What is a good bounce rate?

A bounce is a visit that sees one page and leaves — no second pageview, no click worth recording. Bounce rate is the share of all visits that end that way. It's the most quoted number in web analytics after visitor count, mostly because it sounds like a verdict: they came, they looked, they left.

One definitional landmine before any benchmark: tools don't agree on what a bounce is. Most simple analytics tools use the classic single-page definition above. GA4 quietly redefined it — there, a bounce is any session that isn't 'engaged', meaning it lasted under 10 seconds, had no conversion and under two pageviews. The same site can show 65% bounce in one tool and 35% in another while both are correct by their own rules. Compare a bounce rate only to numbers from the same tool.

For a typical multi-page site, the commonly cited bands: 26–40% is excellent, 41–55% is average, 56–70% is above average but not alarming, and over 70% deserves a look. Most sites live somewhere in the 40s and 50s.

The bands shift hard by site type, though. Blogs and content sites run 65–90% — someone searched a question, read the answer, left satisfied; that's a bounce and a success simultaneously. Single-CTA landing pages run high by design, since the goal is a click-through that often leaves the site. Ecommerce runs low, 20–45%, because shopping is inherently multi-page. And traffic source moves the number too: search visitors bounce less than social visitors, because someone mid-question has more intent than someone mid-scroll.

Bounce rate's dirty secret is how much success it counts as failure. A visitor who lands on a pricing FAQ, reads for four minutes and leaves informed is a bounce. A single-page app where all navigation happens client-side can show a 90% 'bounce rate' while users spend half an hour inside, unless the tool tracks route changes as pageviews. Meanwhile a genuinely broken page — wrong audience, five-second visits — can share the same 75% as a beloved blog.

So the number only becomes a diagnosis with a second signal next to it. High bounce plus long time-on-page is content doing its job. High bounce plus sub-10-second visits is a mismatch between promise and page — an ad writing checks the landing page doesn't cash, a headline answering a different question than the one searched.

First, aim at the right pages: sort by traffic, filter to pages where a second click is actually the goal, and take the high-bounce ones from the top of that list. Lowering the blog's bounce rate is a hobby; lowering the pricing page's is money.

On those pages, the fixes are unglamorous: match the page to the promise that brought the visitor (the ad, the meta description, the social post), get the load time down — every extra second of loading visibly raises abandonment — put the point above the fold, and give the visitor an obvious next step. Then, instead of guessing which of those it is, watch a handful of session recordings of bouncing visitors. Thirty seconds of someone scrolling twice and leaving explains a 78% bounce rate faster than any amount of dashboard-staring.

A good bounce rate is 26–40% for a site that expects multi-page visits, 65–90% is unremarkable for content, and any number is meaningless without knowing the page's job and the tool's definition. Treat it as a smoke detector, not a verdict: it tells you which page to look at, and the conversion rate, time-on-page and a few recordings tell you what's actually on fire.