Bounce rate: what it means, how to calculate it and when it matters.

Few numbers cause as much worry for as little reason as bounce rate. A homepage shows 58 percent, someone calls it high, and a redesign gets discussed before anyone has asked what the visitors came for. Bounce rate is useful, but only once you know how your analytics tool defines it, which page you are looking at and where that traffic came from.

This guide explains what bounce rate means, how to calculate it, how Google Analytics 4 changed the definition and why no single good bounce rate exists, with the one GA4 benchmark we found that names its source. It shows where to find the metric in GA4, Google Ads and Meta Ads Manager, what actually lowers it and the mistakes we see most often. At Viralistic we design websites in Figma and build them in Astro, and we judge every site by what it does for the business: the number of enquiries, their quality and what a new client costs. Bounce rate is a diagnostic in that picture, never the goal.

What is bounce rate?

Bounce rate is the share of visits in which someone arrives on your website and leaves without doing anything there. The visitor lands, looks, and bounces back, usually to the search results or the app they came from. Some guides call a bounce a single-page session, which was true in older analytics tools. Today the definition depends on the tool.

In Google Analytics 4, a bounce is an unengaged session. According to Google’s Analytics help, a session counts as engaged when it meets at least one of three criteria: it lasts longer than 10 seconds, it has a key event (such as a submitted form), or it has 2 or more page or screen views. A session that meets none of them is a bounce. Google describes bounce rate as the opposite of engagement rate.

The bounce rate formula

To calculate bounce rate, divide the number of unengaged sessions by the total number of sessions and multiply the result by 100.

Bounce rate = (unengaged sessions ÷ total sessions) × 100

Because GA4 defines bounce rate as the inverse of engagement rate, the same formula can be written as:

Bounce rate = 100% - engagement rate

A worked example. A service page receives 4,000 sessions in a month, and 2,600 of them are engaged. Engagement rate is 2,600 ÷ 4,000 = 65 percent, so bounce rate is 35 percent. The same 4,000 sessions produce 52 enquiries: 1.3 percent of all sessions, or 2 percent of engaged sessions. That last figure tells you more about the page than the bounce rate does, because it shows whether the people who stay go on to act.

Use the calculator below for your own page. The defaults reproduce the example.

What changed in GA4 compared with Universal Analytics?

Bounce rate in GA4 measures something different from the bounce rate in Universal Analytics. Universal Analytics counted a bounce as a session that sent only a single request to the Analytics server, which in practice meant one page and no further interaction. A visitor who spent four minutes reading an article and then closed the tab was a bounce. GA4 counts that same reader as engaged, because the session lasted longer than 10 seconds.

Standard Universal Analytics properties stopped processing data on July 1, 2023, and from the week of July 1, 2024 even the historical data became inaccessible. If you compare today’s GA4 bounce rate with a figure from an old Universal Analytics export, you are comparing two definitions, not two periods.

A worked example shows how large the gap can be. A long-form guide gets 1,000 sessions. In 800 of them, the visitor views only that guide, so Universal Analytics reports a bounce rate of 80 percent. Of those 800 readers, 500 stay longer than 10 seconds. GA4 counts them as engaged, which leaves 300 bounces and a bounce rate of 30 percent. Nothing about the visitors changed, only the way they were counted.

Adjusting the 10-second timer

Ten seconds is only the default. In GA4 you can raise it under Admin, Data streams, your web data stream and Configure tag settings. Select Show all, then Adjust timer for engaged sessions. For a long sales page or a pricing page, a higher threshold can be more honest: someone who leaves after 12 seconds has not read your argument. Change the timer once if you change it at all, and write down the date. Every change breaks the comparison between the periods before and after.

How other tools define a bounce

If you use another analytics tool alongside GA4 or instead of it, the same behaviour produces a different number. Each tool defines a bounce in its own way.

What is a good bounce rate?

There is no universal good bounce rate. Whether 40 or 70 percent is healthy depends on the page type, the traffic source, the device and what the visitor came to do. You will find average bounce rate tables online with neat ranges such as “26 to 40 percent is excellent”, but they rarely name a source, a year or a measurement method. Under GA4 they also describe a different metric from the one they were built on.

One benchmark does name its data source and period. Databox’s content marketing benchmarks use GA4 data from September 2024 and put the median bounce rate across industries at 44.04 percent. Consulting and professional services, the closest category to most premium service firms, sits at 47.84 percent. The page does not say how many accounts are in the sample, so treat it as a rough middle, not a target.

Traffic source matters more than most benchmarks admit. In Contentsquare’s 2025 Digital Experience Benchmarks, based on 90 billion sessions across 6,000 websites and comparing the fourth quarter of 2024 with that of 2023, businesses that leaned more on paid social saw bounce rates rise by 9.2 percent, page views fall by 8.7 percent and conversions drop by 10.6 percent. Contentsquare counts a single-page visit as a bounce, so its figures do not map one to one onto GA4. The direction is still clear: more paid traffic is not automatically better traffic.

