In short
There is no universally good bounce rate; typical figures run from 20–40% for e-commerce and web applications, 40–60% for service and lead-generation sites, to 60–90% for blogs, news and single-answer pages, where leaving after one page is the expected behaviour. In GA4, bounce rate is the inverse of engagement rate, so a session counts as engaged if it lasts over ten seconds, includes a conversion or views two pages. Judge a page's bounce rate against similar pages on your own site and against what the visitor was meant to do, not against a global average.
What it measures, then and now
In Universal Analytics, a bounce was a session with exactly one page view and no other interaction. If someone read your article for eight minutes and closed the tab, that was a bounce, indistinguishable from someone who arrived, saw the wrong thing, and left in two seconds.
GA4 replaced the metric. Bounce rate is now simply the percentage of sessions that were not "engaged", and a session is engaged if it lasts longer than ten seconds, includes a conversion event, or has two or more page views. The eight-minute reader is no longer a bounce. This makes the GA4 figure lower than the old one for most sites and makes comparisons with pre-2023 benchmarks meaningless.
It also makes the metric slightly more useful, because it now excludes the visitors who clearly got value, and slightly more gameable, because a ten-second threshold is not hard to pass.
Typical ranges by page type
| Page or site type | Typical bounce rate | Why |
|---|---|---|
| E-commerce, web apps | 20–45% | Visitors browse multiple pages by design |
| Service and lead-gen sites | 40–60% | Some read one page and call; some browse |
| Landing pages | 60–90% | One page, one action, then leave; high is normal |
| Blogs, news, single-answer articles | 60–90% | Visitor gets the answer and leaves satisfied |
| Dictionary, reference, tool pages | 80–95% | Look up, done |
These are broad ranges consistent with published benchmarks, expressed in the old definition; GA4 figures run lower. The point of the table is the third column: bounce rate is high where the page's job is done in one visit, and that is not a problem.
The three misreadings
"High bounce rate means the page is bad." For an article that answers a question, a high bounce rate is what success looks like. The reader got the answer. Judging it against an e-commerce benchmark is a category error, and "reducing bounce rate" on such a page usually means adding friction, pop-ups or clickbait, which makes the page worse for the reader while improving the metric.
"Low bounce rate means the site is good." A bounce rate near zero is almost always a tracking bug: the analytics tag firing twice, an event that counts as engagement firing on load, or a redirect that counts as a second page. A real site with a 3% bounce rate has a measurement problem, not a brilliant design.
"Bounce rate is a ranking factor." Google has said repeatedly that it does not use Google Analytics data in ranking. What it may use are its own behavioural signals, such as a searcher clicking a result and quickly returning to the results page. That is related to bounce rate but is not the same thing, and it is not affected by your analytics configuration.
When a high bounce rate is a real problem
On pages that are supposed to lead somewhere. A homepage with a 75% bounce rate is losing three-quarters of visitors before they see a service. A product listing with a 70% bounce rate is failing to get people to products. A landing page with an 85% bounce rate and a 1% conversion rate is not persuading anyone.
The diagnosis is the same as for any conversion problem: the visitor arrived expecting one thing and found another, the page is slow, the next step is unclear, or the traffic source is sending the wrong people. Segment by source and device before concluding anything; a bounce rate that is 40% from search and 85% from a social campaign is a campaign problem, not a page problem.
What to measure instead
Bounce rate is a symptom, and a blunt one. The measures that tell you what happened:
Engagement rate and average engagement time, GA4's replacements, which at least distinguish a two-second exit from an eight-minute read.
Scroll depth. Did the visitor reach the end of the article, or the pricing section, or the form? Far more informative than whether they clicked a second page.
Conversion or next-step rate per page. What the page was for, and whether it did it. This is the number that should replace bounce rate in any report that goes to someone who makes decisions. Our piece on what analytics to actually track sets out a short list.
Return visits and branded searches. The reader who bounced today and came back next week by name is the outcome content is supposed to produce.
Practical guidance
Define what each page is for. Compare its bounce rate with other pages of the same type on your own site, segmented by traffic source. Investigate the outliers: a service page far above its siblings, a landing page far above the campaign's other pages. Ignore the global average entirely, and be suspicious of any figure below 10%. If someone is reporting bounce rate to you as a site-wide health score, ask them for conversion rate by page instead.
If your analytics are producing numbers you do not trust, book a call and we will check the setup; misconfigured tracking is the most common reason a bounce rate looks either wonderful or terrible.
Common questions
Is a 70% bounce rate bad?
For a blog post, a reference page or a landing page, 70% is normal; visitors get what they came for and leave. For a homepage, a product listing or any page whose job is to send people onward, 70% means most visitors are lost before the next step, and it is worth diagnosing. The page's purpose decides, not the number.
What is a good bounce rate in GA4?
GA4 defines bounce rate as the share of sessions that were not engaged, where engaged means over ten seconds, a conversion, or two or more pages. Because that excludes visitors who read a page properly, GA4 bounce rates are lower than the old Universal Analytics figures, and most sites see 25–55%. Compare pages of the same type on your own site rather than against a global figure.
Does bounce rate affect SEO rankings?
Google has said it does not use Google Analytics data for ranking, so your bounce rate as measured in your own analytics has no direct effect. Google may use its own signals about searchers returning to results quickly, which is a related but different behaviour. Improving the page for visitors helps both; manipulating the metric helps neither.
Why is my bounce rate so low?
A bounce rate under about 10% is almost always a tracking error: the analytics tag installed twice so every visit records two page views, an event firing on page load that counts as engagement, or a redirect that registers as a second page. Check the tag installation before celebrating.
How do I reduce bounce rate?
First decide whether you should; on content pages a high rate is normal. On pages that should lead somewhere, match the page to what the traffic source promised, make it fast on phones, put the next step clearly in view, and segment by source to find campaigns sending the wrong visitors. Do not add pop-ups or artificial page splits, which lower the metric and worsen the page.
