What Is a Good Bounce Rate for a Small Business Website?

Someone showed you a 74% bounce rate and implied your website is broken. Before you spend money on that, here is what we measure on real sites. Across 30 small business sites over the 90 days to September 2026 the median bounce rate was 42%, and perfectly healthy companies in the same trade ranged from 18% to 79%. What is a good bounce rate is a question with no honest single-number answer, and the reason why is more useful than the number would have been.

Bounce rate is an instrument reading. It describes what your measurement tool noticed, not what the customer felt. Two sites can behave identically for a visitor and report bounce rates thirty points apart, purely because of how each tool decides a session was worth counting. That is not a criticism of the metric. It is a warning about using it as a grade.

The measured answer: 42% median, 18% to 79% spread

Bots were excluded before anything was counted, sites with fewer than 30 visits in the window were dropped, and Venbit's own properties were left out. Our tracker counts a session as engaged once it passes ten seconds or reaches a second page. Everything else is a bounce. Hold onto that definition, because it explains the strangest thing in the data.

GA4 changed what the word means, so old benchmarks do not apply

If you are comparing your number to a benchmark article written a few years ago, you are comparing two different measurements. Universal Analytics called a bounce any session with a single interaction, full stop. Someone could read your entire about page for four minutes and leave, and that was a bounce. GA4 inverted the whole thing and built it on engagement instead.

Engaged session: a session that meets any of the following criteria. Lasts longer than 10 seconds, has a key event, or has 2 or more screen or page views. Bounce rate: the percentage of sessions that were not engaged.

Read that carefully, because it has a real consequence. Under GA4 the four-minute reader is no longer a bounce, so most sites saw their reported bounce rate drop sharply when they migrated. Nothing about the website changed. The ruler changed. Any benchmark that does not say which definition it used is not a benchmark you can act on.

Our own tracker uses the ten-second and second-pageview tests but does not treat a conversion as automatic engagement, which makes our bounce numbers slightly harsher than GA4's on the same traffic. We are telling you that so you can adjust rather than assume the tools agree. They rarely do.

Average bounce rate for small business websites, by trade

Here is what 30 real small business sites actually reported. Engaged rate is the inverse of bounce rate. Active seconds is median time genuinely spent on the page rather than time with the tab open. Conversion means a form submission or a phone tap.

The caveats matter as much as the figures. Each group holds between 3 and 14 sites. These are medians of specific businesses we work with, measured the same way, not industry averages, and a group of four tells you about four companies. Sites with no conversion goal configured were excluded from the conversion column only.

Notice the group with the lowest bounce rate. Local services and hospitality sit at 29% bounce and 80 seconds of active time, the best-looking engagement numbers on the page, and the lowest conversion rate in the set at 2.9%. People linger, browse the photos, read the menu, and book somewhere else entirely. Engagement is not intent.

The two worst bounce rates belong to two of the best sites

The highest bounce rates we recorded in the window were 79% and 66%. Both are window treatment companies. Both are among the best converting sites we measure. Neither is broken.

Here is what those visits look like. Someone searches for blinds installation, lands on the page, reads the service area and the phone number, taps to call, and is gone in eight seconds. One page. Under ten seconds. By our instrument that is a bounce, and it is also the single best outcome the website could have produced. The visit that ends in a phone call looks statistically identical to the visit that ended in disgust.

Run that pattern across a business where most customers phone rather than fill in forms and you get a 79% bounce rate on a site that is doing its job perfectly. In our data, bounce rate did not predict conversion. Not weakly. Not at all.

What outside benchmarks say, and why they do not settle it

For a comparison that is not ours: Contentsquare, which builds its benchmark program from tens of billions of sessions across thousands of websites, publishes average bounce rates of 43% on desktop, 51% on mobile and 45% on tablet across all industries. Their 2026 benchmark report is drawn from 99 billion sessions across more than 6,500 websites. That is their number, not ours, and their published figure does not state which analytics definition or which year sits behind it, which is exactly the problem with benchmark shopping.