A bounce also means something different on each type of page. This table shows what to look at instead.

The most useful comparison is with yourself: the same page, the same traffic source, the previous period. And always look at what follows. For a premium service business, a 60 percent bounce rate with ten strong enquiries a month beats 35 percent with two.

How to find bounce rate in GA4, Google Ads and Meta

GA4 leaves bounce rate out of most reports by default, and neither Google Ads nor Meta Ads Manager shows it under that name. Here is where to find it and what to use in its place.

Bounce rate in Google Analytics 4

You add bounce rate to a detail report yourself, which requires the Editor or Administrator role.

1. In GA4, go to Reports and open a detail report, such as Engagement, Landing page.
2. Click the pencil icon in the top right to customise the report.
3. Under Report data, click Metrics, then Add metric, and add Engagement rate and Bounce rate.
4. Click Apply and save the changes to the current report.

The Landing page report is the right place, because a bounce is always about the first page of a session. Add Session source / medium as a secondary dimension and you can see which channel drives the bounces on each landing page. To see what visitors do after their first page, build a path exploration with that landing page as the starting point.

Bounce rate in Google Ads

Google Ads does not show a GA4 bounce rate. Once you link GA4 to Google Ads and import site metrics, you can add the Google Analytics columns through the columns icon: % Engaged sessions (GA4), Events / session (GA4) and Avg. engagement duration per session (seconds) (GA4). The first one is the mirror image of bounce rate. It shows, per campaign, ad group or keyword, whether the clicks you pay for bring visitors who stay.

Google’s own help article gives an example. A florist tests two ad groups with 1,000 impressions each. “Birthday bouquets” earns an 8 percent CTR, but 60 percent of its visitors bounce. “Birthday flower arrangements” earns 6 percent, with a 30 percent bounce rate. Judge on click-through rate alone and you back the wrong ad group.

Mind the gap between clicks and sessions as well. When someone clicks your ad and leaves before the page has loaded, Google Ads records a click and Analytics records no session. That visitor never shows up as a bounce anywhere.

Bounce rate in Meta Ads Manager

Meta Ads Manager has no bounce rate. Compare link clicks with landing page views instead. A landing page view only counts when the page actually loads after the click, and it requires the Meta Pixel. A wide gap between the two means people leave before your page appears, which is usually a speed problem. For behaviour after the page loads, use GA4: tag your Meta campaigns with UTM parameters and the Landing page report shows engagement rate and bounce rate per campaign.

Google Search Console as a companion

Google Search Console does not report bounce rate, but it does show which search queries bring people to a page. Put those queries next to the page’s bounce rate in GA4. A high bounce rate on queries that do not match what the page offers points to a mismatch in expectations, not to a bad page.

Why visitors bounce

Visitors almost always bounce for one of five reasons: they expected something else, the page loads too slowly, they cannot tell what you offer, they see no next step, or your tracking is broken.

Mismatched expectations are the biggest cause. An ad, a search snippet or a social post promises something the page does not deliver. Google gives the same advice in its help: if visitors from display ads are bouncing, check that your ads are relevant to your site content. It is also why a broad campaign produces a higher bounce rate than a focused one. You reach more people the page was never written for.

Speed comes second. A visitor staring at a blank screen is gone before your tag loads. Google uses the Core Web Vitals as its measure of a good experience: the largest content appears within 2.5 seconds (LCP), the page responds to a click or tap within 200 milliseconds (INP) and the layout barely shifts (a CLS of 0.1 or less), for at least 75 percent of page loads.

The message is third. Within a few seconds, a visitor should see what you do, who it is for and why it matters. Vague headlines hand that work to the visitor, who usually declines. A lower bounce rate often starts with your brand script, long before anyone touches the page.

A missing next step is fourth. A page without a clear route to your services, a case study or a way to get in touch leaves the visitor a choice between scrolling and leaving. On mobile, pop-ups and cookie banners that cover the first screen make that choice for them.

The fifth reason is technical rather than behavioural. Pages without a tag, a form that is not set up as a key event, or a checkout on another domain distort the picture. Google says as much: if low engagement is widespread, check your measurement code first.

How to reduce bounce rate

Reducing bounce rate is only worth it where bounces cost you enquiries. Start by deciding which pages get traffic you want to keep: your homepage, your service pages and your campaign landing pages. That is where the effort pays. On a guide that answers a question in one go, a bounce is often a satisfied reader.

Then work in a fixed order. First make sure your measurement is right, so that a submitted form, a click on your phone number and a booking count as key events. Next, line up the promise of your ad or search result with the first screen of the page. Make the page fast. Give every page one clear primary action and sensible internal links to whatever the visitor will want to know next. When you are torn between two good options and have enough traffic, run an A/B test, so you know instead of guess. Our guide to conversion rate optimisation explains how to run one reliably.

This table links the signal in your data to the first step we would take.

What a good route to an enquiry looks like before anything is designed is covered in our article on UI and UX design.

