Bounce Rate in GA4: What It Measures Now and How to Reduce It

The first version of this post went up in December 2019. It explained bounce rate the way Universal Analytics measured it and promised ten proven strategies, with a benchmark table for every industry. Somewhere along the way it picked up "2025" in the title without anyone checking whether the metric it described still existed.

It doesn't. Universal Analytics is gone, and the bounce rate you see in Google Analytics 4 today is a different number with the same name.

So this is the rewrite. Less "ultimate guide", more "here's what the number means now and what I'd actually do about it".

Universal Analytics Is Gone, and Its Bounce Rate Went With It

Standard Universal Analytics properties stopped processing data on July 1, 2023. Paying Analytics 360 customers got one extra year. From the week of July 1, 2024, Google shut off access to every UA property, including the API, and deleted the data. If you never exported your old reports, they are not coming back.

That matters for this topic because the old bounce rate was a blunt instrument. In UA, a bounce was a single-page session with no further interaction. Someone could land on your blog post, read every word for four minutes, nod, and close the tab. Bounce. Same as someone who saw the page, hated it, and left after one second.

That's why the old benchmark tables (including the one this post used to have) said blogs "normally" sit at 40 to 60 percent. Of course they did. Reading one article and leaving was counted as failure, even when it was the whole point.

How GA4 Defines Bounce Rate Now

GA4 starts from the opposite direction. It defines an engaged session, and Google's own help page lists three ways a session qualifies. It has to do at least one of these:

• Last longer than 10 seconds
• Trigger a key event (what GA4 used to call a conversion)
• Include 2 or more page or screen views

Engagement rate is the percentage of sessions that were engaged. Bounce rate in GA4 is the percentage of sessions that were not engaged. They always add up to 100.

Quick math: 1,000 sessions, 640 of them engaged. Engagement rate 64 percent, bounce rate 36 percent. That's the whole formula.

Two settings change the result, and almost nobody touches them. The engaged session timer defaults to 10 seconds, and you can raise it (up to 60 seconds) under Admin, Data Streams, your web stream, Configure tag settings, Adjust session timeout. The session timeout itself defaults to 30 minutes. If you change the timer, write down the date, because every report before that day was measured with a different ruler.

Bounce rate also isn't shown in most standard reports by default. Open a detail report such as Pages and screens or Landing page, click Customize report, then Metrics, then Add metric, type "Bounce rate" and apply. Add engagement rate next to it. Seeing both side by side stops you from reading one of them backwards.

Why Old Bounce Rate Benchmarks Are Useless Now

The reader who spent four minutes on your article and left? In UA that was a bounce. In GA4 it's an engaged session, because it lasted longer than 10 seconds. Same visit, opposite verdict.

So if you compare your GA4 bounce rate to a benchmark table written for Universal Analytics, you are comparing two different metrics. Your number will almost always look better, and you'll learn nothing from it. I deleted the industry benchmarks from this post for exactly that reason. I don't have reliable GA4 benchmarks per industry, and I'm not going to recycle UA numbers with a new label.

There's a second problem: your analytics doesn't see everyone. Visitors who decline your cookie banner usually aren't counted, and ad blockers stop the analytics script entirely. I went through how big that gap gets in the invisible problems that quietly cost websites customers, and through the attribution side of it in why half my traffic shows up as "Direct".

What I'd do instead of benchmarks: compare a page against itself over time, and compare pages of the same type against each other. If your service pages sit at 35 percent and one of them sits at 70, that one is your lead.

Is Bounce Rate a Google Ranking Factor?

Google's people have said no for over a decade. Gary Illyes wrote in 2015 that Google doesn't use analytics or bounce rate in search ranking. John Mueller repeated it in a 2022 office-hours session, calling it a misconception that Google looks at the analytics bounce rate when ranking sites.

What Google does have is its own data about clicks on its own search results. That came up in the US antitrust trial under the name Navboost. Someone clicking your result, coming straight back to Google and clicking a competitor is information Google can see without ever touching your GA4 property.