Their range and ours overlap, which is mildly reassuring and completely useless for deciding whether your site is fine. A benchmark can tell you that you are unusual. It cannot tell you that you are wrong.

Judge calls and forms per hundred visits instead

Swap the question. Instead of asking whether your bounce rate is acceptable, ask how many of every hundred visits end in someone contacting you. Across our measured sites with a goal configured the median is 4.5%, so roughly four or five contacts per hundred visits. The full spread runs 0.4% to 19%. The 19% site is a showroom business where nearly everyone arriving is already ready to book.

Then add the number we trust most: the share of visits that reach a contact page or tap the call button. That runs from 1% to 65% across our sites and it tracks lead volume more closely than any other metric we have. It is also easy to reason about. If 3% of visits get anywhere near contacting you, the problem is upstream of the form.

Both of those numbers survive the phone-tap problem that ruins bounce rate, because both of them count the outcome rather than the shape of the session. If you are still working out whether you have enough visits for any of this to matter, that is a separate question with its own numbers.

Why visitors actually leave your website

When a high bounce rate is a real problem, it is almost always one of these, roughly in the order we find them.

1. 1They got what they came for. Phone number, hours, address, service area. This is the most common cause of a scary bounce rate on a local business site and it needs no fixing at all.
2. 2The page was slow on a phone. People leave during the load and never appear as an engaged session. If your site is fine on office wifi and grim on cell data, that specific failure has its own diagnosis.
3. 3The page answered a different question than the search. Someone searched for a price and landed on a company history. Nothing is broken, the match is just wrong.
4. 4No price, no service area, no proof. Visitors will not send a form to a business that has not told them whether it serves their town or works in their budget. Content that answers those questions is the fix, not a redesign.
5. 5The form is broken, too long, or eats submissions silently. More common than anyone wants to believe. Test yours from a phone on cell data and see whether the email actually arrives.
6. 6It is the wrong traffic. A broad ad keyword, a directory listing, or a scraper that slipped past filtering. Bad traffic bounces because it was never yours.
7. 7The site looks untrustworthy on a phone. Text running off screen, an obviously dated layout, no https. That judgment lands in about two seconds and it has visible symptoms.

How to tell which one it was

Bounce rate cannot distinguish satisfied from frustrated, so stop asking it to. What can distinguish them are friction signals and visitor paths. Rage clicks, where someone taps the same spot repeatedly because nothing responds. Dead clicks, where they tap something that looks like a button and is not. Form errors, where they tried to submit and the page refused. Those are behaviours, and behaviours have an obvious reading: nobody rage clicks a page that gave them what they wanted.

Watching the path a single visitor took answers the rest. A visit that goes home, services, contact, and then stops at the form is a completely different story from a visit that loads one page and vanishes, and a bounce rate averages both into a number that describes neither.

This is one of the things Venbit Analytics exists for. It is the first-party analytics we built for our own clients: one script tag, no cookies, bots excluded, a session counted as engaged after ten seconds or a second pageview, and a form submission or a phone tap counted as a conversion. It shows leads, sources including which ad platform a click came from, each visitor's journey, friction signals, heatmaps and Google rankings next to visits and leads, in a web portal and an iOS app. It is free for Venbit clients and there is no public price yet.

It is also not the only way to see this, and pretending otherwise would undercut the point of the article. Microsoft Clarity is free, does heatmaps and rage clicks well, and costs you nothing but a script tag. If all you want is to find out why people are leaving one page, start there.

What to do this week

1. 1Find out how many contacts per hundred visits your site produces. Everything else is secondary to that one ratio.
2. 2Check whether your phone number is a tap-to-call link on mobile. If it is not, some of your bounces are people copying a number by hand and half of them give up.
3. 3Load your top landing page on a phone on cell data and time it.
4. 4Split bounce rate by page rather than reading the site-wide figure. One bad page usually explains the whole number.
5. 5Ignore any benchmark that does not tell you which analytics definition produced it.