The problem is that, as I very difficult to measure. When we improve bounce rate on our pages, we not only affect Google but also the user, and therefore we will not only improve bounce rate but probably links, tweets and mentions of all kinds.
On the other hand, the actions
We take to improve bounce rate will also affect the positioning of our pages, so we do not know what part comes from improving bounce rate and what part comes from on-page changes.
Now, while we don’t have evidence that
Google uses our pages’ bounce time as a office 365 database ranking metric, we do have evidence that it uses the user’s bounce time on the SERP after entering a query.
That is, Google measures the quality of search results bas
On metrics extract from users. It takes into account whether the current results are satisfactory for the user by measuring the time spent in that SERP. If after performing a search, the user quickly jumps to another search after seeing that than you afford give yourself the first search did not show what they want, Google considers this a bounce and uses it as a metric to evaluate the quality of the SERP for that search. Slide from a leak Google document that talks about the learning process and change in SERPs
Navboost: Clicks as a ranking factor
Navboost is one of the most fundamental pieces of Google’s ranking algorithms, having been in use since around 2005.
Its function is basically to collect data on user interaction with search results, especially through their clicks. Its main objective is to show users the most relevant and useful results for their browsing nes. In short, if a result is preferr by users for a given query (through clicks), it is understood that it deserves a better position.
By aggregating all this data and using it
To train other classification algorithms, the aim is to use Machine Learning to enable the algorithms to prict which results will be most satisfactory for the user tg data bas on the entire history collect (which is constantly updat).