Bounce rate
In one sentence
Bounce rate is the percentage of visitors who arrive and leave without any further interaction, and it is one of the most cited and least reliable web metrics because it counts a satisfied departure and a wrong-page exit as the same event.
Not to be confused with Exit intent, or Site search abandonment.
Definition
Bounce rate is the percentage of visitors who arrive and then leave without any further interaction. It is one of the most quoted and least reliable metrics on any analytics dashboard.
The deeper version is that it has become substantially less meaningful over time, and understanding why prevents a great deal of wasted analysis.
The definitional problem
- Traditionally, a bounce is a session with a single interaction event.
- A visitor who arrived, read a page for four minutes, got the answer and left satisfied is a bounce.
- A visitor who arrived, saw the wrong page and left in two seconds is also a bounce.
- The metric cannot distinguish them, which is the core limitation.
How measurement has changed
- Analytics platforms have moved toward engagement-based definitions, counting a session with meaningful interaction, a time threshold or multiple page views as engaged.
- This means bounce rate figures from different periods or platforms are frequently not comparable.
- Any historical bounce rate comparison should be checked for a definition change before it is trusted.
Why high bounce is not necessarily bad
- Single-page content that fully answers a question should have high bounce and high satisfaction.
- Reference pages, contact details and opening hours are all legitimately single-visit.
- A blog post that answers a question completely has done its job when the reader leaves.
What to look at instead
- Scroll depth and time on page, which distinguish reading from leaving.
- Whether the question the visitor likely had could have been answered on the page.
- Return visits and branded search, which indicate the page worked.
- Conversion, where a conversion goal exists.
The interaction with on-site voice
- A voice interaction is an engagement event, so a visitor who speaks to a widget is not a bounce.
- That mechanically improves bounce rate without necessarily improving anything real.
- It is a measurement artifact worth being honest about rather than claiming as a benefit.
Common misconception
That a lower bounce rate is straightforwardly better. It measures single-interaction sessions, and those include both complete failure and complete success. Chasing it down without segmenting by page type optimizes a number that does not correspond to a business outcome.
Why it matters commercially
Bounce rate appears in every analytics dashboard and drives decisions it should not. Knowing its limits prevents optimizing for a figure that does not track a business outcome, and it is easy to move the number without moving anything real.
In voice specifically
A voice interaction registers as an engagement event, so a visitor who speaks to a widget stops counting as a bounce. That lowers the number mechanically without necessarily improving anything real, which is a measurement artifact worth naming rather than selling. The real question is whether the visitor got their answer.
Where AsqVox fits
Voice interactions register as engagement, which mechanically reduces bounce rate. That is a measurement effect rather than an outcome, and AsqVox does not present it as a benefit. The number that matters is whether visitors got their answers, which the transcript shows and the bounce rate does not.
Visual
Two bounces, opposite meanings
One number, two opposite meanings, and no way to separate them from the metric alone.
Statistics
Every figure carries its source and year. Vendor numbers are labelled as vendor numbers, and where no reliable figure exists this page says so rather than borrowing one.
There is no reliable cross-industry bounce rate benchmark worth quoting. Figures vary by industry, page type and traffic source, and critically by the analytics platform definition, which has changed.
-no reliable figureThis is the framing that should lead any use of the metric. Quoting a cross-industry bounce average implies a comparability that does not exist across page types or platform generations.
The bounce rate benchmarks by industry that circulate widely generally trace back to vendor aggregates with undisclosed methodology.
-no reliable figureIf a bounce benchmark is quoted at you, ask which platform definition it used and how the sample was drawn. The trail usually stops at a vendor aggregate that does not say.
Analytics platforms have shifted from single-interaction definitions toward engagement-based ones, which means bounce figures across platform generations are not comparable.
-no reliable figureA definitional change, not a missing number, but it has the same effect: a period-over-period bounce comparison can be entirely invalid if the definition moved between the two points.
Zero-click search reached 68.01 percent of US Google searches in early 2026, up from 60.45 percent in 2024.
68.01%independentSparkToro with Similarweb data, published June 2026, 2026 - Independent third-party measurement, and the one firm figure on this page. It matters here because visitors arriving after an AI answer are later-stage, which changes the on-site behavior a bounce rate is quietly assumed to reflect.
Examples
In practice
A team targets bounce rate reduction and adds interstitial elements that generate interaction events. Bounce rate falls substantially and conversions do not move, because the additional interactions were with the intervention rather than with the content. The metric improved and nothing else did.
The everyday version
Bounce rate counts visitors who looked at one page and left. The problem is that it treats someone who found exactly what they wanted and left happy the same as someone who landed on the wrong page and left immediately. Chasing the number down is usually a waste of time.
Usage
Who says it
Who uses the term
- Marketing and web teams, universally, in analytics reporting.
- Agencies, in performance reporting, sometimes as a proxy for quality.
Where it turns up
In a reporting pack
- In analytics dashboards and marketing reports.
- Occasionally in agency contracts as a performance measure, which is unwise given how easily it is gamed.
Common misuse
What it gets used for that it should not
- Treating lower as better without segmenting by page type.
- Comparing figures across analytics platform generations.
- Optimizing it directly, which is easily gamed and rarely moves a real outcome.
Questions people ask
What is a good bounce rate?
There is no reliable cross-industry benchmark worth quoting. Bounce rate varies by industry, page type and traffic source, and critically by the analytics platform definition, which has changed. A reference page or a blog post that fully answers a question should have a high bounce rate and high satisfaction, so the right target depends entirely on what the page is for.
Is a high bounce rate bad?
Not necessarily. Bounce rate counts single-interaction sessions, which include both complete failure and complete success. A visitor who read a page for four minutes, got their answer and left is a bounce, and so is a visitor who landed on the wrong page and left in two seconds. The metric cannot tell them apart, so a high figure needs to be read against page type before it means anything.
Why is bounce rate unreliable?
Two reasons. It cannot distinguish a satisfied departure from a wrong-page exit, because both are a single interaction. And analytics platforms have shifted from single-interaction definitions toward engagement-based ones, so figures from different periods or platforms are frequently not comparable. Always check a historical bounce comparison for a definition change before trusting it.
What should I look at instead of bounce rate?
Scroll depth and time on page, which separate reading from leaving. Return visits and branded search, which indicate the page worked. And conversion, where a conversion goal exists. Together these tell you whether the visitor got what they came for, which the single bounce number cannot.
Last reviewed 4 August 2026. Written and reviewed by Dhruv Dholakia, founder of AsqVox.