The common assumption is that a high figure means visitors were disappointed. That is sometimes true and frequently is not. A page answering a question completely produces exactly this pattern: somebody arrives from a search, finds what they needed, and leaves satisfied. Counting that as a failure inverts the meaning entirely.
The number only tells you something where a second action was required. A home page whose job is to send visitors to a service page, or a landing page that exists to produce a form submission, should not be leaving people with nothing further to do. On those pages a high figure is a genuine signal. On an article, a directions page, or a set of opening hours, it is noise.
There is a technical detail that has changed how this behaves and many businesses have not noticed. Newer analytics counts engagement rather than bounces, treating a visit as engaged if it lasts beyond a threshold, includes a conversion, or covers more than one page. That means a visitor who read carefully for three minutes and left is no longer counted the same way, and figures from older systems are not comparable with newer ones.
Look for the extremes rather than at the average, because a middling figure across a site tells you nothing actionable. A page at nearly a hundred percent with meaningful traffic is worth investigating, and so is one at a very low figure, since that frequently indicates a tracking error such as the code appearing twice.
Segment before concluding anything. The rate for search traffic on an informational page and for advertising traffic on a landing page describe different situations, and the site average combines them into a number that describes neither. If you look at this at all, look at it per page and per source.
Check the mobile figure separately, since a large gap between devices frequently indicates a genuine problem rather than a behavioural difference. A page performing well on desktop and losing visitors immediately on a phone usually has a layout, speed, or tap target issue, and that comparison is one of the few reliable diagnostics this number provides.
Consider what you would actually change before spending time on it, which is the practical test. If a high figure on a page would cause you to add an internal link, a clearer next step, or a faster load, the number is useful. If you would do nothing differently, it is a statistic rather than information.
Then measure enquiries instead, which is what the business actually depends on. A page with a high bounce rate producing two enquiries a month is outperforming one with excellent engagement producing none, and no amount of interpretation of the first number changes that comparison.
Compare the same page against itself over time rather than against any benchmark, since the figure varies so widely by page type and traffic source that external comparisons are meaningless. A page whose rate rose sharply after a change is telling you something specific, and that is the only version of this number that reliably produces an action.
Check whether the tracking code appears twice if a figure looks impossibly low, since duplicate installation registers a second interaction on every visit and makes every page look highly engaging. That error is common after a site rebuild where the code was added both to the template and through a plugin.