Bounce rate is not a Google ranking factor.

Google cannot see it, so it cannot use it

Bounce rate is calculated inside your analytics, from your analytics tracking. In the classic definition it is the share of visits where someone landed on a page and left without triggering a second interaction the tool could record. That number lives in Google Analytics, or Plausible, or whatever you run. It is not a number Google Search has, because Search does not have your analytics, and it is not going to rank a page on data it cannot read.

People assume a link because both things carry the word Google, but Analytics and Search are separate products with a wall between them for good reason. If rankings were driven by a script any site owner installs and controls, the script would be the first thing manipulated. The whole point of the signals Google does use is that the site cannot trivially fake them.

The "dwell time" half-truth

The sophisticated version of the myth swaps bounce rate for "dwell time", the idea that Google times how long you spend on a result before returning to the search page and ranks accordingly. This is more plausible, because Google can observe a return to its own results, and it has clearly used interaction data in aggregate to evaluate its systems. But that is a long way from a per-page dwell-time dial you can move, and Google has been consistent that there is no such simple stopwatch. Treating "keep them on the page longer" as a ranking tactic is optimising a proxy for a proxy.

The honest position is that time on page is worth caring about for the user's sake, not the ranking's. If people leave instantly because the page is slow, misleading or irrelevant, fix that because it costs you customers. The ranking will follow from the page being genuinely better, not from you having nudged a number that Google was never watching.

What to measure instead

If you care whether a page works, measure the thing you actually want it to do. For a commercial page that is the enquiry or the sale, and improving that is conversion work, tested properly rather than guessed from a bounce figure. For a page whose job is to inform, the useful questions are whether it ranks for the query, whether people who arrive scroll and read, and whether they go on to something else on the site when that is the intent.

Where bounce rate earns a place is as a flag, never a verdict. A page that suddenly bounces far more than it used to is worth a look, because something changed. But the number sends you to investigate, it does not tell you the answer, and it certainly does not explain a ranking. If the page is losing positions, the causes are on the page and in the results, the same place a real ranking drop always lives, and I have written separately about how a real user-experience signal like speed is a tiebreaker rather than the lever people wish it were.

The general lesson

A number being easy to see is not the same as it meaning something, and the easiest numbers to see are the ones most often mistaken for causes. Bounce rate is real, it is in your dashboard, and it is not a ranking factor, not directly and not through a dwell-time back door. Measure what you want the page to achieve, treat the convenient proxies as prompts to investigate, and never let a single ambiguous figure stand in for the judgement it cannot make for you.