So stop treating bounce rate as an SEO lever. Treat it as a symptom. A page that people leave without doing anything is usually a page that didn't give them what the search result promised, and that shows up in rankings eventually, through Google's own signals, not through your dashboard.

When a High Bounce Rate Is Completely Fine

Someone searches your business name, lands on your contact page, copies the phone number in eight seconds and calls you. GA4 calls that a bounce. It was a perfect visit.

Single-answer pages behave the same way. Opening hours, an address, a price, a definition. If the answer is above the fold and the visitor gets it fast, they leave fast. You can fix the measurement by marking phone and email clicks as key events, so those sessions count as engaged. Making the answer harder to find to keep people around is not a fix.

A high bounce rate becomes a problem on pages whose job is to move someone forward: a service page or a landing page you pay to send traffic to. Those are the ones I look at.

What Actually Makes People Leave

The original version of this post had seven reasons and ten strategies, some of them overlapping and some of them dated (Google Optimize, which it recommended for A/B tests, was shut down in September 2023). Here's the shorter list I actually check, in order.

The page is slow

Google published research back in 2017 showing that as mobile load time goes from 1 second to 3 seconds, the probability of a bounce increases by 32 percent. The exact number is old. The direction isn't.

Today I check Core Web Vitals: Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, Cumulative Layout Shift under 0.1. I explain the tools and how to read them in my website speed testing guide. If you're curious how much the platform itself matters, Astro sites pass Core Web Vitals 67 percent of the time and WordPress sites 49 percent, which I unpacked in Astro vs. WordPress performance numbers.

The headline doesn't match the click

Open your page and read the first headline. Does it say what you do and who it's for? Or does it say something like "Solutions that drive results"? A visitor decides in seconds whether they're in the right place, and vague copy sends them back to Google.

For pages that get search traffic, also compare the headline to the query people used to find it. Search Console shows you those queries. If people search "emergency plumber Augsburg" and land on a page titled "Our Philosophy", they're gone. I put this fix at the top of the boring fixes that beat a redesign, because rewriting one headline costs nothing.

The mobile layout fights the visitor

Tiny text and buttons too close together are bad enough. Add a cookie banner and a newsletter pop-up stacked on top of each other before anyone has read a sentence, and the visitor is gone. Google has discouraged intrusive interstitials on mobile for years, and visitors punish them faster than Google does. If you need a pop-up, show it after someone has engaged, not on arrival.

Something is broken

A 404 from an old link, a form that silently fails, an expired SSL certificate, a JavaScript error that kills the menu on one browser. These don't look like bounce rate problems until you check where the high-bounce sessions land. Then they look exactly like one.

There's no obvious next step

If the page answers the question and then just ends, people leave. Relevant internal links and one clear call to action give them somewhere to go. Not five calls to action. One that matches what the page was about.

How I'd Check Your Bounce Rate in GA4

This takes about half an hour and doesn't need any extra tools.

1. Open Reports, Engagement, Landing page and add bounce rate and engagement rate as described above.
2. Sort by sessions. Ignore pages with tiny traffic, the percentages there are noise.
3. Add a comparison for device category. A page that's fine on desktop and terrible on mobile is a layout or speed problem, not a content problem.
4. Look at session source. Paid traffic with a high bounce rate is the expensive kind, because you paid for every one of those visits.
5. Open the worst page on your phone, on mobile data, and use it like a stranger would.

Step five finds more problems than the first four combined. The numbers tell you where to look. They don't tell you what's wrong.

The Number Changed. The Job Didn't.

Bounce rate in GA4 is a cleaner metric than the one I wrote about in 2019. It no longer punishes someone for reading one page and leaving happy. It's still only a pointer to pages worth opening, and it still misses a chunk of your visitors.

I don't think anyone needs ten strategies to reduce bounce rate. You need to know which pages have a problem, why people leave them, and whether leaving was the right outcome in the first place.

If you'd like someone to go through your GA4 data and the pages behind it with you, let's talk. I'll tell you which numbers to care about and which ones to ignore.