How does GA4 calculate bounce rate?
Bounce rate is 100% minus the engagement rate: sessions that weren't engaged, divided by all sessions.
On a Dutch webshop for office furniture, GA4 counted 7,264 sessions in the three months to October 27, 2026, and 57.2% of them were engaged: a bounce rate of 42.8%. Organic Search alone bounced at 26.1%.
Universal Analytics, switched off on July 1, 2023, counted any one-page session as a bounce. In GA4, three minutes on one page count as engaged, if GA4 sees them.
Where do you find bounce rate in GA4?
You add it yourself. Open Reports > Acquisition > Traffic acquisition, click the pencil (Customize report), then Metrics > Add metric > Bounce rate, Apply and Save (Editors and Administrators only, in detail reports). GA4 for SEO covers the rest of the tool.
For an agency, MarketingTracer shows each client site's GA4 numbers: the Bounce Ratio per channel (Analytics > Channels) and the list of events (Analytics > Events).
What is a normal bounce rate?
About half, with a wide spread. AgencyAnalytics put the median GA4 bounce rate at about 45.9% across 150,000+ campaigns on its platform (December 2024). In its August 2025 tables, Advertising & Marketing sits at 61.4%, the United States at 56.2% and Germany at 36.0%.
Most other benchmarks ("40% or lower is good", 26 to 40%, a table often credited to Google) count one-page visits. Contentsquare, whose 2026 numbers Shopify quotes, counts visitors who leave after viewing only one page: a different metric.
GA4's own Benchmarking compares you with your industry on the same definition. Your own history, per type of page, beats both.
Is bounce rate a ranking factor?
No. Google's John Mueller, in 2021: "Google doesn't use that for SEO, so if you're focused on SEO, you can ignore that."
Google does use aggregated and anonymized interaction data to judge whether its results fit the query. Your GA4 property isn't part of that. A page that misses the query loses either way. The queries and clicks before the visit are in Search Console.
Why is my bounce rate so high?
It's a complicated story: make good plans, and always look at what really happens. Start with the tracking, then the channel, and only then the page.
GA4 only counts the 10 seconds when the tag reports them in user_engagement, the event that carries the time on a page. Without it the 10-second rule can't fire, and a long read on one page can count as a bounce. Some of my own sites send a minimal GA4 hit without that event, on purpose: I chose speed over data there.
The first thing I'd check on a client site is that event: is user_engagement listed under Reports > Engagement > Events, and is Average engagement time per session in Traffic acquisition above 0s? If not, find out whether someone chose that. The standard Google tag sends that event itself. A custom or lightweight tag sends only what it was built to send. Until the tag sends it, compare the site only with itself.
Then the channel. On the furniture shop, Direct brought half the sessions, and 56.8% of them bounced. Organic Search, at 26.1%, sits well under the 46% median. Whatever drives those Direct visits, they aren't your SEO: read Organic Search.
Then the page. A bounce rate for the whole site says nothing; split it by page type and look at what each page is for. Readers who leave an article may have found the answer. On the shop, articles bounced at 40.6% and product pages at 35.7%. To the shop's owner I'd say: Your bounce rate is fine. What worries me is 373 product pages where people land and nobody buys: that's where the money is. One black corner desk had 86 sessions at 18.6% bounce and no sale. A page that doesn't match the search loses people, and its queries in Search Console show that. Slow pages lose them before they load; speed is what I write about on corewebvitals.io.
Bounce rate isn't a goal; sales and requests are, and those go in the report. If the client still asks for it, show Organic Search next to the site's own previous period and leave the benchmarks out.