How visitors behave on a site and how they interact with the search engine is a very important factor.
Sources of information:
- Analytics systems (like Google Analytics).
- Browswers (like Google Chrome).
- Plugins to browsers.
It is very natural for search engines to be interested what people do on the sites after they clicked a link in the search results.
Google may have been taking behaviour into account since the may of 2007.
What can paid attention to:
- Click through ratio (CTR) of the domain in the search results.
- CTR of a web page in the search results.
- Standard deviation from the average time spent on a document for a query or a set of queries.
- Average satisfaction – whether people return to the search engine to look for the same query again.
- How many pages of a web site visitors usually read.
- Do visitors interact with a web page?
As soom as it became obvious that behavioral factor has become very important, webmasters immediately started imitate behaviuor (by using special services or even hiring people).
Search engines immediately introduced filters for this malpractice.
Google is reluctantl even to speak about behaviour. But other search engine profess to make allowance behaviour (for example, Yandex).
Vitious circle – dependency on feedback. The more visitors, the better behaviour, the more visitors, the more visitors, the better behaviour.
It is very hard for a young web site to be ranked high. But for a search engine it is important not only show good and trusted sites, but discover good young web sites as well.
The solution is the mathematical task multi-armed bandit problem. Web sites for which a search engine does not have enough data, from time to time get a chance to be visible to users. In short-term period it deteriorates the quality of search results. But in the long run it betters search results.