline makes every fix worth doing: measure in windows. Note the date you changed the page, then compare the 2 weeks after against the 2 weeks before, on the same traffic sources. Bounce and exit numbers wobble day to day, and a fix judged on 48 hours of data will look like a failure half the time for no reason. The window smooths the noise; the segment keeps the comparison honest.

Whichever pattern fits, the next step is an outside read. Run a 5-second test on a stranger, or use the methods in get honest feedback on your website to hear what visitors think before they leave. Or skip the recruiting and run a free website preview: it reads your page the way an impatient visitor does and reports what it found in under a minute.