Bounce rate vs exit rate vs engagement rate

Bounce rate is about the first page of a session, exit rate about the last. Exit rate is the share of a page’s views that were the final view of a session, however many pages came before. Every bounce ends on the page where it started, so every bounce is also an exit, but not every exit is a bounce. A high exit rate on a thank-you page is expected. On a quote form it is an alarm.

Bounce rate rarely tells the story alone. This is how it relates to the metrics around it.

Bounce rate is a diagnostic metric at the top of the funnel. Your KPIs sit at the bottom: enquiries, their quality, what a new client costs and, for ecommerce, ROAS. How to choose and report those is covered in our guide to marketing KPIs.

Common mistakes that distort the number

The most common mistake is steering on the site-wide average. That number blends blog posts, contact pages and campaign pages, and so it blends readers who leave satisfied with visitors who give up. You only learn something when you look per landing page and per channel.

The second is putting old Universal Analytics figures next to GA4 and concluding that the site suddenly performs far better. The behaviour did not change. The definition did.

The third is pushing the number down artificially. Mark a scroll event or an autoplaying video as a key event and your bounce rate drops without a single extra enquiry. A sudden fall to almost zero is usually a tracking error: if your page view fires twice, every session has 2 page views and GA4 counts every session as engaged. Marking first_visit or session_start as a key event does nothing, by the way: GA4 excludes those events from the engaged-session calculation.

The fourth is treating bounce rate as an SEO target. Google’s John Mueller put it plainly on Reddit: “Google doesn’t use that for SEO, so if you’re focused on SEO, you can ignore that.” What does matter for rankings is whether your page answers the query better than the alternatives, which is what our plan to rank higher on Google is about.

The fifth is forgetting who you do not measure. Depending on your consent settings, GA4 measures visitors who decline cookies only partly or not at all. Your bounce rate describes part of your audience, and that part may behave differently from the rest.

How we help

At Viralistic we do not start with bounce rate. We start with what your website has to deliver. We set up measurement so that enquiries, phone clicks and bookings count as key events, and we report what matters: how many enquiries arrive, how good they are and what a new client costs. Bounce rate is our diagnostic. It points us to the pages and channels where something is off.

If the foundation is not healthy yet, that is the first step. A clear brand script, a fast website and solid search visibility solve most bounce problems before a single ad runs. We design in Figma and build in Astro, and webdesign projects start around EUR 4,000. See how we work in our webdesign approach. If you already run campaigns, we look together at which clicks bring visitors who stay. Paid media is an accent on a healthy foundation for us, never a way to pour budget into a leaking bucket. Read more in our marketing approach and in our guide to Google Ads management. One strategist keeps the overview, and a person checks every piece of work, also when AI helped.

Frequently asked questions

What is an ideal bounce rate?

There is no ideal bounce rate that applies to every site. In GA4, bounce rate depends on the page type, the traffic source and the device: a guide bounces differently from a product page, and paid social traffic differently from organic search. Compare each page with its own previous period and judge it alongside enquiries or sales, which show whether the page does its job.

Is a 42% bounce rate good?

For most websites, 42 percent sits close to the middle. Databox’s GA4 benchmark for September 2024 puts the median bounce rate across industries at 44.04 percent, from 35.76 percent for apparel to 48.38 percent for IT services. Whether it is good for you depends on the page, so compare it with its previous period and its enquiries.

Why is my bounce rate so high?

Most often because visitors expected something else: an ad, search snippet or social post promised what the first screen does not show. Slow loading, an unclear message and a missing next step follow. Before you redesign anything, check per landing page and channel, and confirm that forms and phone clicks count as key events.

Is a 20% bounce rate good?

A 20 percent bounce rate can be good, but check your tracking first. Very low bounce rates often come from a page view that fires twice, or from trivial events marked as key events, which make every session look engaged. If the setup is clean and enquiries rise with it, a low bounce rate is a genuinely good sign.

What does a 100% bounce rate mean?

A 100 percent bounce rate means no session on that page was engaged. On a page with very little traffic, that can be chance. On a page with real traffic, it usually signals a tracking problem, such as a missing tag or a form on another domain, or a page that fails to load. Check the measurement before you change the page.

Is it better to have a higher or lower bounce rate?

A lower bounce rate is usually better on pages meant to move visitors forward, such as your homepage, service pages and campaign landing pages. On guides, glossary pages and contact pages, a higher bounce rate often means people found what they needed. Better is the wrong question on its own: ask whether the page produces the enquiries or sales it exists for.

Does bounce rate affect SEO?

No. Google’s John Mueller has said Google does not use bounce rate for SEO and that you can ignore it if SEO is your focus. Many websites do not use Google Analytics at all, which makes it an unlikely signal. What matters is whether your page answers the query well, and a high bounce rate can be a useful hint about that for you.

Is website bounce rate the same as email bounce rate?

No. Website bounce rate is about visitors who leave your site without engaging. Email bounce rate, as reported by tools like Mailchimp, is the share of emails that could not be delivered, for example because an address no longer exists. The two metrics share a name, but they measure different things and call for different fixes.