Facebook is increasing in strength - not only in terms of number of users, but also their activity on the site. How is it possible that Facebook makes us more and more? The answer is simple: he knows what we like and understand us more and more.
Mark Zuckerberg announced on Wednesday that 63 percent that day . of 1.28 billion users logged into Facebook , and the percentage of users who log on at least six days a week exceeded 50 percent .
Facebook has launched a news channel ( news feed ) in 2006 , but the " I Like it " was introduced only a year later. From that moment the service knows what is really like and what we are interested .
Since then, the news channel has changed with with a simple chronological list , the mechanism " able to study us."
Therefore, posts on our " Wallu " are ranked according to what is most likely to be interesting for us . The purpose of this was to ensure that such views baby pictures of our friend will take precedence before knowing the result of the game " Mafia Wars " achieved by another of our friends .
This mechanism , however, was far from perfect. Funny meme collected thousands of high-speed " lajków " when a reliable analysis of military conflict was punished by the algorithms Facebook, because you do not obtain the approval of members .
This meant that " walle " were filled with " junk information", where users truly hated . These posts were " clickable " headlines for gain as much polubień and " clicks " . And these lead to pages filled with spam and ads.
Facebook buys glasses transferring to a virtual reality. For two billion U.S. dollars
For $ 2 billion ... read more »
In the absence of the " I do not like this " algorithms Facebook, had no way of knowing which posts the users are not interested . Finally , a year ago, Facebook has admitted it had a problem with the quality of the content . Service began collecting new data to generate insights that old data - liking of , make comments - could not do .
By Facebookowi in the conversion algorithms are currently three data sources : surveys , A / B testing and research on the time the user spends away from Facebook when you click on the post and what he was doing when he returns .
Facebook testing variations of their algorithms by showing him a small number of users . As it turns out , may be even a thousand different versions of the site for different groups . Facebook collects data from them to see which generate positive change , and that negative reactions.
Every time you log in to Facebook, sites algorithms must choose which of the average of 1,500 posts to put on top of our news channel .
Some speculate that the changes were developed specifically for the degradation of some sites and at the same time to raise the ranking of others. - Facebook determines the quality of content is not the objective ranking system , but according to users' tastes - emphasizes Will Cathcart , who in the company of Menlo PArk oversees product management teams .